The landscape and regulatory potential of eccDNAs in mammalian preimplantation embryos

Ling Wei; Ning Wu; Lu Chen; Tao Wang; Zhipeng Zhu; Leisheng Shi; Xi Xiang; Jie Qiao; Qiang Liu; Xiaolu Zhao; Fengbiao Mao

doi:10.1038/s41467-026-71227-z

. 2026 Mar 31;17:4657. doi: 10.1038/s41467-026-71227-z

The landscape and regulatory potential of eccDNAs in mammalian preimplantation embryos

Ling Wei ^1,^2,^3,^#, Ning Wu ^4,^5,^6,^7,^#, Lu Chen ^4,^5,^6,^7,^#, Tao Wang ^1,^2,³, Zhipeng Zhu ^1,^2,³, Leisheng Shi ^1,^2,³, Xi Xiang ⁸, Jie Qiao ^4,^5,^6,^7,^9,^10,^✉, Qiang Liu ^4,^5,^6,^7,^✉, Xiaolu Zhao ^4,^5,^6,^7,^✉, Fengbiao Mao ^1,^2,^3,^✉

PMCID: PMC13201599 PMID: 41917015

Abstract

Extrachromosomal circular DNA is an emerging regulatory element implicated in genomic stability and gene regulation, yet its role in preimplantation development remains elusive. Here, we report the widespread presence of extrachromosomal circular DNA in preimplantation embryos, characterized by homologous junction sequences and originating from genomic regions enriched for active histone marks and RNA Polymerase II occupancy. Functional perturbations demonstrate that RNA Polymerase II inhibition suppresses extrachromosomal circular DNA production, whereas disruption of the Fanconi anemia pathway elevates it, suggesting that transcription-replication conflicts affect its biogenesis. Notably, extrachromosomal circular DNA levels surge during major zygotic genome activation. Synthetic extrachromosomal circular DNAs carrying putative enhancers for the zygotic genome activation genes Mycn and Egfl7, and the developmental gene Emx1, significantly upregulate the expression of their respective genes upon transfection into fibroblasts and zygotes. Collectively, this study unveils the extrachromosomal circular DNA landscape in preimplantation embryos, elucidates a transcription-replication conflict mechanism underlying its generation, and establishes its regulatory potential during mammalian preimplantation development.

Subject terms: Embryogenesis, Development, Gene regulation

Extrachromosomal circular DNA (eccDNA) is a regulatory element involved in genomic stability and gene regulation. This study identifies widespread eccDNA in early mammalian embryos. These eccDNAs arise from active genomic regions, increase during a major wave of gene activation, and can regulate key developmental genes.

Introduction

Extrachromosomal circular DNA (eccDNA), lacking centromeres and telomeres, is a circular DNA element outside the chromosomes that ranges in size from tens to millions of base pairs^1–3. Large circular DNA molecules spanning hundreds of kilobases to megabases are commonly referred to as extrachromosomal DNA (ecDNA), whereas smaller circular DNA species, typically ranging from tens to several kilobases, are generally termed eccDNA or microDNA⁴. Across various cellular environments, researchers have thoroughly investigated the genomic origins of eccDNA, revealing CpG islands, gene-dense regions, and repetitive elements like long terminal repeats (LTRs), Long interspersed nuclear element-1 (LINE-1), segmental duplications, and satellite DNA as prime sites for eccDNA formation^5–8. In epigenomic studies, higher GC content and dinucleotide patterns suggest that nucleosome packaging may influence the formation of eccDNA molecules^5,9. However, some studies present contrasting findings, reporting a near-random distribution^10,11.

Considering the dynamic and varied nature of eccDNA formation, it seems that DNA damage serves as the primary underlying factor. Several distinct mechanisms have been hypothesized, including the breakage-fusion-bridge (BFB) cycle, translocation deletion amplification processes, episome model, chromothripsis, repeat-based mechanisms, replication slippage, and the fork stalling and template switching (FoSTeS) phenomena⁴. Based on these comprehensive theories, eccDNA is linked to apoptotic events, arising from the stochastic ligation of genomic DNA fragments⁹. Central to these processes are the canonical DNA double-strand breaks (DSBs) repair pathways, such as homologous recombination (HR) and microhomology-mediated end joining (MMEJ)^5,6,12. In addition, researchers have also recognized a comparable repair mechanism to MMEJ, termed alternative end-joining (alt-EJ), which is associated with the biosynthesis of eccDNA derived from retrotransposons in Drosophila¹³.

The role of large, megabase-sized ecDNA in cancer biology is well-established. These molecules exhibit highly open chromatin structures that potently activate the transcription of carried oncogenes¹⁴. This, in turn, drives tumor progression by promoting oncogene overexpression, enhancing genomic heterogeneity, and enabling drug resistance through gene amplification^15–20. In parallel, a distinct population of small eccDNAs has been identified in both normal biological and pathological conditions^8,21. These molecules play diverse roles, including regulating gene expression, participating in DNA damage repair mechanisms¹², and serving as potential biomarkers for disease detection and monitoring^22,23. Notably, these small species are abundantly generated in the germline, as evidenced by their extensive profiles in both humans and mice^11,24.

Early mammalian embryonic development begins with fertilization, which is the fusion of two terminally differentiated gametes: oocytes and sperm. The hallmarks of early embryonic development in most species include a series of cleavage divisions with distinctive cell-cycle dynamics, together with the transition from maternal to zygotic control of gene expression. Following fertilization, the genome within the totipotent embryo remains transcriptionally silent and then initiates the comprehensive transcriptional activation, known as zygotic genome activation (ZGA), which is the first transcriptional event following fertilization in mammals. In mice, ZGA can be divided into two stages: a minor wave of ZGA occurs within 12 hours after fertilization in late one-cell embryos, and a major wave of ZGA, occurring during the mid to late 2-Cell stage^25–27. Early embryonic development is tightly regulated by genetics and epigenetics. Due to drastic and extensive epigenetic reprogramming, early mammalian embryos can undergo extensive DNA damage and repair during the preimplantation stage^28,29. The above process should be accompanied by the production of a substantial amount of eccDNA. However, the presence of eccDNA in mammalian preimplantation embryos, as well as its dynamic changes during this developmental window, has not been thoroughly investigated. Further studies are needed to determine whether there is crosstalk between the generation of eccDNA and key developmental processes such as ZGA and lineage differentiation during embryonic development.

In the current study, we comprehensively characterized the eccDNA landscape during mammalian preimplantation development through an integrated analysis of multiple displacement amplification (MDA)-based whole genome sequencing (WGS), ATAC-seq, RNA-seq, ChIP-seq, and single-cell Repli-seq datasets³⁰. Our analysis indicates that eccDNA participated in the ZGA process and modulated gene expression during preimplantation embryos. Additionally, our findings indicate a correlation between the formation of eccDNA and transcription-replication conflicts (TRCs), offering insights into the underlying mechanisms that may be pivotal to eccDNA formation.

Results

The widespread presence of eccDNA in mouse preimplantation embryos

Circular DNA can be computationally inferred from whole genome sequencing (WGS) data through the analysis of unique split reads and discordant reads³¹. To generate such data, we carried out extensive genomic sequencing using MDA on IVF-derived embryos across six pivotal stages of mouse preimplantation development: zygote, 2-Cell, 4-Cell, 8-Cell, morula, and blastocyst stages. Then we applied the Circle_finder pipeline³¹ to investigate the characteristics of genome-wide eccDNA during mouse preimplantation embryos based on our WGS sequencing data (Fig. 1A). Due to the limitation of the Circle_finder pipeline in distinguishing between extrachromosomal circles and chromosomal segmental tandem duplications^31,32, we implemented a repeat-masked genome (see Methods) to ensure the credibility of detected eccDNA. The number of eccDNAs exhibiting complete overlap between the two replicates at each stage is shown in Supplementary Fig. 1A and Supplementary Data 1. For a more in-depth exploration of this eccDNA distribution, we combined the raw data from replicate samples of each embryonic stage for thorough analysis. The results indicated that a substantial amount of eccDNA was generated during mouse preimplantation development (Supplementary Data 2). Surprisingly, we found that the amount of eccDNA was highest at the 2-Cell stage, which is a prerequisite and key event for successful preimplantation embryo development³³. The large amount of eccDNA produced at the 2-Cell stage might be closely related to the occurrence of ZGA, followed by a subsequent gradual decrease until the morula stage (Fig. 1B). In total, 399,660 unique eccDNAs (Supplementary Data 2) were identified across the six distinct stages, with 104,557 eccDNAs newly generated during the 2-Cell stage compared to the zygote stage (Supplementary Fig. 1B, left). To investigate the potential link between eccDNA biogenesis and ZGA, we analyzed the association between genes harboring newly generated eccDNAs at the 2-Cell stage and previously reported ZGA genes³⁴. This analysis revealed a significant overlap of 1,229 genes (Supplementary Fig. 1B, right; Supplementary Data 3). To assess the statistical significance of this overlap, we performed control analyses by randomly selecting an equivalent number of genes from both the entire mouse genome and a non-ZGA gene set. Fisher’s exact test demonstrated a highly significant overlap of 1229 genes between eccDNA-associated and ZGA gene sets (P value < 0.001; Supplementary Fig. 1B, right), which was not observed in random controls (Supplementary Fig. 1C). This confirms a non-random, highly specific association between eccDNA formation and ZGA. Notably, among the eccDNAs newly generated at the 2-Cell stage, several were located in key ZGA genes such as Pramel5, Ntrk2, and Egfl7. We further validated several ZGA-associated eccDNA candidates from the 2-Cell stage using reverse PCR and Sanger sequencing. These eccDNAs were located in promoter regions, specifically chr4:144,257,675–144,304,941, chr13:58,830,714–58,839,621, and chr2:26,593,442–26,594,001, which we predicted to be associated with Pramel5, Ntrk2, and Egfl7, respectively. Additional eccDNAs were located in intronic (chr13:106,824,086–106,828,831), distal intergenic (chr1:99,465,395–99,467,467), chr8:115,600,974–115,601,341, chr12:13,032,545–13,032,927, and 5′ UTR regions (chr10:82,876,168–82,877,275), which were predicted to be associated with Ipo11, Klf5, Maf, Mycn, and Eid3 (Supplementary Data 4; Supplementary Fig. 1D; Fig. 5E). These findings suggest a significant presence of ZGA-associated eccDNA in mouse 2-Cell embryos, implying a strong connection between eccDNA generation and ZGA processes. Given the lack of prior reports on eccDNA in preimplantation embryos, our subsequent focus was on analyzing its genomic features, sequences, and possible functions, providing valuable insights for future research endeavors.

Fig. 1 — A Schematic overview of the pipeline used to identify eccDNAs across developmental stages: 1C (zygote), 2C (2-Cell), 4C (4-Cell), 8C (8-Cell), MO (morula) and BL (blastocyst). Each sequencing library represents one biological sample consisting of a pool of 100 embryos. Technical replicate sequencing libraries were generated for each sample. B Number of eccDNAs detected at each developmental stage. C Size distribution of eccDNAs across developmental stages. D Bar charts showing the number of eccDNAs gained, lost, or stable at each developmental stage relative to the previous stage. Definitions of the different eccDNA types are provided in the “Methods” section. E Comparison of eccDNA lengths between stage-specific and consistent eccDNAs. Violin plots show the distribution of eccDNA lengths. Statistical significance was assessed using two-sided Wilcoxon rank-sum tests with Benjamini-Hochberg correction for multiple comparisons. ****, adjusted P (P_adj) < 0.0001. F Genomic feature distribution of consistent eccDNAs. G Significantly enriched pathways associated with consistent eccDNAs. Source data are provided as a Source Data file.

Fig. 5 — A PCA of normalized eccDNA (upper panel) and gene expression signals (lower panel) in mouse preimplantation embryos, with expression levels shown as log10-transformed TPM values. B Distinct clusters of eccDNA enrichment across developmental stages were identified using fuzzy c-means clustering based on quantitative analysis of eccDNA abundance. C Gene Ontology (GO) enrichment analysis of genes associated with the different clusters shown in (B). Statistical significance was assessed using a hypergeometric test (two-sided), with P values adjusted for multiple comparisons using the Benjamini-Hochberg method. Enriched GO terms are ranked by −log₁₀(P_adj). D Overlapping genes between the eccDNA-associated genes in Cluster 2 and ZGA genes. Statistical significance was assessed using a two-sided Fisher’s exact test. The odds ratio was 2.00 (95% confidence interval: 1.81-2.20). The P value obtained from Fisher’s exact test is reported as <2.2 × 10⁻¹⁶ due to floating-point precision limits in R. E Validation of eccDNAs corresponding to the genes overlapping in (D) by reverse PCR. Left: schematic illustration of the validated circular DNA structures. Arrows indicate the direction of designed primers, and dashed lines indicate the eccDNA junction site. This graphic was created by the co-author using Adobe Illustrator. Right: DNA bands amplified by reverse PCR were gel-purified and subsequently sequenced. Source data are provided as a Source Data file.

Characteristics of eccDNA in mouse preimplantation embryos

Initially, we confirmed that the size distribution of eccDNA from various embryos follows a consistent pattern (Fig. 1C). Across all stages, there was a notable peak percentage of eccDNA in the 100-10,000 bp range, with percentages decreasing as the length increased. This trend suggests that shorter eccDNA molecules are more abundant during these embryonic stages. Subsequently, the distribution of eccDNA per megabase across each chromosome was examined (Supplementary Fig. 1E), revealing that chromosomes 16 and 6 exhibit the highest eccDNA density. Genome-scale map analysis indicated that eccDNA is predominantly distributed in distal intergenic, introns, and promoter regions (Supplementary Fig. 1F). Next, we evaluated potential correlations between eccDNA frequency and genomic features. No correlation was identified between eccDNA and gene density (Supplementary Fig. 1G, left; R = 0.099, P value = 0.67). A moderate positive correlation was observed between eccDNA and Alu element density although it was not statistically significant (Supplementary Fig. 1G, right; R = 0.402, P value = 0.071). Previous studies have reported variable relationships between eccDNA frequency and genomic features such as gene density and Alu element content^8,11. In our dataset from early preimplantation embryos, eccDNA frequency did not exhibit a significant correlation with these genomic features. These differences are likely attributable to cell-type-specific mechanisms of eccDNA formation and to differences in genomic hotspots of eccDNA generation.

