Metagenomic analysis of vitamins B and K₂ biosynthesis in chicken gut microbiota across laying periods

Abstract

Background

The gut microbiota are crucial for synthesizing vitamins vital for chicken health and production, including vitamins B and K₂. However, the microbial pathways and temporal dynamics of these vitamins during different laying periods are not well understood, limiting targeted strategies to support poultry health and production. Clarifying these processes is essential for optimizing nutrition and enhancing poultry productivity.

Results

This study investigated the metagenomic landscape of microbe-driven vitamin biosynthesis with the aim of elucidating the chicken gut microbiome’s potential to produce vitamins B and K₂ across various laying periods. We collected and analyzed 26,053 chicken gut genomes from diverse sources, yielding 14,121 medium-quality, non-redundant genomes for downstream analysis. Genome clustering analysis identified 2,920 species-level genome bins, predominantly from Baccetota. The gene catalog contained approximately 15.09 million non-redundant genes, of which 1.90 million were associated with the biosynthesis of vitamins B and K₂. These genes were predominantly distributed among the phyla Bacillota, Bacteroidota, Pseudomonadota and Actinomycetota. Among the 14,121 non-redundant genomes, 3,453 high-quality genomes were identified as capable of de novo synthesizing at least one vitamin. Importantly, 7.67% of these genomes were capable of synthesizing five or more vitamins, while 33.85% could synthesize only one. The comparative genomic analysis of cobalamin biosynthesis underscores the dominance of the anaerobic pathway, with Bacillota emerging as a key contributor.

Conclusions

The findings highlight the microbiome’s crucial role in vitamin biosynthesis, showing substantial taxonomic and temporal variations. This study suggests that microbial involvement plays a pivotal role in vitamin synthesis, which could inform microbiota-based nutritional strategies to support poultry health and productivity.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-025-04296-7.

Background

Laying hens are a crucial source of high-quality protein, and their productivity plays a significant role in global food securit [1]. Gut health is essential for poultry productivity [2, 3], as the gastrointestinal (GI) tract is responsible for nutrient digestion and absorption. Disruptions to gut health impair immune development, nutrient uptake, and overall productivity, particularly in laying hens [4, 5]. The poultry GI tract hosts diverse microbial communities, predominantly bacteria [6], which are vital for maintaining GI homeostasis and protecting against pathogens. Beneficial bacteria form a protective barrier, produce essential nutrients such as vitamins and short-chain fatty acids and release antimicrobial compounds [7–9]. The microbiota supports metabolism, nutrient digestion, and immune function, which is especially crucial for egg production in laying hens [2, 10]. Disruptions in the gut microbiota can harm feed efficiency and productivity, leading to issues such as immune dysregulation, reduced fat digestibility, and mucosal degradation [5, 11].

The structure and composition of the poultry gut microbiota shift in response to various factors, with microbial community functions varying according to age, GI tract location, and diet [12, 13]. The microbiota functions as a “metabolic organ,” with its composition and activity highly responsive to dietary inputs [14]. Diet directly shapes the microbiome, with both macronutrients (carbohydrates, fats, and proteins) and micronutrients (vitamins and minerals) influencing gut microbial composition [15, 16]. Dietary vitamins can influence the microbiota’s balance [17, 18]. In laying hens, antioxidantssuch as vitamins C and E protect cells from oxidative damage, boosting production, immunity, and antioxidant levels, while also enriching beneficial gut microbiota in older hens [19]. On the other hand, gut microbes can synthesize essential vitamins, such as B vitamins, including biotin, cobalamin, folates, and riboflavin, as well as vitamin K, which are critical for metabolic and cellular functions that support poultry health. Understanding the composition, vitamin biosynthetic functions, and variations across different gut regions of the microbiota can provide valuable insights for developing targeted interventions to maximize microbiome benefits for poultry health and productivity. However, previous studies have rarely investigated how these vitamin biosynthetic functions vary over time, especially across different laying stages in hens. This limits our understanding of how the microbiota dynamically contributes to host nutrition and production.

Recent advances in metagenomics have enabled researchers to map the microbial composition, diversity, and functional capacities of the gut microbiome with unprecedented depth and precision [20, 21]. In this study, we employed a comprehensive metagenomic approach to cluster and analyze over 19,000 genomes from the chicken gut, uncovering significant microbial species diversity and genetic functions associated with vitamin biosynthesis. We hypothesize that the gut microbiota’s potential for vitamin biosynthesis differs across laying periods, both in terms of species composition and gene abundance. This study provides a genome-resolved perspective that links microbial taxonomy and function to specific egg-laying stages, which has not been comprehensively addressed in previous research. By cataloging these genomes, the study identified key microbial players and gene distributions for B and K₂ vitamin synthesis, examining both taxonomic and temporal patterns within the gut. Certain bacterial phyla, such as Bacillota and Bacteroidota, make particularly significant contributions to the biosynthesis of these essential vitamins. Moreover, their roles exhibit significant differences across different laying periods, such as the second peak laying period and the late laying period. The study further delves into the relationship between microbial vitamin synthesis and the laying periods in hens, focusing on identifying shifts in microbial diversity and gene abundance during high productivity phases. The data provide a foundational genomic catalog that enhances our understanding of vitamin biosynthesis in the chicken gut, with the potential to influence poultry nutrition strategies.