Through a comparative analysis of eccDNA dynamics between two consecutive stages (see Methods), we observed a continuous generation of novel eccDNA along with the retention of eccDNA from earlier developmental stages. Approximately 3,500 stable eccDNAs were persistent between the two stages, indicating a degree of stability in the eccDNA landscape (Fig. 1D). Nevertheless, substantial stage-specific differences were also observed, likely reflecting the distinct biological processes regulating eccDNA generation at each developmental stage. Notably, during the transitions from the zygote to the 2-Cell stage and subsequently to the 4-Cell stage, eccDNA exhibited significant dynamic changes, characterized by marked production and reduction (Fig. 1D). These changes may reflect regulatory mechanisms associated with ZGA.

The intersection of eccDNA across all developmental stages revealed 2,285 eccDNAs that were consistently present throughout, which we refer to as consistent eccDNAs. The number of these consistent eccDNAs was relatively small compared to stage-specific eccDNAs (Supplementary Fig. 2B). They tended to be longer than eccDNAs from individual stages (Fig. 1E, and Supplementary Fig. 2C), and were predominantly enriched in promoter regions relative to the overall eccDNA distribution across stages (Fig. 1F and Supplementary Fig. 1F). Notably, a recent study³⁵ reported a higher frequency of DSBs in promoter regions, suggesting a potential mechanistic link to eccDNA generation. Functional enrichment analysis using the GREAT tool revealed that genes associated with these consistent eccDNAs were significantly enriched in biological pathways related to G-protein coupled receptor signaling and insulin-activated receptor activity (Fig. 1G). Taken together, these observations suggest that consistent eccDNAs may be preferentially associated with regulatory genomic regions and signaling-related genes, thereby potentially contributing to transcriptional regulatory landscapes during preimplantation development.

Formation mechanisms of eccDNA during mouse preimplantation stages

Understanding the characteristics of eccDNA junction sites is crucial, as it provides direct insights into the mechanisms of circularization and identifies the genomic regions susceptible to eccDNA formation^5,36. For each detected eccDNA, we analyzed the 20 bp upstream and downstream sequences from both the start and end points, labeling them as Sequences I and II according to their genomic orientations (Fig. 2A). Our analysis revealed that both I and II Sequences contained 2- to 40-base direct repeats (Supplementary Data 5). We identified direct repeats of more than 7 bp around junction sites in approximately 30% of eccDNAs (Fig. 2B), in contrast to only ~9% in simulated control sequences (Supplementary Fig. 2A). Additionally, the length of these direct repeats positively correlated with the size of the eccDNA (Fig. 2C). Consistently, it has been demonstrated that longer DNA fragments exhibit increased spatial separation between free ends, necessitating longer homologous sequences to effectively facilitate the process of recombination²⁴. Similarly, several studies have also proposed that the presence of microhomology at eccDNA junctions suggests that MMEJ may contribute to the circularization of eccDNA^6,24,36.

We found that several TF motifs, including GATA4, FOXO1, and SOX15, were enriched in both Sequences I and II at the 2-Cell stage (Fig. 2D). These TFs exhibit high transcription levels and translation efficiency during the mouse 2-Cell stage, playing crucial roles in ZGA and early development³⁷. Studies have shown that GATA4 binding sites are enriched in active histone marks³⁸, and GATA4 has been demonstrated to promote chromatin decompaction and nucleosome eviction³⁹. Similarly, FOXA/FOXO family members are classified as pioneer TFs that bind nucleosomal DNA and promote chromatin opening^40,41. These TFs' binding regions are associated with high transcriptional activity and increased topological stress, making them more susceptible to DNA breakage. Therefore, the presence of TF motifs at eccDNA junctions could reflect both regulatory activity and DNA fragility.

Previous research has indicated that microDNA from adult mouse tissues tends to be enriched at promoters with activating or bivalent markers, as well as regions occupied by RNA Polymerase II (RNA Pol II)⁷. Additionally, the strong association of eccDNA in sperm with euchromatin²⁴ prompted us to investigate the relationship between eccDNA and transcriptionally active histone modifications, such as H3K27ac and H3K4me3. Reanalyzing public ChIP-seq data for H3K27ac, H3K4me3, and Pol II in mouse embryos^42–46, we found that across all analyzed stages of preimplantation development, eccDNA regions exhibited significantly higher enrichment of active histone modifications H3K27ac and H3K4me3 compared with control regions (Fig. 2E, F), as well as stronger Pol II signals relative to controls (Fig. 2G). We then categorized the reads into different types, including split reads, discordant reads, and concordant reads, and found that these two active histone modifications and Pol II binding signal were most enriched in split and discordant reads (Supplementary Fig. 2D), suggesting that these active transcription markers do indeed originate from eccDNA. We further analyzed the GC content of the regions surrounding eccDNA junction sites across developmental stages (Supplementary Fig. 2E). We found that the GC content in these regions ranges from 0.39 to 0.41, indicating an AT-rich characteristic. Notably, the 2-Cell stage showed the lowest GC content, suggesting a higher AT content at this stage. Previous studies have shown that AT-rich regions tend to form open chromatin structures and are associated with nucleosome-free regions and transcriptional activation⁴⁷. Therefore, we speculate that eccDNAs preferentially arise from such AT-rich regions, which may exhibit a more open and flexible chromatin configuration, making them more susceptible to DNA breakage and subsequent circularization.

The initial DNA replication in the zygote is essential for triggering mitosis in the preimplantation embryo. This process follows a specific temporal sequence known as the replication timing (RT) program, dividing the genome into early and late replicating regions³⁰. By utilizing the most recent genome-wide replication timing atlas of mouse embryos provided by Nakatani et al.³⁰, we delved into the complex replication patterns during early embryos. Our analysis revealed that regions overlapped with eccDNA exhibited notably higher RT values in early stages compared to those without eccDNA overlapping, indicating their robust replication activity (Fig. 2H). However, the RT values of eccDNA regions declined by the morula stage, suggesting a diversification of the sources contributing to eccDNA as development progresses.

Suppression of transcription led to a marked decrease in both the quantity and abundance of eccDNA

To further investigate the influence of transcription on the production of eccDNA, we utilized α-amanitin to inhibit RNA Pol II-mediated transcriptional activity in zygote (Supplementary Fig. 3A), and then performed MDA-based WGS and RNA-seq experiments to systematically compare their genomic and transcriptomic landscapes between the control and the blocked embryos (Fig. 3A). Firstly, we analyzed transcriptome data to determine whether inhibiting transcription affected ZGA. A strong correlation was observed between the two biological replicates within each group’s RNA-seq samples, underscoring the reliability and accuracy of the gene expression data obtained (Supplementary Fig. 3B). Our differential expressed gene (DEG) analysis revealed 4269 DEGs (P_adj ≤ 0.05; fold change ≥ 2) between the blockage and control group, with 491 genes showing upregulation and 3,778 genes displaying downregulation (Fig. 3B; Supplementary Data 6). To further explore these changes, we compared them with previously reported ZGA genes and found that 1,399 ZGA genes were downregulated (Fig. 3C), confirming that the transcriptional block impacted genes crucial for the normal progression of ZGA in embryos. In terms of eccDNA, we found a significant decrease of eccDNA quantity in the inhibition group (Fig. 3D). To further explore the dynamic changes of eccDNA, we developed a pipeline to quantify eccDNA based on its relative abundance (Supplementary Fig. 3C) (see Methods). Our analysis revealed a significant downregulation of a large amount of eccDNA (n = 4,607) following inhibition (Fig. 3E). These results indicate that the quantity and abundance of eccDNA were both significantly reduced after transcriptional inhibition with α-amanitin. Next, we explored the relationship between downregulated genes and reduced eccDNA by joint analysis of eccDNA and RNA-seq data. We observed that the reduction of eccDNA abundance was obviously correlated with the downregulation of their associated genes (Fig. 3F). With regard to key ZGA genes, we further confirmed that the reduction in eccDNA abundance notably occurred in regions associated with repressed ZGA genes, such as Btg3, Ccdc126, Mycn, and Sox11 (Supplementary Fig. 3D; Supplementary Data 7). To investigate the relationship between transcriptional repression and eccDNA reduction, we compared genes that were downregulated at the RNA-seq level with those associated with decreased eccDNA signals. This analysis revealed a highly significant overlap between the two gene sets (P value < 2.2 × 10⁻¹⁶) (Fig. 3G), indicating a strong association between eccDNA loss and gene downregulation. In contrast, a control analysis using randomly selected gene sets of the same size from the genome showed no significant overlap with the RNA-seq downregulated genes (P value = 0.47689) (Supplementary Fig. 3E), highlighting the specificity of the observed relationship. Then we performed gene ontology (GO) enrichment analysis on the intersecting genes. The top 20 enriched pathways were closely associated with the ZGA process, including RNA metabolic processes and nucleic acid metabolic processes (Fig. 3H), suggesting that eccDNA-associated transcriptional changes may play a pivotal role in regulating early preimplantation development.

Fig. 3 — A A schematic representation of the α-amanitin treatment. B Scatter plots display the RNA-seq analysis of differential gene expression between control and inhibited embryos. Each group includes two biologically independent samples, with two technical replicate measurements per sample. Differentially expressed genes were identified using significance thresholds of fold change ≥ 2 and P_adj < 0.05. C Scatter plots illustrate the RNA-seq analysis of ZGA gene expression in both control and inhibited embryos, applying the same significance criteria as in (B). D Box plot showing the number of eccDNAs in embryos treated with α-amanitin (AMA) compared with control embryos (H₂O). Each group includes two biologically independent samples, with two technical replicate measurements per sample. The centre line indicates the median; box limits represent the interquartile range (25th-75th percentiles); whiskers extend to 1.5 × IQR. Statistical significance was assessed using a two-sided Wilcoxon rank-sum test. P = 2.86 × 10⁻²; statistical significance: *P < 0.05. E Scatter plots present the differential eccDNA in control and inhibited embryos, with significance determined by a fold change ≥ 1.5 and a P_adj < 0.05. F Correlation analysis between the RNA-seq expression levels of genes and the corresponding eccDNA abundance (DAE: Differentially Abundant eccDNA, DEG: Differentially Expressed Genes). G 301 genes were simultaneously downregulated at both the eccDNA and transcript levels. Statistical significance was assessed using a two-sided Fisher’s exact test. The odds ratio was 2.44 (95% confidence interval: 2.15–2.77). The P obtained from Fisher’s exact test is reported as <2.2 × 10⁻¹⁶ by R due to floating-point precision limits. H GO analysis was conducted on the 301 genes obtained from (G), focusing on the enrichment of the top 20 pathways. Source data are provided as a Source Data file.

Increased TRCs lead to elevated levels of eccDNA, both in terms of its quantity and abundance

Given the established role of the Fanconi anemia (FA) pathway in resolving TRCs⁴⁸, we aimed to investigate the functional connection between TRCs and eccDNA biogenesis in preimplantation embryos by disrupting the FA pathway. To this end, we treated zygotes with the FA pathway inhibitor ML323 at varying concentrations. ML323 impaired mouse preimplantation development in a concentration-dependent manner: exposure to 50 µM significantly reduced developmental rates across embryonic stages, whereas 100 µM led to widespread arrest at the zygote stage (Fig. 4A). Consistent with this phenotype, the proportion of developmentally arrested embryos was significantly higher in ML323-treated groups than in DMSO controls (Fig. 4B). To confirm the on-target effect of ML323, we treated zygotes with 50 µM ML323 for 8 hours and performed immunofluorescence staining for γH2A.X, a well-established marker of DNA damage. We observed a marked accumulation of γH2A.X foci in both pronuclei (Fig. 4C), indicating the accumulation of unresolved DNA damage, which is a hallmark of TRCs, and consequently validating the effective impairment of the FA pathway. Following MDA-based WGS of both 50 µM ML323-treated and control 2-Cell stage embryos, we observed an overall increase in both the quantity and abundance of eccDNA in the treated group. Although the average number of eccDNA per sample did not change significantly (Fig. 4D), a subset of eccDNAs was significantly upregulated (Fig. 4E). Gene association analysis revealed that these upregulated eccDNAs were significantly enriched in ZGA-related genes, with 155 genes overlapping a known ZGA gene set³⁴ (Fig. 4F). To rule out the possibility that this overlap occurred by chance, we performed a randomization test by sampling an equal number of genes 1000 times. The mean number of overlapping genes with the ZGA set in these random trials was significantly lower than 155 (Fig. 4G), confirming that the observed enrichment is biologically meaningful. These results suggest that FA pathway inhibition exacerbates TRCs, thereby elevating eccDNA production, which in turn may disrupt the regulation of ZGA.

Fig. 4 — A Developmental rates of mouse embryos treated with different concentrations of the FA pathway inhibitor ML323. Embryos were collected from six female mice per experiment, and the experiment was repeated independently three times with similar results. Data from two independent experiments were used for analysis and presentation, comprising a total of 191 embryos collected from 12 female mice. Statistical analysis between groups was performed using two-way ANOVA followed by Tukey’s multiple-comparisons test (two-sided; P_adj for multiple comparisons). Data are presented as mean ± SEM, Statistical significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; n.s. indicates P > 0.05. B Representative images and morphological analysis of embryos cultured for 96 h after hCG administration (scale bar = 50 µm). C Left: Immunofluorescence detection of γH2A.X signals in zygotes from DMSO control and ML323-treated groups (scale bar = 20 µm). Right: Quantitative analysis of γH2A.X fluorescence intensity in the female and male pronuclei of the zygotes. Data were analyzed using a two-tailed unpaired Student’s t-test; the mean values of DMSO-Female, ML323-Female, DMSO-Male, and ML323-Male were 10.70, 19.71, 15.78, and 30.75, respectively. Box plots show the mean (centre line), and whiskers represent the minimum to maximum values. Data are presented as mean ± SEM, Statistical significance is presented as *P < 0.05, **P < 0.01. D Changes in the average number of eccDNAs detected per sample following FA pathway inhibition. E Volcano plot displaying eccDNA abundance changes following FA pathway perturbation. Differential expression was assessed using a two-sided exact test in edgeR, with P values adjusted for multiple comparisons using the Benjamini-Hochberg method. Each point represents a gene: up (red) indicates log₂ fold change of at least 0.585 and P_adj < 0.05; down (blue) indicates log₂ fold change of at most −0.585 and P_adj < 0.05; gray indicates non-significant changes. F Significant overlap between genes associated with upregulated eccDNAs and ZGA gene sets. Statistical significance was assessed using a two-sided Fisher’s exact test (P = 4.4 × 10⁻⁷). G Histogram showing the distribution of overlap counts between ZGA genes and randomly sampled gene sets matched in size to the upregulated eccDNA-associated gene set in (F) across 1000 randomizations. The blue dashed line indicates the mean overlap count from random sampling, and the red dashed line represents the observed overlap count of 155 genes from (F) between FA pathway perturbation upregulated eccDNA-associated genes and ZGA genes. The fine yellow dashed lines indicate the mean ± 1 standard deviation. Empirical P value calculated from 1000 randomizations is <0.001 (0/1000), indicating that the observed overlap is highly unlikely to occur by chance. Source data are provided as a Source Data file.

Potential roles of eccDNA during preimplantation development

To study the characteristics and potential biological functions of eccDNA, we quantified all eccDNA based on its relative abundance (see Methods). Principal components analysis (PCA) revealed that embryos at similar developmental stages clustered closely, whereas those at different stages separated as expected, demonstrating high data reproducibility (Fig. 5A, upper, Supplementary Fig. 4A). To systematically analyze the associations between the dynamic changes of eccDNA and RNA expression, we collected RNA-seq datasets from preimplantation developmental stages^43,49. Spearman’s correlation analysis showed a similar pattern with intra-group similarity surpassing inter-group similarity (Supplementary Fig. 4B). In contrast to the dynamic changes of eccDNA, PCA of RNA-seq data revealed a continuum of variation during mouse preimplantation development (Fig. 5A). Both the transcriptome and eccDNA atlas diverged significantly from the 2-Cell stage, coinciding with the occurrence of ZGA. These findings imply a synergistic regulatory interplay between eccDNA dynamics and gene expression.

Next, we utilized the fuzzy c-means algorithm⁵⁰ to delineate the profiles of eccDNA dynamic changes across all developmental stages. Our analysis revealed six distinct clusters, each representing unique temporal patterns of eccDNA with diverse regulatory dynamics. The heatmap demonstrated that distinct clusters of eccDNA enrichment were present across each of the six developmental stages. (Fig. 5B; Supplementary Data 8). To better understand the functional relevance of these clusters, we conducted GO enrichment analysis for each cluster. The enrichment analysis of eccDNA-associated genes in each module revealed comparable biological functions, including RNA Pol II transcription, RNA metabolic processes, reproduction, and cell differentiation (Fig. 5C), aligning with pivotal biological processes during embryonic development. Notably, Cluster 1 showed activity in cellular differentiation processes, suggesting enrichment for cell development consistent with the meiotic-to-mitotic transition characteristic of the zygote (Fig. 5C). In terms of Cluster 2, the associated genes in this module showed enrichment in the regulation of transcription by RNA Pol II and reproduction (Fig. 5C), manifesting significant transcription activity during ZGA. Moreover, a significant overlap of 569 genes was observed between eccDNA-associated genes in Cluster 2 and known ZGA genes (P value < 2.2 × 10^-¹⁶; Fig. 5D). In contrast, control analyses using randomly selected gene sets of the same size from either the whole genome (P value = 0.52937; Supplementary Fig. 4C) or from non-ZGA genes (P value = 0.67519; Supplementary Fig. 4D) showed no significant enrichment, underscoring the specificity of the observed association. Furthermore, several eccDNAs corresponding to genes within this overlapping set were randomly selected and experimentally validated (Fig. 5E). These results suggest that eccDNAs present specifically at the 2-Cell stage are preferentially associated with ZGA-related genes. eccDNA in Cluster 3 was highly expressed in 4-Cell embryos, and showed enrichment of cellular developmental process, protein localization, and regulation of nucleic acid-templated transcription (Fig. 5C). Corresponding genes in Cluster 4 showed high activity in 8-Cell embryos and were enriched for the establishment of localization and nucleic acid metabolic processes (Fig. 5C). These findings enhanced our understanding of eccDNA dynamics during preimplantation development and link eccDNA to key developmental genes, such as Rfpl4b and Zfp352 in the 2-Cell stage, Tspan1, Xist, Sall1, Tbx20, Pax6 and Gjb3 in the 8-Cell stage, and Zfp51 in the blastocyst stage^43,51. Taken together, the above analyses revealed dynamic changes of eccDNA during mouse preimplantation development, and the eccDNA dynamics at distinct stages could potentially serve as indicators of lineage specification.

Although eccDNAs have been shown to influence gene expression in certain pathological contexts such as cancer, their functional relevance during preimplantation development remains poorly understood. To explore the potential regulatory role of eccDNAs, we first compared the expression levels of eccDNA-associated genes to those of all other genes across developmental stages. We observed a significantly higher expression of eccDNA-associated genes (Fig. 6A), suggesting that eccDNAs may contribute to transcriptional activation during preimplantation development.

Fig. 6 — A Violin plot showing the distribution of gene expression levels between eccDNA-associated genes and other genes across developmental stages. Box plots are overlaid on the violins, with the centre line representing the median, box bounds representing the 25th–75th percentiles, and whiskers extending to 1.5 × IQR. Sample sizes for each group are indicated above the boxes. Statistical significance was assessed using a two-sided unpaired Student’s t-test (P < 2.2 × 10⁻¹⁶). B Heatmap displaying the abundance of eccDNA and its associated gene expression levels across developmental stages. Values were calculated as log₁₀(TPM + 1) and subsequently standardized by row (z-score). C GO analysis of genes marked by different eccDNA associations, with P values adjusted using the Benjamini and Hochberg method. D Genes overlapping between eccDNA⁺ genes and ZGA genes. Statistical significance was assessed using a two-sided Fisher’s exact test. The odds ratio was 1.65 (95% confidence interval: 1.29–2.09). The exact P value from the test was 6.37 × 10⁻⁵. E Violin plot showing expression levels of genes overlapping eccDNA⁺ and ZGA gene sets across developmental stages. Box plots are overlaid, with the center line representing the median, the box bounds representing the 25th-75th percentiles, and whiskers extending to 1.5 × IQR. Statistical significance was assessed using two-sided Wilcoxon rank-sum tests, with Benjamini-Hochberg adjustment for multiple comparisons. P_adj for comparisons with the 2C stage are: 2C vs 1C, P = 1.03 × 10⁻⁵; 2C vs 4C, P = 1.9 × 10⁻²; 2C vs 8C, P = 6.67 × 10⁻³; 2 C vs MO, P = 1.29 × 10⁻⁵. F Violin plot showing expression levels of the genes in the overlap of eccDNA⁻ and ZGA gene sets across developmental stages. Box plots are overlaid, with the centre line representing the median, the box bounds representing the 25th–75th percentiles, and whiskers indicating the minimum and maximum values. Statistical significance was assessed using two-sided Wilcoxon rank-sum tests, with Benjamini-Hochberg adjustment for multiple comparisons. P_adj for comparisons with the 2C stage are: 2C vs 1C, P = 0.00342; 2C vs 4C, P = 0.859; 2C vs 8 C, P = 0.655; 2C vs MO, P = 0.655. Statistical significance is presented as *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. G Left, XcmI restriction enzyme digestion of Linear A, Linear B, and the LAMA-synthesized circular product, analyzed by agarose gel electrophoresis, confirming successful synthesis of eccEmx1. Right, detection of eccEmx1 transfection into NIH 3T3 cells using inverse PCR. H Left, PflMI restriction enzyme digestion of Linear A, Linear B, and the LAMA-synthesized circular product, analyzed by agarose gel electrophoresis, confirming successful synthesis of eccEgfl7. Right, detection of eccEgfl7 transfection into NIH 3T3 cells using inverse PCR. I qRT‑PCR analysis of *Emx1* mRNA expression levels in 2 C embryos and NIH 3T3 cells before and after transfection with eccEmx1, respectively. The experiment was repeated three times independently. Box plots show the mean (centre line), and whiskers represent the minimum to maximum values. Data are presented as mean ± SEM. Individual data points are plotted as markers. A two-tailed paired Student’s t-test was used for NIH 3T3 cells, and a two-tailed unpaired Student’s t-test was used for 2C embryos. *P < 0.05, **P < 0.01. J qRT‑PCR analysis of *Egfl7* mRNA expression levels in 2C embryos and NIH 3T3 cells before and after transfection with eccEgfl7, respectively. K Model depicting the biogenesis and proposed functions of eccDNAs during mammalian preimplantation development. It illustrates their generation from genomic DNA via transcription-replication conflicts and associated recombination processes (homologous sequences are indicated in red), and suggests the potential regulatory roles of the resulting eccDNA in embryonic development. Source data are provided as a Source Data file.

To further investigate this possibility, we analyzed the correlation between eccDNA abundance and gene expression using Spearman’s correlation coefficient across developmental stages. Genes whose expression levels positively correlated with their corresponding eccDNAs were defined as eccDNA⁺ genes, but as eccDNA^- genes otherwise. This analysis identified 1,230 eccDNA⁺ genes and 610 eccDNA^- genes consistently across all stages (Fig. 6B). GO analysis revealed that eccDNA^- genes were primarily enriched in biological processes such as negative regulation of cellular processes, establishment of localization, and negative regulation of metabolic process (Fig. 6C, top). In contrast, eccDNA⁺ genes were significantly enriched in pathways essential for ZGA, including positive regulation of cell differentiation, MAPK signaling, and RNA polymerase II-mediated transcription (Fig. 6C, bottom). Importantly, eccDNA⁺ genes showed significant overlap with known ZGA genes³⁴ (P value = 6 × 10⁻⁵; Fig. 6D), whereas random gene sets of the same size drawn from the genome showed no such enrichment (P value = 0.7681; Supplementary Fig. 4E). These overlapping genes, hereafter referred to as eccDNA⁺ ZGA genes, exhibited markedly elevated expression specifically at the 2-Cell stage compared to all other developmental stages (Fig. 6E). To determine whether this expression pattern was attributable solely to ZGA dynamics, we performed a parallel analysis using eccDNA^- ZGA genes, defined as ZGA genes overlapping with eccDNA^- genes. Unlike eccDNA⁺ ZGA genes, the eccDNA^- ZGA genes only exhibited significantly increased expression at the 2-Cell stage relative to the zygote stage, but not when compared to later stages (Fig. 6F). This distinction supports a specific association between eccDNA and enhanced transcription of a subset of ZGA genes, suggesting that eccDNA may play a role in fine-tuning gene expression during genome activation in preimplantation development.

Based on the above findings, we next asked whether eccDNA could enhance gene expression during preimplantation development. To test this, we synthesized eccDNA molecules eccMycn, eccEgfl7, and eccEmx1 using Ligation-Assisted Minicircle Amplification (LAMA)⁵² (Supplementary Fig. 4F, G and Fig. 6G, H). These constructs carry potent enhancers of the ZGA genes Mycn and Egfl7, as well as the developmental gene Emx1, identified in our initial mapping. The eccDNAs were then transfected into the mouse embryonic fibroblast cell line NIH 3T3 and/or zygote. We found that the mRNA levels of these genes were significantly higher than those in the control group (Fig. 6I, J, Supplementary Fig. 4H). Together, these results suggest that eccDNA contributes to ZGA and may play a broader role in regulating gene expression during preimplantation development.

Collectively, through a dynamic analysis of eccDNAs in mouse preimplantation embryos, we provide a comprehensive landscape, propose potential mechanisms for their biogenesis, and suggest their possible regulatory functions, as summarized in our model (Fig. 6K). As detailed in this study, embryonic eccDNAs are enriched in regions marked by active transcription and replication signals, indicating hotspots for both processes. The frequent collisions between transcription and replication machineries are known to cause genomic instability, including R-loop formation and DNA damage^53–55, which have been proposed as potential triggers for eccDNA biogenesis in other systems⁵⁶. However, the precise molecular pathways leading from TRCs to eccDNA formation in preimplantation embryos remain to be fully elucidated. Multiple replication-associated mechanisms, including fork stalling, template switching, and replication slippage, may underlie the generation of eccDNAs from these genomic regions^7,57,58. Importantly, while error-prone repair of DNA double-strand breaks (e.g., via NHEJ, MMEJ, or SSA) can generate eccDNAs, such events often cause insertions, deletions, or loss of heterozygosity (LOH)^59,60. Given that large-scale LOH is highly deleterious in early embryos, it is unlikely to be a major source of the eccDNAs observed in healthy development. Taken together, we propose a working model wherein eccDNAs during preimplantation development arise primarily from replication-associated processes triggered by transcription-replication conflicts, without necessarily invoking catastrophic LOH. To elucidate the exact mechanisms (such as R-loop mediation and replication slippage) operating during preimplantation development will be an important goal for future research.

Characteristics of eccDNAs during preimplantation development are conserved between human and mouse

Researchers have validated that the ATAC-seq experiment is a feasible and sensitive method for detecting eccDNA³¹. We also collected public ATAC-seq datasets covering key stages of mouse preimplantation development to identify eccDNA (Fig. 7A). The collected stages ranged from early cleavage (2-Cell, 4-Cell, 8-Cell) to the blastocyst stage, represented by the inner cell mass (ICM), and also included embryonic stem cells (ESCs) derived from the ICM. A total of 20,384 eccDNAs were detected (Supplementary Fig. 5A; Supplementary Data 9). We found that the quantity of eccDNA reaches its peak in the ICM cells of the blastocyst stage, which corresponds to the ICM undergoing rapid cell divisions and cell fate decisions⁶¹. To verify the robustness of our eccDNA identification pipeline from ATAC-seq data, we directly confirmed the presence of eccDNA in mouse 2-Cell and 4-Cell embryos using reverse PCR (Supplementary Fig. 5B, 5C). These eccDNAs had a large specificity (Supplementary Fig. 5D) with only eight eccDNAs overlapping across all stages. The sizes of eccDNA varied from tens of bases to tens of megabases, with the majority of them being around ~1,000 bp (Supplementary Fig. 5E). The chromosomal distribution of eccDNA revealed the presence of eccDNA sequences originating from all chromosomes (Supplementary Fig. 5F). The eccDNA detected in the ATAC-seq data was primarily distributed in the promoter region (Supplementary Fig. 5G), displaying a different pattern compared to the eccDNA mainly located in the distal intergenic region identified in WGS data (Supplementary Fig. 1F). The difference can be attributed to experimental technique, as ATAC-seq specifically aims to enrich open chromatin regions in the genome. Furthermore, H3K27ac, H3K4me3, and Pol II signals were highly enriched within eccDNA regions identified from the ATAC-seq dataset (Supplementary Fig. 5H). Together, these results indicated that the identification of eccDNA in ATAC-seq datasets was reliable, and the characteristics of these eccDNAs were consistent with our observations from WGS experimental datasets.