Methods

Data collection

All data analyzed in this study were obtained from publicly available repositories, and no new samples were collected by the authors. A total of 26,053 chicken gut microbiome genomes were collected from diverse chicken sources, and 40 metagenomic sequencing samples were specifically obtained from fecal samples of Hy-Line Brown laying hens across different laying stages. Specifically, we obtained 12,339 metagenome-assembled genomes (MAGs) from the National Microbiology Data Center (NMDC, (https://nmdc.cn/icrggc/)), 979 MAGs from the ENA database (PRJEB55375, PRJEB55374), and 6,786 MAGs from the figshare database ((https://dx.doi.org/10.6084/m9.figshare.24901884), (http://dx.doi.org/10.6084/m9.figshare.24901878), and (10.6084/m9.figshare.24681096.v1)). Additionally, we retrieved 5,949 MAGs and 40 metagenomic sequencing samples of Hy-Line variety brown laying hens at different laying stages from the NCBI database (Project ID: PRJNA1099794). The fecal samples were collected across five laying stages: the early laying period (HE, 120 days of age), first peak laying period (HPI, 187 days), second peak laying period (HPII, 253 days), third peak laying period (HPIII, 367 days), and late laying period (HL, 436 days). According to the information provided by the original data contributors, all samples were collected from the same breed (Hy-Line brown) under standardized management conditions, and no antibiotic treatment was administered during the sampling period. These consistent conditions help minimize variation caused by host genetics or antimicrobial exposure, although certain metadata limitations inherent in public datasets may still exist.

Given the diverse origins of these datasets, we adopted a unified quality control strategy to mitigate potential biases introduced by differences in sequencing platforms, host breeds, and rearing conditions. Detailed genome filtering and dereplication procedures are described in the following section.

Preprocessing and quality control of genomes

We used the predict function in CheckM2 (v1.0.1) [22] with default settings to assess the completeness and contamination levels of the 26,053 genomes collected. Genomes with completeness ≥ 50%, contamination < 5%, and completeness—5 × contamination ≥ 50% were retained, following widely accepted thresholds for medium-quality MAGs [23]. Based on these criteria, 18,229 genomes were selected for further analysis. These genomes were taxonomically classified using the classify_wf workflow in GTDB-Tk (v2.3.2) [24], with default parameters except that ANI-based filtering (ani_screen) was skipped using the –skip_ani_screen option. Strain-level dereplication was performed using dRep (v3.4.5) [25] with parameters (-pa 0.9, -sa 0.99, -nc 0.30, -cm larger), which are optimized for high-resolution clustering of closely related genomes. Additionally, open reading frames (ORFs) for these genomes were predicted using Prodigal (v2.6.3) [26] (options: -p single). Clustering was then performed with the easy-cluster workflow from MMseqs2 [27, 28] (options: –split-mode 2 –split-memory-limit 150G –cov-mode 2 -c 0.9 –min-seq-id 0.95 –cluster-mode 2 –cluster-reassign 1 –kmer-per-seq 200 –kmer-per-seq-scale 0.8), resulting in a non-redundant microbial gene catalog containing 15,092,300 genes.

Phylogenetic tree construction and functional annotation

We used dRep (v3.4.5) (33) (options: “-pa 0.9 -sa 0.95 -nc 0.30 -cm larger –S_algorithm fastANI”) to estimate the average nucleotide identity (ANI) between genomes with the same species-level classification, identifying 2,920 species-level genome bins (SGBs) as known species (ANI > 95%) [29]. Subsequently, we constructed a maximum likelihood tree for these genomes using the protein-coding sequences predicted by Prodigal with PhyloPhlAn (v0.99) [30]. Finally, the visualization was performed using the iTOL (v6.9.1) platform [31].

We searched the Kyoto Encyclopedia of Genes and Genomes (KEGG) database [32] using DIAMOND BLASTP (v2.1.8.162) [33] for functional annotation of 14,121 genomes. Functional annotation used DIAMOND BLASTP with a minimum score of 60, query coverage ≥ 70%, and up to 5 hits per query, ensuring confident and specific alignments. As a result, we selected 3,453 genomes predicted to be capable of de novo biosynthesis of vitamins B and K₂. Subsequently, the same method was employed to construct and visualize the maximum likelihood tree.