Fig. 7 — A Diagram illustrating the identification of eccDNAs from public ATAC-seq datasets. B The number of eccDNA identified at specific developmental stages. C Length distribution of eccDNA. D Distribution of eccDNA across various genomic features. E Histogram showing the distribution of rate ratio (left) and odds ratio (right) from 1000 random permutations of human eccDNAs and shared de novo mutations (DNMs). Black vertical lines indicate the observed rate ratio and odds ratio for autism spectrum disorder (ASD) relative to controls. Statistical significance was assessed using a one-sided permutation test, with empirical P values calculated as the proportion of random permutations equal to or exceeding the observed values. F Upset plot of eccDNA from human early embryos. G Enrichment of eccDNAs marked with H3K4me3, and H3K27ac. Source data are provided as a Source Data file.

Due to ethical constraints and challenges in obtaining human embryos, we further demonstrated the widespread presence of eccDNA during human preimplantation development by analyzing public ATAC-seq datasets (Fig. 7A). The analyzed stages included the zygote, 2-Cell, 4-Cell, and 8-Cell stages, as well as the blastocyst stage represented by the inner cell mass (ICM) and its derivative embryonic stem cells (ESCs). Overall, we detected a total of 5,475 eccDNAs in human embryos, underscoring the broad distribution of eccDNA during the initial stages of human development (Fig. 7B; Supplementary Data 10). The highest abundance of eccDNA was observed in ICM cells at the blastocyst stage, consistent with the above analysis in mouse. Regarding their sizes, eccDNA exhibited a diverse range, varying from tens of bases to tens of megabases, with the majority falling around the 1,000 bp mark (Fig. 7C). Notably, chromosomes 17 and 19 exhibit relatively high eccDNA density (Supplementary Fig. 5I). The distribution of eccDNA in the designated genomic elements is consistent with that obtained from mouse ATAC-seq data (Fig. 7D).

Human developmental disorders, such as autism spectrum disorder (ASD), are strongly influenced by genetic and epigenetic perturbations⁶². Numerous large-scale sequencing studies have demonstrated that de novo mutations (DNMs) play a major role in ASD etiology⁶³. Meanwhile, eccDNA has been increasingly recognized as an important contributor to gene regulation and chromatin organization, with potential impacts on genome stability and cellular processes^1,8. To investigate whether eccDNAs may intersect with genomic regions relevant to neurodevelopmental pathogenesis, we analyzed the overlap between eccDNA regions and DNMs implicated in ASD. DNMs associated with ASD were obtained from the Gene4Denovo database⁶⁴. Our analysis revealed that ASD-related DNMs significantly overlapped with eccDNA identified in human preimplantation embryos, compared with simulated random controls (Fig. 7E). Although this observation suggests a non-random spatial association between eccDNA-forming regions and DNMs involved in neurodevelopmental disorders, the eccDNAs analyzed in this section were identified based on ATAC-seq data, which inherently biases detection toward open chromatin regions enriched for gene activity. Additionally, eccDNA populations are highly heterogeneous across developmental stages. Further investigation using stage-specific eccDNA datasets and functional validation will be essential to determine whether eccDNA formation contributes to the etiology of human developmental disorders.

eccDNA from human ATAC-seq datasets also exhibited remarkable specificity, with only three overlapping sequences detected (Fig. 7F). Similarly, analysis of active histone modifications during human preimplantation stages, including H3K27ac and H3K4me3, showed elevated signals over these regions (Fig. 7G).

These findings suggest a widespread presence of eccDNA during mammalian embryonic development. The uniformity in characteristics and distribution patterns of eccDNA in both human and murine systems strongly supports the hypothesis that eccDNA may serve crucial physiological roles during mammalian preimplantation embryonic development.

Discussion

Our study examines the distribution patterns, characteristics, and the dynamic changes of eccDNA during human and mouse preimplantation embryonic development. Through multi-omic analyses, we proposed that eccDNA in preimplantation embryos may arise from intense TRCs and R-loop formation during transcription, providing valuable insights into the mechanism of eccDNA generation. Furthermore, we revealed that the production of eccDNA at the 2-Cell stage is coupled with ZGA events, suggesting a potential regulatory role for eccDNA during the ZGA process. The above findings deepen our understanding of eccDNA production and function during mammalian embryonic development. In addition, our research also offers a valuable resource of eccDNAs for embryonic development studies.

Because eccDNA detection can be susceptible to technical and analytical artifacts, we systematically evaluated multiple potential sources of bias. To evaluate the effect of sequencing depth, we randomly sampled different proportions of sequencing reads and observed that eccDNA detection had not fully reached saturation in most stages, whereas the 8-Cell and morula stages showed comparatively higher saturation levels (Supplementary Fig. 6A). Variations in eccDNA patterns across identical stages could be attributed to a combination of factors. These include the random processes governing eccDNA generation, the molecular heterogeneity present in preimplantation embryos, and the limitations of WGS as a non-specific method. These factors collectively complicate the identification and quantification of eccDNA signals within the extensive sequencing data³². High-throughput technologies frequently overlook low-frequency eccDNA and are vulnerable to contamination by linear chromosomal DNA.

A central concern in eccDNA research is the potential for artifactual DNA circles to arise during in vitro amplification, such as through template-switching events inherent to certain polymerases. To mitigate this, we employed several strategies. First, we utilized a repeat-masked genome during our computational pipeline to prevent the misidentification of tandem repeats as eccDNAs. Second, the requirement for both split reads and discordant reads for eccDNA calling provides orthogonal validation from a single sequencing library. Most importantly, the biological relevance of our eccDNA set is supported by its non-random genomic features: its specific enrichment at loci marked with active histone modifications (H3K4me3, H3K27ac) and RNA Pol II, its dynamic changes in response to Pol II inhibition and FA pathway disruption, and its significant association with ZGA genes and processes. While no single method can completely eliminate all potential artifacts, the consistent and biologically meaningful patterns observed across multiple, independent datasets and experimental perturbations provide confidence that the majority of the reported eccDNAs are unlikely to represent technical artifacts.

To exclude the possibility that the eccDNAs identified in this study originated from residual maternal components, such as discarded polar bodies, we analyzed MDA-amplified data from metaphase II (MII) oocytes in another unpublished study of ours. These results revealed minimal overlap between eccDNAs detected in MII oocytes and those identified across embryonic stages (Supplementary Fig. 6B). Although a slightly higher overlap was observed at the 8-Cell stage, the contribution of MII-derived eccDNAs was limited and did not significantly affect our overall conclusions. Furthermore, our study primarily focuses on early developmental stages, particularly around the 2-Cell stage, where the influence of MII-derived eccDNAs is negligible. In addition, we collected eccDNAs from sperm reported in a previously published study²⁴ and performed overlap analysis with eccDNAs identified at various embryonic stages. This analysis revealed minimal overlap (Supplementary Fig. 6C), suggesting that the eccDNAs detected in our study are unlikely to originate from sperm.

To further assess the genomic stability of early embryos and the potential origin of eccDNA, we performed structural variant (SV) analysis (including deletion, duplication, inversion, and translocation) in our MDA-based WGS datasets using LUMPY⁶⁵. The results revealed a low number of SVs across all developmental stages, suggesting overall genomic stability (Supplementary Fig. 6D). Moreover, overlap analysis between the identified eccDNAs and SV regions showed minimal intersection, indicating that the eccDNAs are not derived from large structural rearrangements (Supplementary Fig. 6E). We also performed copy number variation (CNV) analysis with Ginkgo⁶⁶. No significant segmental copy number gains were observed across developmental stages (Supplementary Fig. 6F), suggesting that the genome structure remained largely stable throughout preimplantation development. Given that WGS data can also be analyzed using the AmpliconArchitect (AA) pipeline⁶⁷, we further applied AA to detect potential large circular ecDNAs across embryonic stages. Although multiple amplicons were detected in each sample, the vast majority were classified as either linear amplifications or invalid amplicons, with no definitive ecDNA events identified (Supplementary Data 11). In some samples, such as those from the 2-Cell stage, cyclic structures were observed; however, these were likewise not classified as ecDNA. To illustrate the structural complexity observed in specific cases, we further examined representative amplicons from the 4-Cell stage, which contained more complex rearrangements. For example, 4C_amplicon7 displayed features consistent with a breakage-fusion-bridge (BFB) pattern (Supplementary Fig. 6G), whereas 4C_amplicon11 represented a typical linear amplification (Supplementary Fig. 6H). The absence of ecDNA events across all samples, together with the overall lack of significant CNVs, further supports the structural integrity of the embryonic genome and validates the specificity of eccDNAs identified through our detection pipeline. Together, these analyses support that the identified eccDNAs predominantly reflect developmentally regulated biological signals rather than technical artifacts.

A limitation of this study is that all embryos were generated by IVF and cultured in vitro. We followed an optimized culture protocol and included only morphologically normal embryos. In addition, embryos were collected at strictly matched developmental stages and processed in parallel to minimize experimental variation. Nevertheless, we cannot completely exclude the possibility that some eccDNA features are influenced by the IVF procedure itself. Importantly, our key findings, such as the timing and pattern of ZGA-associated transcriptional activation, are highly consistent with previously reported data based on both IVF and in vivo-derived embryos, supporting the biological validity of our results. Our decision to use IVF embryos was driven by practical considerations, including the need for precise developmental staging and sufficient material for sequencing, requirements that are technically challenging to achieve with in vivo-derived embryos. Nevertheless, we acknowledge that naturally fertilized embryos represent the physiological baseline, and the lack of direct comparison with in vivo-derived embryos is a limitation. Future studies using in vivo flushed embryos will be important to validate and extend our findings.

In addition, certain limitations of this study should be acknowledged. The scarcity of preimplantation embryo samples and ethical constraints presented challenges in obtaining an adequate number of embryos for Circle-seq assay. This limitation hindered our ability to directly identify eccDNA in human embryos using this established technology, a well-known technical bottleneck in embryonic development research. In addition, our analysis data strongly suggests that eccDNA may have crucial physiological functions during embryonic development. However, due to current methodological limitations, technology for knocking out specific eccDNA has not yet been developed. Therefore, there is an urgent need to develop low-input, high-sensitivity eccDNA sequencing technologies and state-of-the-art methods for loss-function study of specific eccDNA. Such advancements will greatly promote our in-depth study of the physiological functions and regulatory mechanisms of eccDNA in mammalian preimplantation embryonic development.

Methods

Animals and ethics

All animal experiments were conducted in accordance with relevant ethical regulations and were approved by the Institutional Animal Welfare and Ethics Committee of Peking University Third Hospital (approval number A2024145).

Five-week-old female mice and three-month-old male mice (C57BL/6J,) were provided by the Department of Laboratory Animal Science, Peking University Health Science Center (Beijing, China). Mice were reared in a temperature-controlled room (a 12-hours light–dark period), with a temperature of 18–23 °C and 40–60% humidity, with food and water ad libitum under specific pathogen-free conditions.

To ensure sufficient embryos for all experimental stages and replicates, approximately 140 mice were used in total, including 70 females and 70 males. Each breeding pair yields ~20 gametes or embryos. For this study, we planned to collect 100 gametes/embryos per stage (oocyte, zygote, 2-Cell, 4-Cell, 8-Cell, morula, blastocyst), with two experimental replicates, totaling 1,400 gametes/embryos. This number provides adequate biological replicates and ensures strain stability.

Sex was considered in the study design (oocytes were collected from female mice and sperm from male mice), but individual experimental data were not disaggregated by sex.

Following the completion of experiments, mice were euthanized by cervical dislocation, and carcasses were disposed of in designated biomedical waste containers, in accordance with institutional guidelines.

Oocyte collection, in vitro fertilization and embryo culture

For the collection of ovulated oocytes, five-week-old C57BL/6J female mice were intraperitoneally injected with 5 IU pregnant mare’s serum gonadotropin (PMSG, Ningbo Second Hormone Factory, China) followed by 5 IU human chorionic gonadotropin (hCG, Ningbo Second Hormone Factory, China) 46-48 hours later. The MII oocytes were collected from oviducts 14-16 hours after hCG injection and then digested away from the cumulus with 0.3 mg/ml bovine testicular hyaluronidase in M2 medium for 1 minute. Ovulated oocytes were then transferred into fresh M2 medium (Sigma-Aldrich, M7167) at 37 °C in 5% CO₂.

Sperm was released from the cauda epididymides of 3-month-old male mice with proven fertility in human tubal fluid (HTF, Millipore, MR-070-D) medium in oil and capacitated for 1 hour at 37 °C in 5% CO₂. Ovulated MII oocytes harvested as described above were then incubated in 250 μL HTF with 5.14 mM calcium containing capacitated sperm (～4 × 10⁵/mL) for 6-8 hours at 37 °C in a humidified atmosphere of 5% CO₂. Upon fertilization, fertilized zygotes with clear pronuclei were transferred into fresh KSOM (Millipore, MR-106-D) medium for in vitro culture. Embryos at different stages were counted and collected at 40 hours (2-Cell stage), 54 hours (4-Cell stage), 72 hours (8-Cell stage), 84 hours (morula), and 96 hours (blastocyst) post hCG injection. Images were captured using an Axio Observer 3 microscope (Zeiss) with ZEN Blue Lite imaging software.