Preprocessing of raw data

We used fastp (v0.23.0) [34] to perform quality control on the raw reads from 40 fecal samples of Hy-Line variety brown laying hens (options: -q 20 -u 30 -n 5 -y -Y 30 -l 80 –trim_poly_g). Next, Bowtie2 (v2.5.0) [35] was used to align the filtered reads to the host genome (NCBI RefSeq assembly: GCF_016699485.2) to obtain clean reads for further analysis (options: –end-to-end –mm –fast).

Mapping and quantification of microbial genomes in fecal samples

High-quality reads (20 million reads) from 40 fecal samples were mapped to the microbial gene catalog using Bowtie2 (v2.5.0) [35] with the –end-to-end and –fast options to prioritize speed and maximal exact matching. The read counts were converted to transcripts per million (TPM), and the relative abundances of KEGG orthologous groups (KOs) were calculated based on gene abundance. Taxonomic profiles at the phylum and genus levels were determined by summing the abundances of all genes assigned to each category, with KO profiles derived in a similar manner. Functional roles in vitamin biosynthesis pathways were represented by KOs sharing the same functions, and the abundance of each functional role was calculated by summing the abundances of the associated KOs. The pathway abundance for vitamins B and K₂ was determined by summing the abundances of all KOs involved in these pathways.

Statistical analyses and visualization

Data analyses were conducted using R version 4.2.2. Taxonomic and functional gene abundance data were used to calculate the Simpson, Chao1 and Shannon indices. β-diversity was assessed through Principal Coordinate Analysis (PCoA) based on Bray–Curtis distance, and group differences were evaluated using permutational multivariate analysis of variance (PERMANOVA). The Wilcoxon rank-sum test was utilized to evaluate the significance of differences in diversity indices, taxonomic units, and functional gene feature abundances among groups. To account for multiple comparisons, P values were adjusted using the Benjamini–Hochberg false discovery rate (FDR) correction. Adjusted P values < 0.05 were considered statistically significant. The Min–Max normalization of the data was performed using the ‘caret’ package (v6.0.94) [36]. Rarefaction curves were created with the ‘vegan’ package (v2.6–4) [37]. The Sankey diagram was constructed using the ‘networkD3’ package (v0.4) [38]. Random forest analysis was conducted using the ‘rfPermute’ package (v2.5.2) [39]. All other visualizations were generated using the ‘ggplot2’ package (v4.2.3) [40].

Results

Large-scale genomic and gene data collection of the chicken gut microbiota

To obtain comprehensive genomic data on the chicken gut microbiota, we collected a total of 26,053 genomes. After applying quality filters (completeness ≥ 50%, contamination < 5%, and completeness—5 × contamination ≥ 50) and removing redundant sequences based on a 99% ANI threshold, a final dataset of 14,121 genomes met the quality standards and was used in subsequent analyses.

For classification of the chicken gut microbiota, genomes were clustered at a 95% ANI threshold, resulting in 2,920 SGBs (Fig. 1; Supplementary Table 1). Genome sizes ranged from 202,222 to 7,292,068 bp (mean: 2,124,699 bp), with completeness from 50.25% to 100% (mean: 88.22%) and contamination rates from 0 to 4.99% (mean: 1.11%). Using the GTDB-Tk [41], for annotation, these 2,920 SGBs were classified into 20 phyla, 35 classes, 88 orders, 198 families, and 744 genera. Bacillota emerged as the most dominant phylum, comprising 56.0% of the SGBs, followed by Bacteroidota (15.7%), Pseudomonadota (8.7%), and Actinomycetota (7.3%). At the genus level, the most abundant genera were Alistipes (1.7%), Mediterraneibacter (1.7%), Cryptobacteroides (1.5%) and Phocaeicola (1.1%). Additionally, we constructed a gene catalog containing 15,092,300 non-redundant genes, with an average length of 761.5 bp, all containing complete ORFs.

Fig. 1.

Phylogenetic relationships and genomic characteristics of 2,920 genomes from the chicken gut microbiome. The phylogenetic tree branches are color-coded by phylum-level classification. The first outer ring provides genus-level classification with distinct color codes for genera. The second outer ring presents genome size as a bar chart. The third outer ring indicates contamination levels for each genome, serving as a measure of data quality. The fourth outer ring represents genome completeness, reflecting the estimated coverage of each genome. The data highlight the diversity, taxonomic composition, and quality metrics of the genomes in the chicken gut microbiome, supporting further exploration of microbial contributions to gut health and function

Temporal changes in vitamins B and K₂ biosynthesis genes in chicken gut microbiota

To explore the temporal variation in the abundance of vitamin biosynthesis genes related to vitamins B and K₂, across different laying stages, we conducted Kyoto Encyclopedia of Genes and Genomes (KEGG) functional annotation of the gene catalog. A total of 1.90 million genes and 191 KOs were identified as involved in the biosynthesis of vitamins B (biotin, cobalamin, folate, niacin, pantothenate, pyridoxine, riboflavin, thiamine) and vitamin K₂ (menaquinone) (Supplementary Table 2). As the sample size increased, the rarefaction curve plateaued, indicating that the data volume was sufficient to capture the diversity of genes associated with vitamin biosynthesis capabilities mediated by the chicken gut microbiota during the laying periods (Fig. 2A).

Fig. 2.

Temporal dynamics of vitamin B and K₂ biosynthesis genes. A The rarefaction curve analysis illustrates the relationship between gene relative abundance accumulation and sample size expansion. B The total relative abundance of microbial genes involved in the biosynthesis of vitamins B and K₂ across the laying cycle is shown. The horizontal line represents the median, the whiskers indicate the minimum and maximum values within 1.5 times the interquartile range, and the boundary lines denote the lower and upper quartiles. C-D Raincloud plots combine dot, box, and distribution plots. The distribution plot shows the probability density of the data, while the dot plot displays the distribution of individual sample data points. The box plots illustrate the Shannon index and Simpson index of microbial genes involved in vitamin B and K₂ biosynthesis. Statistical significance was assessed using the Wilcoxon rank-sum test: * P < 0.05. E  The results of the Bray–Curtis principal coordinates analysis (PCoA) reveal microbial community structures at different laying stages. Ellipses represent 95% confidence intervals. Boxplots above and to the right display the sample scores for PCoA1 and PCoA3, with boxes indicating medians and quartiles. Error bars extend to the most extreme values within 1.5 times the interquartile range. The heatmap in the upper right corner highlights significant differences between groups along the first and third axes. Significance levels were determined using the Wilcoxon rank-sum test: * P < 0.05, ** P < 0.01, *** P < 0.001

Throughout the laying period, the abundance of genes involved in riboflavin biosynthesis was the highest, peaking during the second laying peak (HPII), which may reflect increased vitamin demand to support metabolic activity during this critical phase of egg production. In contrast, genes involved in menaquinone biosynthesis remained the least abundant (Fig. 2B). We further assessed the diversity of these genes across different laying periods using the Shannon and Simpson indices, both of which showed a significant decrease during HPII (Fig. 2C and D, P < 0.05). PCoA and PERMANOVA also revealed significant group differences (Fig. 2E; Supplementary Fig. 1). These results may suggest significant shifts in the gut microbiota community structure and function during HPII, which may influence the capacity for vitamin biosynthesis.

To further characterized the vitamin biosynthesis process in laying hens across different periods by examining the distribution of genes involved in vitamins B and K₂ synthesis across related pathways in the gut microbiota (Supplementary Figs. 2–4). The relative abundance of genes involved in the biosynthetic pathways of these vitamins varied dynamically across different laying periods.

Species distribution of vitamin biosynthesis genes

We explored the species distribution of genes associated with vitamin biosynthesis across different laying periods. As the sample size increased, the rarefaction curve plateaued, indicating that the data had reached saturation (Fig. 3A). No significant differences were observed in the Shannon or Chao1 indices among the laying periods (Fig. 3B and C), suggesting that overall gene diversity may have remained relatively stable. However, principal coordinates analysis (PCoA) revealed distinct clustering patterns among groups (Fig. 3D; P = 0.005), and PERMANOVA further confirmed significant differences in community structure between HPII and the adjacent laying periods, HPI and HPIII (Fig. 3E; P < 0.01). Interestingly, the phyla Bacillota and Bacteroidota harbor genes involved in the biosynthesis of nine different vitamins, with both exhibiting a high abundance of vitamin biosynthesis-related genes (Fig. 3F). Bacillota maintained consistently high levels of vitamin biosynthesis genes throughout the laying period, especially for biotin and pantothenate biosynthesis. A similar trend was observed for Bacteroidota. The relative abundance of these genes in Bacillota during HPII was significantly higher than in other stages (Supplementary Fig. 5; P < 0.05). In contrast, Chlamydiota, Deinococcota, Eremiobacterota, and Thermoproteota harbored genes solely for biotin biosynthesis, while Planctomycetota contained only genes for pantothenate biosynthesis.

Fig. 3.