Multiple displacement amplification (MDA), library construction, and sequencing

Following the manufacturer’s recommendation, whole metagenome amplification was performed on 6 samples (zygote, 2-Cell, 4-Cell, 8-Cell, morula, and blastocyst) using MDA with the REPLI-g Single Cell kit (Qiagen, 150345). Briefly, for cell lysis and lysed genomic DNA (gDNA) denaturation, 4 µL cell material (supplied with PBS) was mixed carefully by flicking the microcentrifuge tube with 3 µL buffer D2 (1 M DTT and buffer DLB, denaturation buffer). Cell lysis was incubated for 10 minutes at 65 °C, and 3 µL stop solution was added. To the total denatured gDNA (10 µL), 40 µL of master mix (9 µL H₂O, 29 µL REPLI-g sc reaction buffer Solution and 2 µL REPLI-g sc DNA polymerase) was added to the reaction. This solution (50 µL) was gently mixed and incubated at 30 °C for 8 hours and heat-inactivated at 65 °C for 3 minutes. The gDNA concentration and quality were measured by Qubit 4.0 Fluorometer (Thermo Fischer Scientific, USA), and the Integrity of the DNA sample was evaluated by agarose gel electrophoresis (Agarose concentration 1%, voltage 120 V, 45 minutes). After quality control, the library was prepared using 0.2 µg gDNA as template according to the Watchmaker DNA Library Prep Kit with Fragmentation (7K0019-096). DNA concentration of the library was measured in Qubit 4.0, and fragment size was assessed using the Fragment Analyzer system. Q-PCR was performed using ABI Quant Studio 12 K Flex to accurately quantify the effective concentration of the library (the effective concentration of the library > 10 nM). The cluster generation and sequencing were performed on the Novaseq 6000 S4 platform at Annoroad Gene Technology Company (Beijing, China), using the NovaSeq 6000 S4 Reagent kit V1.5.

Candidate eccDNA validation by PCR

eccDNAs of interest were verified using reverse PCR and Sanger sequencing. The reaction system is 25 µL, consisting of 1 µL DNA product, 2 µL forward primer (5 µM), 2 µL reverse primer (5 µM), 12.5 µL 2 × Rapid Taq main mixture (Vazyme), and the corresponding amount of ddH₂O. The hot cycle PCR procedure was 95 °C, 2 minutes, 35 cycles of (95 °C 15 seconds, 55 °C 15 seconds, 72 °C 5 seconds), 72 °C for 5 minutes and 4 °C hold.

eccDNA synthesis via Ligation-Assisted Minicircle Amplification (LAMA)

The synthetic eccDNA was produced using the LAMA method⁵². The random 700 bp DNA sequence for control was generated with “random DNA sequence generator” (http://www.faculty.ucr.edu/~mmaduro/random.htm) web tools with a 50% guanine–cytosine content. The half-complementary linear DNA fragments of each eccDNA and its corresponding PCR primers were synthesized in GENEWIZ. Sequences of linear DNA fragments and PCR primers were listed in the Supplementary Data 12. LAMA reaction involves mixing equal amounts of linear A and B amplification products with Taq DNA ligase (NEB) and buffer. The thermal cycle system used is 5 minutes at 95 °C, (95 °C 20 seconds, 4 °C 1 minute, 55 °C 20 minutes) with 10 cycles. The residual linear DNA was removed by exonuclease digestion of V (NEB), and the LAMA reaction products were purified by magnetic beads (Vazyme). Then the synthetic eccDNA was digested by certain restriction enzymes using the linear fragments as controls to verify its circular structure. EccDNA with only one linear band after restriction enzyme digestion was selected for subsequent experiments.

Transcription inhibition assay

For inhibition of both minor and major ZGA, embryos were exposed to 0.1 mM α-amanitin (MedChemExpress, HY-19610) from the zygote stage at 6 hours post in vitro fertilization (IVF) until the 2-Cell stage at 35 hours post-IVF for Smart-seq2 (n = 15 for each group) and WGS (n = 100 for each group) analysis. The same volume of solvent (H₂O) was used as a control treatment for the embryos. We used microscopy to verify the effect of α-amanitin on embryonic development at 40 hours post-IVF, when the control embryos had reached the 4-Cell stage.

scRNA-seq library preparation and sequencing

The RNA-seq libraries were constructed from both control and inhibition embryos following the Smart-seq2 protocol⁶⁸. Cells were lysed using a hypotonic lysis buffer (Amresco, M334), and polyadenylated mRNAs were captured with PolyT primers. After incubating the lysate for approximately 5 minutes at 72 °C, reverse transcription reactions were carried out using the Smart-seq2 approach. The cDNA underwent pre-amplification and purification with AMPure XP beads, after which it was fragmented using a Covaris system and prepared for sequencing with the Illumina TruSeq library kit. Sequencing was performed on NovaSeq™ X Plus sequencers following the manufacturer’s guidelines. Each sample was analyzed in two separate biological replicates.

eccDNA microinjection and RT-qPCR in NIH 3T3 cells

Mouse embryonic fibroblasts NIH 3T3, purchased from the Central Laboratory of Peking University Third Hospital, were cultured in DMEM medium (HyClone) supplemented with 10% fetal bovine serum (FBS) (Biological Industries (BI)) and 1% penicillin/streptomycin (Invitrogen). Cells were cultured in an incubator at 37 °C with 5% CO₂. Lipomaster 2000 transfection reagent (Vazyme) was used. For each transfection, 500 ng eccDNA and 1.5 μL Lipofectamine 2000 were diluted with Opti-MEM (Gibco) to a total volume of 25 μL, then the diluted DNA was added to the diluted lipofectamine in a ratio of 1:1. After incubation for 10 minutes, the transfer solution was uniformly injected into the 60,000 adherent cells in 24-well plates. Each group was performed with three technical replicates. Total RNA was extracted with FastPure Cell/Tissue Total RNA Isolation Kit V2 (Vazyme) 48 hours after transfection, and was reverse transcribed into cDNA according to the instructions of HiScript III 1st strand cDNA Synthesis Kit (+ gDNA wiper) (Vazyme). qPCR was carried out using ChamQ Universal SYBR qPCR Master Mix (Vazyme). All primers used in the qPCR experiments were listed in Supplementary Data 12.

eccDNA transfection and RT-qPCR in embryo samples

Zygotes were pre-incubated in M2 medium before microinjection. Using an Eppendorf Transferman NK2 micromanipulator, approximately 5 pL of eccDNA solution (300 ng/µL) was microinjected into the cytoplasm of each zygote. Post-injection, embryos were cultured in KSOM medium at 37 °C under 5% CO₂. cDNA was synthesized from embryonic microsamples following established protocols⁶⁹. Briefly, groups of five or ten embryos were collected, washed three times in PBS containing 0.2% BSA, and lysed in 2 μL of lysis buffer (0.2% Triton X-100 supplemented with RNase inhibitor). Random primers and dNTP mix were added to the lysate, followed by hybridization in a PCR thermal cycler. Reverse transcription was performed using PrimeScript II Reverse Transcriptase (Takara) according to the manufacturer’s instructions. Quantitative PCR analysis was carried out with Power SYBR Green PCR Master Mix (Applied Biosystems) on an Applied Biosystems Real-Time PCR system.

Immunofluorescence staining and confocal microscopy

For immunofluorescence, embryos were fixed in 4% paraformaldehyde (in PBS) for 30 minutes at room temperature (RT), permeabilized with 0.3% Triton X-100 (in PBS) for 15 minutes, and blocked with 1% BSA (in PBS) for 1 hour. Samples were then incubated overnight at 4 °C with primary antibodies against phospho-H2A.X (γH2A.X, #9718S; Cell Signaling Technology; 1:400 dilution in blocking solution). After three washes in PBS, zygotes were incubated with fluorophore-conjugated secondary antibodies containing Hoechst 33342 for 30 minutes at RT. Finally, samples were washed, mounted on glass slides using SlowFade Gold Antifade Mountant (Life Technologies), and imaged on a Zeiss LSM980 confocal microscope. Fluorescence intensity was quantified using ImageJ software.

Embryo culture with ML323 inhibitor

ML323 (SML1177, Sigma-Aldrich) is a potent and selective inhibitor of the USP1–UAF1 deubiquitinase complex^70,71. Previous studies in somatic cells demonstrated that treatment with a final concentration of 30 μM for 3–6 hours effectively inhibits USP1–UAF1 activity^72,73. For embryo culture, ML323 was first dissolved in dimethyl sulfoxide (DMSO) to prepare a 10 mM stock solution. Prior to use, the stock was diluted with KSOM embryo culture medium to working concentrations of 10 μM, 50 μM, and 100 μM. These working solutions were equilibrated overnight for more than 6 hours in an incubator maintained at 37 °C with 5% CO₂.

Female C57BL/6 J mice were intraperitoneally injected with 5 IU pregnant mare serum gonadotropin (PMSG). After 48 hours, 5 IU human chorionic gonadotropin (hCG) was administered, and females were subsequently paired with adult males for mating. Approximately 20 hours post-coitum, fertilized zygotes were collected, thoroughly washed, and transferred into KSOM medium containing different concentrations of ML323 for continued culture. DMSO was used as the control. Embryo development rates and immunofluorescence staining were recorded and compared across the treatment groups.

The effect of ML323 concentration was first tested using in vivo fertilized embryos, and consistent results were confirmed with IVF embryos. Due to the limited yield from in vivo fertilization, we proceeded with IVF-derived embryos for subsequent WGS. For WGS, embryos generated by IVF at the 2-Cell stage were used, with 100 embryos analyzed per group.

Public low-input ATAC-seq data of embryos

All ATAC-seq data used in this study were obtained from publicly available datasets reported by Wu et al.^43,74, covering human (six consecutive stages) and mouse (five consecutive stages). These stages comprise the zygote (only in humans), 2-Cell, 4-Cell, 8-Cell, and blastocyst, represented by the inner cell mass (ICM)-along with ICM-derived embryonic stem cells (ESCs). These publicly available datasets were used for downstream analyses without additional experimental manipulation.

Identification and filtering of eccDNA

Reference genomes for human and mouse were from Ensembl GRCh38 Release 107 and Ensembl GRCm38 Release 99. To ensure the quality of the eccDNA identified, we used the masked genomic DNA (interspersed repeats and low complexity regions are masked by replacing repeats with ‘N’s) and then also masked the blacklist (https://github.com/Boyle-Lab/Blacklist/). BWA-MEM (v0.7.18-r1243-dirty)⁷⁵, with the default setting, was used to map paired-ended reads to the hg38 and mm10 genome builds. Using Samblaster (v0.1.26)⁷⁶ to collect split reads and discordant reads. The complete process to identify eccDNA was accessible through the GitHub https://github.com/pk7zuva/Circle_finder. To improve detection accuracy, the obtained eccDNAs were filtered to remove those mapped to unplaced genomic scaffolds (e.g., chrGL*, chrJH*), mitochondrial genes, and regions longer than 50 Mb. Downstream analyses focused primarily on relatively short eccDNAs by applying strict filtering criteria: fragment sizes were limited to 100-1,000,000 bp, and each eccDNA was required to have at least 1 split read and a combined total of ≥4 split plus discordant reads (Supplementary Fig. 3C).

Quantification of eccDNA

We conducted a quantitative analysis of eccDNA by devising a counting approach. Specifically, we classified reads into split reads, discordant reads, and concordant reads based on the above pipeline. 1) For split reads, we used the BEDTools v2.30.0⁷⁷ suite pairToBed (do pairToBed -type both) to output reads that are aligned to both ends of eccDNA and were on the same direction chain of DNA. Then we filtered these reads to obtain preliminary eccDNA reads and ultimately calculated the number of split reads supporting each eccDNA. 2) For concordant reads, we employed “bedtools multicov” to calculate the number of concordant reads that overlap with eccDNA, and subsequently tallied the number of concordant reads that corroborate each eccDNA. 3) For discordant reads, both reads must be located entirely within eccDNA regions and have opposite orientations, with one read on the positive strand and the other on the negative strand. Following this criterion, we then counted the number of discordant reads that support each eccDNA.

Based on these three distinct types of sequencing reads, we defined the following variables: s_i, representing the number of split reads supporting eccDNA, i, in each sample; d_i, indicating the number of discordant reads supporting eccDNA, i, in each sample; and c_i, representing the number of concordant reads supporting eccDNA, i, in each sample. We also introduced r, which is defined as:

r = \frac{s_{i} + d_{i}}{s_{i} + d_{i} + c_{i}}

r represented the proportion of split and discordant reads among all three types of reads. Next, we used r to calculate the abundance of eccDNA, i:

E_{i} = r \times c_{i} + s_{i} + d_{i}

In this formula, we multiplied r with c_i to get an estimated contribution from concordant reads towards eccDNA support, and then added the actual numbers of split and discordant reads. Ultimately, E_i was used to calculate Transcripts per Million (TPM) values, which then represent the abundance of eccDNA. To assess the effect of eccDNA length on abundance estimation, we performed correlation analysis between eccDNA length and its abundance across all stages. The analysis revealed no significant or consistent correlation (Supplementary Fig. 4I), indicating that eccDNA size does not systematically bias quantification. This supports the robustness of our abundance estimation method.

Gene expression data processing

Public RNA-seq data were obtained from the Gene Expression Omnibus (GEO) GSE66582⁴³ and GSE44183⁴⁹. For all datasets, raw reads were trimmed using fastp (version 0.23.2) to remove the primer sequences as well as low-quality bases with the parameters -q 20 -u 30, followed by mapping to the ENSEMBL reference genome (corresponding to hg38 or mm10) by HISAT2 (version 2.2.1) software⁷⁸ with the default settings. The reads mapped to each gene were counted using featureCounts⁷⁹ (Version 2.0.1) under the parameters -p -a gtf -g gene_id, basing on the gene annotation files from Ensembl. For downstream analysis, we standardized the read counts by calculating TPM, and then the normalized counts were log10-transformed following the addition of a pseudocount of 1.

Analysis of public chromatin datasets

The published H3K4me3 ChIP-seq data for human 4-Cell embryos, 8-Cell embryos, and ICM, along with the H3K27ac ChIP-seq data of human 8-Cell embryos and ICM, were collected from GSE124718⁴⁶. The H3K4me3 ChIP-seq data for mouse early embryos per stage were pulled from GSE71434⁴². The H3K27ac ChIP-seq data for mouse 2-Cell embryos, 8-Cell embryos, and ESC were curated from GSE66390 and GSE72784, respectively^43,44. The Pol II datasets of mouse early development were from GSE135457⁴⁵. All reads were trimmed using fastp (version 0.23.2) with parameters -z 4 -q 20 -u 30. Following trimming, reads were aligned to the GRCh38 or GRCm38 reference genome using Bowtie 2 (version 2.3.5.1)⁸⁰, then sorted and indexed using SAMtools (version 1.15) with default parameters⁸¹. For downstream analysis, “bamCoverage” function in DeepTools (version 2.5.7)⁸² was utilized to generate normalized RPKM.bw files. Subsequently, DeepTools was employed for heatmap visualization using the ‘computeMatrix’ and ‘plotProfile’ functions.

Overlaps of eccDNAs from human preimplantation development with de novo mutations

We used “bedtools shuffle” to generate a random set of regions that match the lengths of eccDNA 1000 times, which was used as a control. Then, “bedtools intersect” was used to calculate the number of overlaps between these regions and the target.

Genomic feature acquisition

To determine chromosomal features per Mb relative to the recorded number of eccDNA per Mb, we extracted genomic features, such as chromosome length and the number of genes, from GRCh38 (Ensembl release 107) or GRCm38 (Ensembl release 99). Additionally, the number of Alu elements on each chromosome was obtained from Homer (http://homer.ucsd.edu/homer/).

Definition of gained, lost, stable, stage-specific, and consistent eccDNAs

To explore the dynamics of eccDNA, eccDNA presents at a specific stage but absent in the preceding stage were designated as “gain” eccDNAs for that stage. Conversely, eccDNAs detected at a given stage but absent in the subsequent stage were categorized as “loss” eccDNAs for that stage. Additionally, eccDNAs present at a stage and also observed in the preceding stage were classified as “stable” eccDNAs for that stage. All comparisons were performed using a stringent criterion of 100% overlap between eccDNA genomic coordinates. “Stage-specific eccDNAs” refer to those detected only at the current stage, with no coordinate overlap with eccDNAs from other stages. “Consistent eccDNAs” are defined as those present across all developmental stages, exhibiting complete coordinate overlap.

Genomic distribution and gene annotation of eccDNA

To delineate the genomic localization of eccDNA, eccDNA was designated according to the priority order (promoter > 5’ UTR > 3’ UTR > exon > intron > downstream > intergenic), facilitated by the ChIPseeker R package (v1.34.1)⁸³ in cases where a single eccDNA encompassed multiple genomic features. The identification of eccDNA-associated genes was carried out by determining the nearest genes to eccDNA using the ChIPseeker package.

Gene ontology (GO) enrichment analysis

GO enrichment analysis was conducted utilizing the ‘enrichGO’ functions of the ClusterProfiler R package (v4.6.2)⁸⁴. The P values were adjusted using the Benjamini-Hochberg correction to account for multiple comparisons, focusing on statistically significant enrichment.

Characterization of the eccDNA junction sites and motif-enrichment analysis

BEDTools v2.30.0 was employed to extract DNA sequences, comprising 20 base pairs upstream and downstream of every eccDNA junction site. Motif enrichment analyses were conducted using HOMER (version 4.11)⁸⁵ findMotifsGenome.pl algorithm with the parameters ‘-size given and -mask’, leading to known enrichment results and de novo enrichment results; the former were selected and utilized.

Fuzzy C-means clustering

eccDNA obtained from six consecutive embryonic stages was clustered into distinct groups using the Mfuzz R package (v2.58.0), employing the fuzzy c-means algorithm⁸⁶, allowing for a comprehensive understanding of temporal expression trends.

Correlation between eccDNA and RNA expression

In the case of dynamic changes of eccDNA and RNA expression, the input consisted of log10-transformed TPM values derived from each stage. A correlation analysis was performed between eccDNA and RNA expression levels in shared genes, utilizing Spearman’s correlation coefficient with a cutoff of P value < 0.01, and a correlation coefficient of 0.7 for positive (or negative −0.7) were employed.

Structural variant (SV) calling in preimplantation embryo MDA samples

SVs in genomes from various developmental stages were identified using the lumpyexpress function of LUMPY (v0.2.13)⁶⁵ with default parameter settings. We then quantified the major SV types, including inversions, translocations, deletions, and duplications. To assess their association with eccDNAs, overlap analysis was performed using a 90% reciprocal overlap threshold as the criterion for counting shared regions.

Inference of copy number variation

We employed Ginkgo⁶⁶ to analyze genome-wide copy number values (CNVs) on MDA-based WGS datasets using 100-kb genomic bins. Subsequently, we performed statistical analysis on the CNVs of each region across different developmental stages.

AmpliconArchitect (AA) analysis

AmpliconSuite (version 1.4.0)⁶⁷ was used to process MDA-based WGS datasets, including the reconstruction of focal amplifications and potential circular amplicons. BAM files were first processed with PrepareAA, and seed intervals were defined based on regions of high copy number, as inferred from CNV profiles. AA was then run with default parameters to reconstruct the architecture of amplified regions and to determine the presence of cyclic structures.

Statistics & reproducibility

No statistical method was used to predetermine sample size. Sample sizes were determined based on embryo availability and previous published studies. No data were excluded from the analyses. The experiments were not randomized. The investigators were not blinded to allocation during experiments and outcome assessment. Statistical analyses were performed using R (version 4.2.1). Principal component analysis was conducted using the prcomp function in R. Data visualization was performed using the ggplot2 package (version 3.4.3).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information^{(3.9MB, pdf)}

41467_2026_71227_MOESM2_ESM.pdf^{(361.6KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1^{(894.2KB, xlsx)}

Supplementary Data 2^{(12.4MB, xlsx)}

Supplementary Data 3^{(581.8KB, xlsx)}

Supplementary Data 4^{(11KB, xlsx)}

Supplementary Data 5^{(20MB, xlsx)}

Supplementary Data 6^{(1.3MB, xlsx)}

Supplementary Data 7^{(1.2MB, xlsx)}

Supplementary Data 8^{(8.3MB, xlsx)}

Supplementary Data 9^{(617.3KB, xlsx)}

Supplementary Data 10^{(175KB, xlsx)}

Supplementary Data 11^{(15.4KB, xlsx)}

Supplementary Data 12^{(12.1KB, xlsx)}

Reporting Summary^{(103.7KB, pdf)}

Transparent Peer Review file^{(1.6MB, pdf)}

Source data

Source Data^{(46.2MB, xlsx)}

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant Nos. 82288102 to JQ, 32470894 to QL, 32170493 to XLZ, 32470835 to FBM, 32400703 to LC), the National Key Research and Development Program of China (Grant No. 2025YFC2708100 to QL), the Beijing Natural Science Foundation (Grant Nos. L248056 to FBM, 7242169 to XLZ, 7244435 to LC), the fellowship of China National Postdoctoral Program for Innovative Talents (Grant No. BX20230031 to LC), the Key Clinical Projects of Peking University Third Hospital (Grant No. BYSYZD2024025 to XLZ), and the State Key Laboratory of Female Fertility Promotion, Center for Reproductive Medicine, Department of Obstetrics and Gynecology, Peking University Third Hospital (Grant Nos. BYSYSZKF2024002 to XLZ, BYSYSZKF2025008 to FBM).

Author contributions

F.B.M., X.L.Z., Q.L. and J.Q. conceived and designed the study. L.W. collected the data, did the analysis, and wrote the paper. N.W. and L.C. performed the experiments. TW did the analysis of eccDNA overlapping with de novo mutations. LSS tested the AA pipeline. ZPZ tested the Ginkgo pipeline. F.B.M., X.L.Z., Q.L., J.Q. and X.X. performed a critical reading of the manuscript and modified it. All authors improved the manuscript and approved the submission.

Peer review

Peer review information

Nature Communications thanks Joan Barau and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

Raw sequencing datasets of MDA and scRNA-seq reported in this paper have been deposited in the Genome Sequence Archive (GSA) in the National Genomics Data Center, China National Center for Bioinformation, Chinese Academy of Sciences, under accession number CRA019281. Publicly available ATAC-seq and other datasets used in this study were obtained from the Gene Expression Omnibus (GEO) under the following accession codes: GSE66582, GSE44183, GSE124718, GSE71434, GSE66390, GSE72784, GSE135457. Source data are provided with this paper.

Code availability

All scripts used for eccDNA quantification in this study are publicly available at GitHub: https://github.com/duck-rong/eccDNA_quantification_scripts. No software was developed beyond these scripts; all analyses were performed using these scripts together with publicly available tools as described in the Methods section.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Ling Wei, Ning Wu, Lu Chen.

Contributor Information

Jie Qiao, Email: jie.qiao@263.net.

Qiang Liu, Email: liuqiangtaian2008@163.com.

Xiaolu Zhao, Email: xiaolu_zhao@163.com.

Fengbiao Mao, Email: fengbiaomao@bjmu.edu.cn.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-71227-z.