Species distribution of vitamin biosynthesis genes. A  The rarefaction curve analysis shows the relationship between genomic relative abundance accumulation and the increase in sample size. B-C  Raincloud plots combine dot, box, and distribution plots. The distribution plot displays the probability density of the data, while the dot plot represents the distribution of sample data points. Box plots illustrate the Shannon index and Chao1 index of microbial genomes involved in vitamin B and K₂ biosynthesis. Statistical significance was determined using the Wilcoxon rank-sum test. D  Principal coordinates analysis (PCoA) revealed compositional shifts in microbial genomes associated with vitamins B and K₂ biosynthesis. Samples are plotted along the first and second principal coordinates (PCoA1 and PCoA2), with the percentage of explained variance provided. Ellipses indicate the 95% confidence intervals for each group. The bar plots at the top and right display the distribution of samples along PCoA1 and PCoA2, reflecting sample density within each group. E  The heatmap, based on permutational multivariate analysis of variance (PERMANOVA), illustrates the significance of differences between groups. Statistical significance was assessed using the Wilcoxon rank-sum test: * P < 0.05; ** P < 0.01; *** P < 0.001. F The bubble plot shows the phylum-level taxonomic distribution of genomes involved in vitamin B and K₂ biosynthesis throughout the entire laying period. The bubbles are colored according to their respective groups and sized based on transcripts per million (TPM) values. Abbreviations for vitamins B and K₂ are as follows: thiamine (Thi), riboflavin (Rib), niacin (Nia), pantothenate (Pan), pyridoxine (Pyr), biotin (Bio), folate (Fol), cobalamin (Cob), and menaquinone (Men)

At the genus level, genes involved in biotin biosynthesis were more abundant in Lactobacillus and Ligilactobacillus. (Supplementary Fig. 6A). Among the top 15 most abundant genera, Mediterranea and Turicibacter showed a significant decrease in relative abundance during HPII compared to other peak laying periods (HPI and HPIII) (Supplementary Fig. 6B, P < 0.05). In contrast, the relative abundances of Lactobacillus, Limosilactobacillus, and Ruminococcus were higher during HPII.

In addition, we used a random forest model to identify the key phyla involved in vitamin biosynthesis across different laying stages. The results indicated that Bacillota, Deinococcota, Spirochaetota and Eremiobacterota play critical roles in vitamin biosynthesis during various laying periods (Supplementary Fig. 7, P < 0.01).

Species-specific distribution of de novo vitamins B and K₂ biosynthesis genes

In this study, the biosynthesis of each vitamin was associated with one or more KEGG pathway maps. A genome was considered potentially capable of de novo vitamin biosynthesis if it contained all the KOs required for at least one complete biosynthetic route defined within the KEGG map for that vitamin. Here, we focused on the de novo biosynthesis of vitamins, with particular attention to the species distribution of genes involved in vitamin synthesis in the chicken gut microbiota. To investigate the de novo biosynthetic capacity for vitamins B and K₂ in the gut microbiota of laying hens, we identified 3,453 genomes capable of synthesizing at least one vitamin (Supplementary Table 3). These genomes exhibited high quality, with completeness ranging from 90.01% to 100% (mean: 96.40%) and contamination levels between 0% and 4.99% (mean: 1.02%). Genome sizes varied from 1,081,356 bp to 7,292,068 bp (mean: 2,537,816 bp), and GC content ranged from 25 to 74% (mean: 49.1%) (Fig. 4A). This indicates that a substantial portion of the chicken gut microbiota harbors the genetic capacity to synthesize essential vitamins, potentially contributing directly to host nutritional requirements.

Fig. 4.

The species-specific distribution of 3,543 genomes capable of de novo biosynthesis of vitamins B and K₂. A  The phylogenetic tree illustrates the evolutionary relationships of 3,453 genomes capable of independently biosynthesizing at least one vitamin de novo. Each branch is color-coded according to the phylum-level classification of the genomes. The colors of the first to ninth outer rings indicate the vitamins that these genomes can synthesize, while the tenth to thirteenth outer rings represent the completeness, contamination rate, size, and GC content of these genomes, respectively. Abbreviations for vitamins B and K₂ are as follows: thiamine (Thi), riboflavin (Rib), niacin (Nia), pantothenate (Pan), pyridoxine (Pyr), biotin (Bio), folate (Fol), cobalamin (Cob), and menaquinone (Men). B  The Sankey diagram illustrates the classification of 3,453 genomes across various taxonomic levels and their distribution of vitamins that can be independently biosynthesized de novo

The number of genomes capable of de novo biosynthesis for each vitamin is as follows: biotin (n = 613), cobalamin (n = 339), folate (n = 444), menaquinone (n = 315), niacin (n = 2,346), pantothenate (n = 1,821), pyridoxine (n = 813), riboflavin (n = 1,612), and thiamine (n = 290) (Supplementary Table 4). Interestingly, two genomes were found capable of independently synthesizing seven vitamins (excluding thiamine and menaquinone), while 66 genomes could synthesize six vitamins, and 197 genomes could synthesize five vitamins. However, no genomes were identified that could independently synthesize eight or more vitamins.

In the chicken gut microbiota, the biosynthesis of various vitamins was primarily driven by the phyla Bacteroidota and Bacillota (Fig. 4B, Supplementary Table 4). These two phyla contribute significantly to the biosynthesis of niacin, pantothenate, pyridoxine, and riboflavin, with Bacteroidota contributing 34.1%, 45.1%, 20.9%, and 31.9%, and Bacillota contributing 29.1%, 13.3%, 19.9%, and 38.7%, respectively, based on the number of genomes. The biosynthesis of biotin is mainly attributed to Bacillota (26.9%) and Pseudomonadota (22.6%). For menaquinone, the primary contributors are Campylobacterota (61.9%) and Bacteroidota (19.3%), while folate biosynthesis is predominantly derived from Bacteroidota (44.5%) and Pseudomonadota (27.0%). The genes responsible for cobalamin biosynthesis are overwhelmingly sourced from Bacillota (87.3%), and those for thiamine synthesis are mainly derived from Bacteroidota (67.5%).

Cobalamin biosynthesis in gut microbiota during laying stages, focusing on HPII

We investigated the distribution of genes involved in the de novo biosynthesis of cobalamin within the gut microbiota of laying hens. Cobalamin biosynthesis occurred through both aerobic and anaerobic pathways (Fig. 5; Supplementary Fig. 3). The aerobic pathway was particularly constrained by the low abundance of key functional genes CobG (TPM: 0.0017) and CobF (TPM: 0.0016) (Supplementary Table 5), which may suggest that the aerobic pathway is likely underutilized or less efficient in this setting. Interestingly, both biosynthetic pathways are limited by the extremely low abundance of the CobR gene (TPM: 0.00033) (Supplementary Table 5), suggesting it may be underutilized or less efficient in the chicken gut environment, potentially reflecting adaptation to anaerobic conditions prevalent during specific laying stages.

Fig. 5.

The distribution of genes involved in the de novo biosynthesis of cobalamin. Large circles represent the functional roles in the cobalamin biosynthetic pathway, with circular stacked bar charts inside the circles showing the distribution of these genes across different phylum-level classifications throughout the entire laying period. The color-coded bars within each chart correspond to the gene distribution for each phylum. Small circles represent the metabolites involved in the cobalamin biosynthesis process

Diversity analyses revealed that the Chao1 index of genes involved in cobalamin biosynthesis was significantly lower during the HPII compared to the HPIII and HL (Supplementary Fig. 8A; P < 0.05). Additionally, β-diversity analysis demonstrated significant differences between HPII and other egg-laying stages (Supplementary Fig. 8B and C, P < 0.01), which may suggest that the microbial community structure shifts during HPII. This shift could result in the predominance of specific microbial taxa over others, potentially altering the functionality and efficiency of cobalamin biosynthesis. Over 50% of the genes responsible for cobalamin biosynthesis were derived from Bacillota, followed by Actinomycetota, Bacteroidota and Methanobacteriota (Supplementary Fig. 9). At the genus level, most of these genes came from Lactobacillus, Ligilactobacillus, Limosilactobacillus, and Ornithinicoccus (Supplementary Fig. 10). The CobNST genes were predominantly sourced from Bacteroidota, while the CobC genes were mainly derived from Actinomycetota, followed by Bacillota (Fig. 5). In contrast, CobG genes were primarily sourced from Actinomycetota, with a secondary contribution from Pseudomonadota. CobF genes were exclusively derived from Actinomycetota (Supplementary Fig. 11). CobR genes were restricted to Bacillota, Actinomycetota, and Pseudomonadota, which may serve as a key factor limiting the efficiency of cobalamin biosynthesis in the gut microbiota of laying hens.

Discussion

The gut microbiota plays a crucial role in synthesizing essential vitamins. In this study, we provide an in-depth metagenomic analysis of the chicken gut microbiome’s role in vitamin biosynthesis with particular emphasis on identifying microbial genomes involved in biosynthesis of vitamins B and K₂ during the different laying period in hens.