References

1.Yang, L. et al. Extrachromosomal circular DNA: biogenesis, structure, functions and diseases. Signal Transduct. Target Ther.7, 342 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Zhao, X. et al. CircleBase: an integrated resource and analysis platform for human eccDNAs. Nucleic Acids Res.50, D72–D82 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Wei, L. et al. CircleBase V2: an eccDNA annotation platform across cancers and species. Nucleic Acids Res.54, D66–D77 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Wu, N. et al. Innovative insights into extrachromosomal circular DNAs in gynecologic tumors and reproduction. Protein Cell15, 6–20 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Shibata, Y. et al. Extrachromosomal microDNAs and chromosomal microdeletions in normal tissues. Science336, 82–86 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Moller, H. D., Parsons, L., Jorgensen, T. S., Botstein, D. & Regenberg, B. Extrachromosomal circular DNA is common in yeast. Proc. Natl. Acad. Sci. USA112, E3114–3122 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Dillon, L. W. et al. Production of extrachromosomal MicroDNAs IS Linked To Mismatch Repair Pathways And Transcriptional Activity. Cell Rep.11, 1749–1759 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Moller, H. D. et al. Circular DNA elements of chromosomal origin are common in healthy human somatic tissue. Nat. Commun.9, 1069 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Wang, Y. et al. eccDNAs are apoptotic products with high innate immunostimulatory activity. Nature599, 308–314 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Moller, H. D., Ramos-Madrigal, J., Prada-Luengo, I., Gilbert, M. T. P. & Regenberg, B. Near-random distribution of chromosome-derived circular DNA in the Condensed genome of pigeons and the larger, more repeat-rich human genome. Genome Biol. Evol.12, 3762–3777 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Henriksen, R. A. et al. Circular DNA in the human germline and its association with recombination. Mol. Cell82, 209–217 e207 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Paulsen, T. et al. MicroDNA levels are dependent on MMEJ, repressed by c-NHEJ pathway, and stimulated by DNA damage. Nucleic Acids Res.49, 11787–11799 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Yang, F. et al. Retrotransposons hijack alt-EJ for DNA replication and eccDNA biogenesis. Nature620, 218–225 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Wu, S. et al. Circular ecDNA promotes accessible chromatin and high oncogene expression. Nature575, 699–703 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Noer, J. B., Horsdal, O. K., Xiang, X., Luo, Y. & Regenberg, B. Extrachromosomal circular DNA in cancer: history, current knowledge, and methods. Trends Genet.38, 766–781 (2022). [DOI] [PubMed] [Google Scholar]
16.Lv, W. et al. Extrachromosomal circular DNA orchestrates genome heterogeneity in urothelial bladder carcinoma. Theranostics14, 5102–5122 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Koche, R. P. et al. Extrachromosomal circular DNA drives oncogenic genome remodeling in neuroblastoma. Nat. Genet.52, 29–34 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Zhu, Y. et al. Oncogenic extrachromosomal DNA functions as mobile enhancers to globally amplify chromosomal transcription. Cancer Cell39, 694–707 e697 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Nathanson, D. A. et al. Targeted therapy resistance mediated by dynamic regulation of extrachromosomal mutant EGFR DNA. Science343, 72–76 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Lv. W, et al. Spatial-temporal diversity of extrachromosomal DNA shapes urothelial carcinoma evolution and tumor-immune microenvironment. Cancer Discov.10.1158/2159-8290.CD-24-1532 (2025). [DOI] [PubMed]
21.Ling, X. et al. Small extrachromosomal circular DNA (eccDNA): major functions in evolution and cancer. Mol. Cancer20, 113 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Shi, B., Yang, P., Qiao, H., Yu, D. & Zhang, S. Extrachromosomal circular DNA drives dynamic genome plasticity: emerging roles in disease progression and clinical potential. Theranostics15, 6387–6411 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Zhou, M. et al. Plasma extrachromosomal circular DNA as a potential diagnostic biomarker for nodular thyroid disease. Clin. Transl. Med.14, e1740 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Hu, J. et al. Microhomology-mediated circular DNA formation from oligonucleosomal fragments during spermatogenesis. Elife12, RP87115 (2023). [DOI] [PMC free article] [PubMed]
25.Schulz, K. N. & Harrison, M. M. Mechanisms regulating zygotic genome activation. Nat. Rev. Genet.20, 221–234 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Vastenhouw, N.L., Cao, W.X., Lipshitz, H.D. The maternal-to-zygotic transition revisited. Development146, dev161471 (2019). [DOI] [PubMed]
27.Abe, K. I. et al. Minor zygotic gene activation is essential for mouse preimplantation development. Proc. Natl. Acad. Sci. USA115, E6780–E6788 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Messerschmidt, D. M., Knowles, B. B. & Solter, D. DNA methylation dynamics during epigenetic reprogramming in the germline and preimplantation embryos. Genes Dev.28, 812–828 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Reik, W., Dean, W. & Walter, J. Epigenetic reprogramming in mammalian development. Science293, 1089–1093 (2001). [DOI] [PubMed] [Google Scholar]
30.Nakatani, T. et al. Emergence of replication timing during early mammalian development. Nature625, 401–409 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Kumar, P. et al. ATAC-seq identifies thousands of extrachromosomal circular DNA in cancer and cell lines. Sci. Adv.6, eaba2489 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Fan, X. et al. SMOOTH-seq: single-cell genome sequencing of human cells on a third-generation sequencing platform. Genome Biol.22, 195 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Flach, G., Johnson, M. H., Braude, P. R., Taylor, R. A. & Bolton, V. N. The transition from maternal to embryonic control in the 2-cell mouse embryo. EMBO J.1, 681–686 (1982). [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Chen, Z. & Zhang, Y. Loss of DUX causes minor defects in zygotic genome activation and is compatible with mouse development. Nat. Genet.51, 947–951 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Kang, X. et al. Extrachromosomal DNA replication and maintenance couple with DNA damage pathway in tumors. Cell188, 3405–3421 e3427 (2025). [DOI] [PubMed] [Google Scholar]
36.Paulsen, T., Kumar, P., Koseoglu, M. M. & Dutta, A. Discoveries of extrachromosomal circles of DNA in normal and tumor cells. Trends Genet34, 270–278 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Ji, S. et al. OBOX regulates mouse zygotic genome activation and early development. Nature620, 1047–1053 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
38.He, A. et al. Dynamic GATA4 enhancers shape the chromatin landscape central to heart development and disease. Nat. Commun.5, 4907 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Cirillo, L. A., Lin, F. R., Cuesta, I., Friedman, D., Jarnik, M. & Zaret, K. S. Opening of compacted chromatin by early developmental transcription factors HNF3 (FoxA) and GATA-4. Mol. Cell9, 279–289 (2002). [DOI] [PubMed] [Google Scholar]
40.Yan, J., Xu, L., Crawford, G., Wang, Z. & Burgess, S. M. The forkhead transcription factor FoxI1 remains bound to condensed mitotic chromosomes and stably remodels chromatin structure. Mol. Cell Biol.26, 155–168 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Lalmansingh, A. S., Karmakar, S., Jin, Y. & Nagaich, A. K. Multiple modes of chromatin remodeling by Forkhead box proteins. Biochim Biophys. Acta1819, 707–715 (2012). [DOI] [PubMed] [Google Scholar]
42.Zhang, B. et al. Allelic reprogramming of the histone modification H3K4me3 in early mammalian development. Nature537, 553–557 (2016). [DOI] [PubMed] [Google Scholar]
43.Wu, J. et al. The landscape of accessible chromatin in mammalian preimplantation embryos. Nature534, 652–657 (2016). [DOI] [PubMed] [Google Scholar]
44.Dahl, J. A. et al. Broad histone H3K4me3 domains in mouse oocytes modulate maternal-to-zygotic transition. Nature537, 548–552 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Liu, B. et al. The landscape of RNA Pol II binding reveals a stepwise transition during ZGA. Nature587, 139–144 (2020). [DOI] [PubMed] [Google Scholar]
46.Xia, W. et al. Resetting histone modifications during human parental-to-zygotic transition. Science365, 353–360 (2019). [DOI] [PubMed] [Google Scholar]
47.Lorch, Y., Maier-Davis, B. & Kornberg, R. D. Role of DNA sequence in chromatin remodeling and the formation of nucleosome-free regions. Genes Dev.28, 2492–2497 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Yang, Y. et al. Transcription-replication conflicts in primordial germ cells necessitate the Fanconi anemia pathway to safeguard genome stability. Proc. Natl. Acad. Sci. USA119, e2203208119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Xue, Z. et al. Genetic programs in human and mouse early embryos revealed by single-cell RNA sequencing. Nature500, 593–597 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Futschik, M. E. & Carlisle, B. Noise-robust soft clustering of gene expression time-course data. J. Bioinform. Comput Biol.3, 965–988 (2005). [DOI] [PubMed] [Google Scholar]
51.Xu, R. et al. Stage-specific H3K9me3 occupancy ensures retrotransposon silencing in human pre-implantation embryos. Cell Stem Cell29, 1051–1066 e1058 (2022). [DOI] [PubMed] [Google Scholar]
52.Paulsen, T., Shibata, Y., Kumar, P., Dillon, L. & Dutta, A. Small extrachromosomal circular DNAs, microDNA, produce short regulatory RNAs that suppress gene expression independent of canonical promoters. Nucleic Acids Res.47, 4586–4596 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Gan, W. et al. R-loop-mediated genomic instability is caused by impairment of replication fork progression. Genes Dev.25, 2041–2056 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Huertas, P. & Aguilera, A. Cotranscriptionally formed DNA:RNA hybrids mediate transcription elongation impairment and transcription-associated recombination. Mol. Cell12, 711–721 (2003). [DOI] [PubMed] [Google Scholar]
55.Palmerola, K. L. et al. Replication stress impairs chromosome segregation and preimplantation development in human embryos. Cell185, 2988–3007 e2920 (2022). [DOI] [PubMed] [Google Scholar]
56.Chen S, et al. The urinary eccDNA landscape in prostate cancer reveals associations with genome instability and vital roles in cancer progression. J. Adv. Res.77, 637–652 (2025). [DOI] [PMC free article] [PubMed]
57.Tang, L. et al. Circular single-stranded DNA as switchable vector for gene expression in mammalian cells. Nat. Commun.14, 6665 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Joubert, P. M. & Krasileva, K. V. The extrachromosomal circular DNAs of the rice blast pathogen Magnaporthe oryzae contain a wide variety of LTR retrotransposons, genes, and effectors. BMC Biol.20, 260 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Tei, C. et al. Comparative analysis of multiple DNA double-strand break repair pathways in CRISPR-mediated endogenous tagging. Commun. Biol.8, 749 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Sfeir, A., Tijsterman, M. & McVey, M. Microhomology-MEdiated End-joining Chronicles: Tracing The Evolutionary Footprints Of Genome Protection. Annu. Rev. Cell Dev. Biol.40, 195–218 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Chazaud, C., Yamanaka, Y., Pawson, T. & Rossant, J. Early lineage segregation between epiblast and primitive endoderm in mouse blastocysts through the Grb2-MAPK pathway. Dev. Cell10, 615–624 (2006). [DOI] [PubMed] [Google Scholar]
62.Werling, D. M. et al. An analytical framework for whole-genome sequence association studies and its implications for autism spectrum disorder. Nat. Genet.50, 727–736 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Iossifov, I. et al. The contribution of de novo coding mutations to autism spectrum disorder. Nature515, 216–221 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Zhao, G. et al. Gene4Denovo: an integrated database and analytic platform for de novo mutations in humans. Nucleic Acids Res.48, D913–D926 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Layer, R. M., Chiang, C., Quinlan, A. R. & Hall, I. M. LUMPY: a probabilistic framework for structural variant discovery. Genome Biol.15, R84 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Garvin, T. et al. Interactive analysis and assessment of single-cell copy-number variations. Nat. Methods12, 1058–1060 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Deshpande, V. et al. Exploring the landscape of focal amplifications in cancer using AmpliconArchitect. Nat. Commun.10, 392 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Picelli, S., Faridani, O. R., Bjorklund, A. K., Winberg, G., Sagasser, S. & Sandberg, R. Full-length RNA-seq from single cells using Smart-seq2. Nat. Protoc.9, 171–181 (2014). [DOI] [PubMed] [Google Scholar]
69.Chen, L. et al. NAT10-mediated mRNA N(4)-acetylation is essential for the translational regulation during oocyte meiotic maturation in mice. Sci. Adv.11, eadp5163 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Oka, Y., Bekker-Jensen, S. & Mailand, N. Ubiquitin-like protein UBL5 promotes the functional integrity of the Fanconi anemia pathway. EMBO J.34, 1385–1398 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Dexheimer, T.S. et al. Discovery of ML323 as a Novel Inhibitor of the USP1/UAF1 Deubiquitinase Complex. In: Probe Reports from the NIH Molecular Libraries Program) (2010).
72.Liang, Q. et al. A selective USP1-UAF1 inhibitor links deubiquitination to DNA damage responses. Nat. Chem. Biol.10, 298–304 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Yu, Z. et al. USP1-UAF1 deubiquitinase complex stabilizes TBK1 and enhances antiviral responses. J. Exp. Med.214, 3553–3563 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
74.Wu, J. et al. Chromatin analysis in human early development reveals epigenetic transition during ZGA. Nature557, 256–260 (2018). [DOI] [PubMed] [Google Scholar]
75.Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics25, 1754–1760 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
76.Faust, G. G. & Hall, I. M. SAMBLASTER: fast duplicate marking and structural variant read extraction. Bioinformatics30, 2503–2505 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
77.Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics26, 841–842 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
78.Pertea, M., Kim, D., Pertea, G. M., Leek, J. T. & Salzberg, S. L. Transcript-level expression analysis of RNA-seq experiments with HISAT, StringTie and Ballgown. Nat. Protoc.11, 1650–1667 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
79.Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics30, 923–930 (2014). [DOI] [PubMed] [Google Scholar]
80.Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods9, 357–359 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
81.Danecek, P, et al. Twelve years of SAMtools and BCFtools. Gigascience10, giab008 (2021). [DOI] [PMC free article] [PubMed]
82.Ramirez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res.44, W160–165 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
83.Yu, G., Wang, L. G. & He, Q. Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics31, 2382–2383 (2015). [DOI] [PubMed] [Google Scholar]
84.Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS16, 284–287 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Heinz, S. et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell38, 576–589 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
86.Kumar, L. M EF. Mfuzz: a software package for soft clustering of microarray data. Bioinformation2, 5–7 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(3.9MB, pdf)}

41467_2026_71227_MOESM2_ESM.pdf^{(361.6KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1^{(894.2KB, xlsx)}

Supplementary Data 2^{(12.4MB, xlsx)}

Supplementary Data 3^{(581.8KB, xlsx)}

Supplementary Data 4^{(11KB, xlsx)}

Supplementary Data 5^{(20MB, xlsx)}

Supplementary Data 6^{(1.3MB, xlsx)}

Supplementary Data 7^{(1.2MB, xlsx)}

Supplementary Data 8^{(8.3MB, xlsx)}

Supplementary Data 9^{(617.3KB, xlsx)}

Supplementary Data 10^{(175KB, xlsx)}

Supplementary Data 11^{(15.4KB, xlsx)}

Supplementary Data 12^{(12.1KB, xlsx)}

Reporting Summary^{(103.7KB, pdf)}

Transparent Peer Review file^{(1.6MB, pdf)}

Source Data^{(46.2MB, xlsx)}

Data Availability Statement

[CR1] 1.Yang, L. et al. Extrachromosomal circular DNA: biogenesis, structure, functions and diseases. Signal Transduct. Target Ther.7, 342 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Zhao, X. et al. CircleBase: an integrated resource and analysis platform for human eccDNAs. Nucleic Acids Res.50, D72–D82 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Wei, L. et al. CircleBase V2: an eccDNA annotation platform across cancers and species. Nucleic Acids Res.54, D66–D77 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Wu, N. et al. Innovative insights into extrachromosomal circular DNAs in gynecologic tumors and reproduction. Protein Cell15, 6–20 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] 5.Shibata, Y. et al. Extrachromosomal microDNAs and chromosomal microdeletions in normal tissues. Science336, 82–86 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Moller, H. D., Parsons, L., Jorgensen, T. S., Botstein, D. & Regenberg, B. Extrachromosomal circular DNA is common in yeast. Proc. Natl. Acad. Sci. USA112, E3114–3122 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR7] 7.Dillon, L. W. et al. Production of extrachromosomal MicroDNAs IS Linked To Mismatch Repair Pathways And Transcriptional Activity. Cell Rep.11, 1749–1759 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR8] 8.Moller, H. D. et al. Circular DNA elements of chromosomal origin are common in healthy human somatic tissue. Nat. Commun.9, 1069 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Wang, Y. et al. eccDNAs are apoptotic products with high innate immunostimulatory activity. Nature599, 308–314 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Moller, H. D., Ramos-Madrigal, J., Prada-Luengo, I., Gilbert, M. T. P. & Regenberg, B. Near-random distribution of chromosome-derived circular DNA in the Condensed genome of pigeons and the larger, more repeat-rich human genome. Genome Biol. Evol.12, 3762–3777 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.Henriksen, R. A. et al. Circular DNA in the human germline and its association with recombination. Mol. Cell82, 209–217 e207 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Paulsen, T. et al. MicroDNA levels are dependent on MMEJ, repressed by c-NHEJ pathway, and stimulated by DNA damage. Nucleic Acids Res.49, 11787–11799 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.Yang, F. et al. Retrotransposons hijack alt-EJ for DNA replication and eccDNA biogenesis. Nature620, 218–225 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Wu, S. et al. Circular ecDNA promotes accessible chromatin and high oncogene expression. Nature575, 699–703 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Noer, J. B., Horsdal, O. K., Xiang, X., Luo, Y. & Regenberg, B. Extrachromosomal circular DNA in cancer: history, current knowledge, and methods. Trends Genet.38, 766–781 (2022). [DOI] [PubMed] [Google Scholar]

[CR16] 16.Lv, W. et al. Extrachromosomal circular DNA orchestrates genome heterogeneity in urothelial bladder carcinoma. Theranostics14, 5102–5122 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Koche, R. P. et al. Extrachromosomal circular DNA drives oncogenic genome remodeling in neuroblastoma. Nat. Genet.52, 29–34 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Zhu, Y. et al. Oncogenic extrachromosomal DNA functions as mobile enhancers to globally amplify chromosomal transcription. Cancer Cell39, 694–707 e697 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Nathanson, D. A. et al. Targeted therapy resistance mediated by dynamic regulation of extrachromosomal mutant EGFR DNA. Science343, 72–76 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Lv. W, et al. Spatial-temporal diversity of extrachromosomal DNA shapes urothelial carcinoma evolution and tumor-immune microenvironment. Cancer Discov.10.1158/2159-8290.CD-24-1532 (2025). [DOI] [PubMed]

[CR21] 21.Ling, X. et al. Small extrachromosomal circular DNA (eccDNA): major functions in evolution and cancer. Mol. Cancer20, 113 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Shi, B., Yang, P., Qiao, H., Yu, D. & Zhang, S. Extrachromosomal circular DNA drives dynamic genome plasticity: emerging roles in disease progression and clinical potential. Theranostics15, 6387–6411 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Zhou, M. et al. Plasma extrachromosomal circular DNA as a potential diagnostic biomarker for nodular thyroid disease. Clin. Transl. Med.14, e1740 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Hu, J. et al. Microhomology-mediated circular DNA formation from oligonucleosomal fragments during spermatogenesis. Elife12, RP87115 (2023). [DOI] [PMC free article] [PubMed]