The comprehensive genomic analysis of the chicken gut microbiota revealed significant microbial diversity and functional capacity, with 14,121 genomes clustered into 2,920 SGBs. This diversity spans 20 phyla, 35 classes, 88 orders, 198 families, and 744 genera, highlighting the complex and varied microbial ecosystem in chickens.l Bacillota emerged as the dominant phylum (56.0%), suggesting a critical role in gut health, while other prominent phyla, such as Bacteroidota, Pseudomonadota, and Actinomycetota, likely contribute to carbohydrate and protein metabolism. This result is consistent with previous studies showing that Bacillota and Bacteroidota, are most abundant in the chicken cecal microbiome, representing over 90% of the bacterial population [il42, 43]. The abundance of Bacillota is beneficial to the health of the bird because of its anti-inflammatory effects [44, 45]. The gene catalog contains over 8 million non-redundant genes with complete ORFs. By prioritizing quality, such as high completeness and low contamination, and minimizing redundancy, this dataset provides a robust foundation for understanding the role of the chicken gut microbiota in health and development. These findings offer valuable insights into the microbial composition and functional potential of the chicken gut microbiota, paving the way for future studies on its role in poultry health and nutrition.

The gut microbiota in poultry changes dynamically, and its diversity and richness are affected by age, breed, diet, feeding conditions, and the environment [46, 47]. The analysis of over 1.90 million genes and 191 KOs related to vitamin biosynthesis revealed dynamic changes in the chicken gut microbiota across laying stages. Riboflavin biosynthesis genes were consistently abundant throughout all stages, reaching their highest levels during the HPII period, likely reflecting increased demand for riboflavin due to its critical role in energy metabolism and antioxidant defense at peak production [48, 49]. In contrast, the persistently low abundance of menaquinone biosynthesis genes suggests limited microbial contribution, possibly compensated by dietary sources or host synthesis [50, 51]. These findings underscore the importance of microbial functions in meeting metabolic demands during key laying stages.

The vitamin biosynthesis-related genes in the chicken gut microbiota exhibited significant changes across different laying stages, particularly during HPII. Variations in Shannon and Simpson indices, as well as distinct clustering in PCoA analysis, suggest a functional reorganization of the microbiota during peak laying periods. This may be associated with changes in the taxonomic composition of the gut microbiota across different production stages in laying hens [52]. These findings indicate that the chicken gut microbiota may dynamically adjust its vitamin biosynthesis functions according to the laying cycle to meet metabolic demands and support nutrient synthesis during critical periods [53–55].

Next, we examined the species distribution of genes involved in vitamin biosynthesis across different laying stages. Although Shannon and Chao1 indices remained stable, indicating no significant changes in overall α-diversity, PCoA and PERMANOVA revealed distinct clustering during HPII. This suggests a temporal shift in the composition of vitamin-producing taxa. Such changes may reflect physiological adaptations or unmeasured factors—such as stress, metabolic fluctuations, or hormonal changes—associated with peak laying. These findings indicate that the gut microbiota may adjust its functional capacity in response to increased host demands during critical stages of production [52, 55]. Previous studies have identified Bacillota and Bacteroidota as the dominant phyla in the gut microbiota of laying hens [56]. This findings further highlight the importance of these key phyla in vitamin biosynthesis. Specifically, Bacillota, Bacteroidota and Actinomycetota were identified as major contributors to vitamin biosynthesis. Bacillota consistently exhibited high gene abundance across all laying stages, particularly in the synthesis of biotin and pantothenate [8], while Bacteroidota contributed to niacin and riboflavin production. At the genus level, Lactobacillus and Ligilactobacillus harbored abundant biotin biosynthesis genes, whereas Ornithinicoccus mainly contributed to pantothenate synthesis.y Interestingly, during HPII, Lactobacillus abundance increased as Ornithinicoccus declined, possibly suggesting a microbial shift toward genera more capable of supporting host metabolic demands [55]. Given their biosynthetic capacity and probiotic potential, Lactobacillus strains may serve as functional additives to enhance nutrient synthesis and gut health in laying hens [57]. Similarly, vitamin-producing members of Bacteroidota may offer dietary intervention value during peak laying periods [58].

This study highlights the critical role of the chicken gut microbiota in vitamin biosynthesis. A total of 3,453 high-quality genomes were identified with the ability to synthesize various vitamins. However, no single genome could produce all vitamins independently. This finding indicates a complex and collaborative process involving multiple species [59]. The phyla Bacteroidota and Bacillota dominate the biosynthesis of niacin, pantothenate, pyridoxine, and riboflavin, while biotin is mainly synthesized by Bacillota and Pseudomonadota, and menaquinone by Campylobacterota. Folate and thiamine production rely on Bacteroidota, and cobalamin biosynthesis is almost exclusively carried out by Bacillota. This functional specialization highlights the metabolic complementarity of the gut microbiota, providing a multi-species system that supports host nutrition and intestinal health [60, 61]. Modulating key gut microbiota may offer new strategies for the development of probiotics or functional feed additives [60].

Cobalamin, or vitamin B₁₂, is a unique vitamin that cannot be synthesized by animals, plants, or fungi; it is exclusively produced by microorganisms, especially anaerobic bacteria [62]. Comparative genomics indicates that anaerobic pathways dominate cobalamin biosynthesis, with Bacillota as a major contributor. Over 50% of cobalamin biosynthesis genes originate from Bacillota, followed by Actinomycetota, Bacteroidota and Methanobacteriota. This may suggests that maintaining a healthy balance of these microbial groups is likely crucial for sustaining adequate levels of cobalamin biosynthesis. The aerobic pathway of cobalamin biosynthesis is likely limited by the low abundance of key genes such as cobG and cobF. This pattern is also observed in ruminants [63]. Moreover, cobR was particularly scarce during HPII, which may further restrict cobalamin production in the gut microbiota of laying hens.This suggests that endogenous synthesis may be insufficient to meet the hens’ nutritional demands, especially during peak production phases [64–67]. Laying hens likely rely on dietary cobalamin or specific bacteria with more active biosynthetic pathways. Previous studies show that cobalamin supplementation improves egg production, eggshell quality, and gut health, highlighting the importance of dietary support [68, 69]. In ruminants, cobalamin biosynthesis is largely mediated by genera such as Prevotella, Methanobrevibacter and Clostridium [63]. In contrast, our findings in laying hens identify Lactobacillus, Ligilactobacillus and Limosilactobacillus as key contributors. These observations suggest that different host species may rely on distinct microbial taxa for cobalamin production.

Collectively, these data highlight the potential influence of microbial diversity, and egg-laying cycles on the biosynthesis capabilities of chicken gut microbiota of essential vitamins. The study results provide a foundation for further analysis of microbial contributions to improved gut function and health through tailored microbial and nutritional interventions, possibly through probiotics or dietary adjustments. In poultry production, stressors such as heat, pathogens, and housing conditions may alter gut microbial composition and potentially affect vitamin biosynthesis [58, 70]. Although our study was based on controlled conditions, the stability of vitamin-producing taxa under stress remains unclear and warrants further investigation. Several limitations should be noted, including inconsistent rearing conditions among datasets, lack of direct vitamin measurements, and limited analysis of microbial interactions. Additionally, metagenomic assembly may miss low-abundance or high-GC genomes, potentially underrepresenting some taxa. Nevertheless, the large dataset, quality control procedures, and focus on gene-level functional annotation help mitigate these biases. Future studies should incorporate experimental validation to confirm microbial functions and assess their resilience under practical production challenges.

Conclusion

This study employed metagenomic approaches to comprehensively investigate the role of gut microbiota in the biosynthesis of vitamins B and K₂ across different laying periods in hens. By collecting and processing a large dataset, 14,121 genomes and a comprehensive gene catalog were obtained. The results revealed dynamic changes in the relative abundance, diversity, and species distribution of vitamin biosynthesis genes throughout the laying cycle, with distinct patterns observed for riboflavin and menaquinone biosynthesis genes. The second laying peak emerged as a critical turning point. A total of 3,453 genomes with vitamin biosynthetic capabilities were identified, highlighting metabolic complementarity and cooperation among microbial taxa. Additionally, the anaerobic cobalamin biosynthesis pathway and the scarcity of key genes indicate that hens may require external supplementation during specific stages. Certain Lactobacillus and Bacteroidota strains with vitamin biosynthesis potential may serve as candidates for probiotic development or dietary supplementation strategies. Overall, this study provides insights into the relationship between the vitamin biosynthetic potential of gut microbiota and the laying cycle, and supports microbiota-informed nutritional interventions such as probiotic supplementation or feed adjustment to enhance vitamin production during peak laying periods.

Abbreviations

HE

The early laying period

HPI

The first peak laying period

HPII

The second peak laying period

HPIII

The third peak laying period

HL

The late laying period

Authors’ contributions

Z.Q.G conceptualized the study and wrote the original draft. J.W.S contributed to the methodology, software development, and also wrote the original draft. Y.Q curated the data and participated in the review and editing of the manuscript. T.Y provided critical resources and contributed to the review and editing process. H.C supplied essential resources and assisted with manuscript review and editing. L.H.Y offered conceptualization resources, contributed to the methodology and software components, and was involved in reviewing and editing the manuscript. H.M.E conceptualized the study, supervised the project, and played a key role in writing the original draft. All authors read and approved the final version of the manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

All data used in this study are publicly available. Specifically, 12,339 MAGs were retrieved from the National Microbiology Data Center (NMDC, (https://nmdc.cn/icrggc/)), 979 MAGs from the ENA database (PRJEB55375, PRJEB55374), 6,786 MAGs from Figshare ((https://dx.doi.org/10.6084/m9.figshare.24901884), (http://dx.doi.org/10.6084/m9.figshare.24901878), (http://doi.org/10.6084/m9.figshare.24681096.v1)