[CR25] 25.Schulz, K. N. & Harrison, M. M. Mechanisms regulating zygotic genome activation. Nat. Rev. Genet.20, 221–234 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Vastenhouw, N.L., Cao, W.X., Lipshitz, H.D. The maternal-to-zygotic transition revisited. Development146, dev161471 (2019). [DOI] [PubMed]

[CR27] 27.Abe, K. I. et al. Minor zygotic gene activation is essential for mouse preimplantation development. Proc. Natl. Acad. Sci. USA115, E6780–E6788 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Messerschmidt, D. M., Knowles, B. B. & Solter, D. DNA methylation dynamics during epigenetic reprogramming in the germline and preimplantation embryos. Genes Dev.28, 812–828 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Reik, W., Dean, W. & Walter, J. Epigenetic reprogramming in mammalian development. Science293, 1089–1093 (2001). [DOI] [PubMed] [Google Scholar]

[CR30] 30.Nakatani, T. et al. Emergence of replication timing during early mammalian development. Nature625, 401–409 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Kumar, P. et al. ATAC-seq identifies thousands of extrachromosomal circular DNA in cancer and cell lines. Sci. Adv.6, eaba2489 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Fan, X. et al. SMOOTH-seq: single-cell genome sequencing of human cells on a third-generation sequencing platform. Genome Biol.22, 195 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Flach, G., Johnson, M. H., Braude, P. R., Taylor, R. A. & Bolton, V. N. The transition from maternal to embryonic control in the 2-cell mouse embryo. EMBO J.1, 681–686 (1982). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Chen, Z. & Zhang, Y. Loss of DUX causes minor defects in zygotic genome activation and is compatible with mouse development. Nat. Genet.51, 947–951 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR35] 35.Kang, X. et al. Extrachromosomal DNA replication and maintenance couple with DNA damage pathway in tumors. Cell188, 3405–3421 e3427 (2025). [DOI] [PubMed] [Google Scholar]

[CR36] 36.Paulsen, T., Kumar, P., Koseoglu, M. M. & Dutta, A. Discoveries of extrachromosomal circles of DNA in normal and tumor cells. Trends Genet34, 270–278 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Ji, S. et al. OBOX regulates mouse zygotic genome activation and early development. Nature620, 1047–1053 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.He, A. et al. Dynamic GATA4 enhancers shape the chromatin landscape central to heart development and disease. Nat. Commun.5, 4907 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Cirillo, L. A., Lin, F. R., Cuesta, I., Friedman, D., Jarnik, M. & Zaret, K. S. Opening of compacted chromatin by early developmental transcription factors HNF3 (FoxA) and GATA-4. Mol. Cell9, 279–289 (2002). [DOI] [PubMed] [Google Scholar]

[CR40] 40.Yan, J., Xu, L., Crawford, G., Wang, Z. & Burgess, S. M. The forkhead transcription factor FoxI1 remains bound to condensed mitotic chromosomes and stably remodels chromatin structure. Mol. Cell Biol.26, 155–168 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR41] 41.Lalmansingh, A. S., Karmakar, S., Jin, Y. & Nagaich, A. K. Multiple modes of chromatin remodeling by Forkhead box proteins. Biochim Biophys. Acta1819, 707–715 (2012). [DOI] [PubMed] [Google Scholar]

[CR42] 42.Zhang, B. et al. Allelic reprogramming of the histone modification H3K4me3 in early mammalian development. Nature537, 553–557 (2016). [DOI] [PubMed] [Google Scholar]

[CR43] 43.Wu, J. et al. The landscape of accessible chromatin in mammalian preimplantation embryos. Nature534, 652–657 (2016). [DOI] [PubMed] [Google Scholar]

[CR44] 44.Dahl, J. A. et al. Broad histone H3K4me3 domains in mouse oocytes modulate maternal-to-zygotic transition. Nature537, 548–552 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR45] 45.Liu, B. et al. The landscape of RNA Pol II binding reveals a stepwise transition during ZGA. Nature587, 139–144 (2020). [DOI] [PubMed] [Google Scholar]

[CR46] 46.Xia, W. et al. Resetting histone modifications during human parental-to-zygotic transition. Science365, 353–360 (2019). [DOI] [PubMed] [Google Scholar]

[CR47] 47.Lorch, Y., Maier-Davis, B. & Kornberg, R. D. Role of DNA sequence in chromatin remodeling and the formation of nucleosome-free regions. Genes Dev.28, 2492–2497 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Yang, Y. et al. Transcription-replication conflicts in primordial germ cells necessitate the Fanconi anemia pathway to safeguard genome stability. Proc. Natl. Acad. Sci. USA119, e2203208119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Xue, Z. et al. Genetic programs in human and mouse early embryos revealed by single-cell RNA sequencing. Nature500, 593–597 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] 50.Futschik, M. E. & Carlisle, B. Noise-robust soft clustering of gene expression time-course data. J. Bioinform. Comput Biol.3, 965–988 (2005). [DOI] [PubMed] [Google Scholar]

[CR51] 51.Xu, R. et al. Stage-specific H3K9me3 occupancy ensures retrotransposon silencing in human pre-implantation embryos. Cell Stem Cell29, 1051–1066 e1058 (2022). [DOI] [PubMed] [Google Scholar]

[CR52] 52.Paulsen, T., Shibata, Y., Kumar, P., Dillon, L. & Dutta, A. Small extrachromosomal circular DNAs, microDNA, produce short regulatory RNAs that suppress gene expression independent of canonical promoters. Nucleic Acids Res.47, 4586–4596 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.Gan, W. et al. R-loop-mediated genomic instability is caused by impairment of replication fork progression. Genes Dev.25, 2041–2056 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Huertas, P. & Aguilera, A. Cotranscriptionally formed DNA:RNA hybrids mediate transcription elongation impairment and transcription-associated recombination. Mol. Cell12, 711–721 (2003). [DOI] [PubMed] [Google Scholar]

[CR55] 55.Palmerola, K. L. et al. Replication stress impairs chromosome segregation and preimplantation development in human embryos. Cell185, 2988–3007 e2920 (2022). [DOI] [PubMed] [Google Scholar]

[CR56] 56.Chen S, et al. The urinary eccDNA landscape in prostate cancer reveals associations with genome instability and vital roles in cancer progression. J. Adv. Res.77, 637–652 (2025). [DOI] [PMC free article] [PubMed]

[CR57] 57.Tang, L. et al. Circular single-stranded DNA as switchable vector for gene expression in mammalian cells. Nat. Commun.14, 6665 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Joubert, P. M. & Krasileva, K. V. The extrachromosomal circular DNAs of the rice blast pathogen Magnaporthe oryzae contain a wide variety of LTR retrotransposons, genes, and effectors. BMC Biol.20, 260 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR59] 59.Tei, C. et al. Comparative analysis of multiple DNA double-strand break repair pathways in CRISPR-mediated endogenous tagging. Commun. Biol.8, 749 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Sfeir, A., Tijsterman, M. & McVey, M. Microhomology-MEdiated End-joining Chronicles: Tracing The Evolutionary Footprints Of Genome Protection. Annu. Rev. Cell Dev. Biol.40, 195–218 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR61] 61.Chazaud, C., Yamanaka, Y., Pawson, T. & Rossant, J. Early lineage segregation between epiblast and primitive endoderm in mouse blastocysts through the Grb2-MAPK pathway. Dev. Cell10, 615–624 (2006). [DOI] [PubMed] [Google Scholar]

[CR62] 62.Werling, D. M. et al. An analytical framework for whole-genome sequence association studies and its implications for autism spectrum disorder. Nat. Genet.50, 727–736 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR63] 63.Iossifov, I. et al. The contribution of de novo coding mutations to autism spectrum disorder. Nature515, 216–221 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR64] 64.Zhao, G. et al. Gene4Denovo: an integrated database and analytic platform for de novo mutations in humans. Nucleic Acids Res.48, D913–D926 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR65] 65.Layer, R. M., Chiang, C., Quinlan, A. R. & Hall, I. M. LUMPY: a probabilistic framework for structural variant discovery. Genome Biol.15, R84 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR66] 66.Garvin, T. et al. Interactive analysis and assessment of single-cell copy-number variations. Nat. Methods12, 1058–1060 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR67] 67.Deshpande, V. et al. Exploring the landscape of focal amplifications in cancer using AmpliconArchitect. Nat. Commun.10, 392 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR68] 68.Picelli, S., Faridani, O. R., Bjorklund, A. K., Winberg, G., Sagasser, S. & Sandberg, R. Full-length RNA-seq from single cells using Smart-seq2. Nat. Protoc.9, 171–181 (2014). [DOI] [PubMed] [Google Scholar]

[CR69] 69.Chen, L. et al. NAT10-mediated mRNA N(4)-acetylation is essential for the translational regulation during oocyte meiotic maturation in mice. Sci. Adv.11, eadp5163 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR70] 70.Oka, Y., Bekker-Jensen, S. & Mailand, N. Ubiquitin-like protein UBL5 promotes the functional integrity of the Fanconi anemia pathway. EMBO J.34, 1385–1398 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR71] 71.Dexheimer, T.S. et al. Discovery of ML323 as a Novel Inhibitor of the USP1/UAF1 Deubiquitinase Complex. In: Probe Reports from the NIH Molecular Libraries Program) (2010).

[CR72] 72.Liang, Q. et al. A selective USP1-UAF1 inhibitor links deubiquitination to DNA damage responses. Nat. Chem. Biol.10, 298–304 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR73] 73.Yu, Z. et al. USP1-UAF1 deubiquitinase complex stabilizes TBK1 and enhances antiviral responses. J. Exp. Med.214, 3553–3563 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR74] 74.Wu, J. et al. Chromatin analysis in human early development reveals epigenetic transition during ZGA. Nature557, 256–260 (2018). [DOI] [PubMed] [Google Scholar]

[CR75] 75.Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics25, 1754–1760 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR76] 76.Faust, G. G. & Hall, I. M. SAMBLASTER: fast duplicate marking and structural variant read extraction. Bioinformatics30, 2503–2505 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR77] 77.Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics26, 841–842 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR78] 78.Pertea, M., Kim, D., Pertea, G. M., Leek, J. T. & Salzberg, S. L. Transcript-level expression analysis of RNA-seq experiments with HISAT, StringTie and Ballgown. Nat. Protoc.11, 1650–1667 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR79] 79.Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics30, 923–930 (2014). [DOI] [PubMed] [Google Scholar]

[CR80] 80.Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods9, 357–359 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR81] 81.Danecek, P, et al. Twelve years of SAMtools and BCFtools. Gigascience10, giab008 (2021). [DOI] [PMC free article] [PubMed]

[CR82] 82.Ramirez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res.44, W160–165 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR83] 83.Yu, G., Wang, L. G. & He, Q. Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics31, 2382–2383 (2015). [DOI] [PubMed] [Google Scholar]

[CR84] 84.Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS16, 284–287 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR85] 85.Heinz, S. et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell38, 576–589 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR86] 86.Kumar, L. M EF. Mfuzz: a software package for soft clustering of microarray data. Bioinformation2, 5–7 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

The landscape and regulatory potential of eccDNAs in mammalian preimplantation embryos

Ling Wei

Ning Wu

Lu Chen

Tao Wang

Zhipeng Zhu

Leisheng Shi

Xi Xiang

Jie Qiao

Qiang Liu

Xiaolu Zhao

Fengbiao Mao

Abstract

Introduction

Results

The widespread presence of eccDNA in mouse preimplantation embryos

Fig. 1. Characteristics of eccDNAs during mouse preimplantation embryos.

Fig. 5. Distinct clusters of eccDNA reveal key regulatory events during preimplantation development.

Characteristics of eccDNA in mouse preimplantation embryos

Formation mechanisms of eccDNA during mouse preimplantation stages

Fig. 2. Molecular signatures suggestive of eccDNA biogenesis mechanisms.

Suppression of transcription led to a marked decrease in both the quantity and abundance of eccDNA

Fig. 3. The effect of transcriptional inhibition on eccDNA biogenesis.

Increased TRCs lead to elevated levels of eccDNA, both in terms of its quantity and abundance

Fig. 4. Fanconi anemia pathway perturbation increases eccDNA quantity and abundance.

Potential roles of eccDNA during preimplantation development

Fig. 6. Potential roles of eccDNA in regulating gene expression during preimplantation development.

Characteristics of eccDNAs during preimplantation development are conserved between human and mouse

Fig. 7. eccDNA characteristics are conserved during human preimplantation development.

Discussion

Methods

Animals and ethics

Oocyte collection, in vitro fertilization and embryo culture

Multiple displacement amplification (MDA), library construction, and sequencing

Candidate eccDNA validation by PCR

eccDNA synthesis via Ligation-Assisted Minicircle Amplification (LAMA)

Transcription inhibition assay

scRNA-seq library preparation and sequencing

eccDNA microinjection and RT-qPCR in NIH 3T3 cells

eccDNA transfection and RT-qPCR in embryo samples

Immunofluorescence staining and confocal microscopy

Embryo culture with ML323 inhibitor

Public low-input ATAC-seq data of embryos

Identification and filtering of eccDNA

Quantification of eccDNA

Gene expression data processing

Analysis of public chromatin datasets

Overlaps of eccDNAs from human preimplantation development with de novo mutations

Genomic feature acquisition

Definition of gained, lost, stable, stage-specific, and consistent eccDNAs

Genomic distribution and gene annotation of eccDNA

Gene ontology (GO) enrichment analysis

Characterization of the eccDNA junction sites and motif-enrichment analysis

Fuzzy C-means clustering

Correlation between eccDNA and RNA expression

Structural variant (SV) calling in preimplantation embryo MDA samples

Inference of copy number variation

AmpliconArchitect (AA) analysis

Statistics & reproducibility

Reporting summary

Supplementary information

Source data

Acknowledgements

Author contributions

Peer review

Peer review information

Data availability

Code availability

Competing interests

Footnotes

Contributor Information

Supplementary information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES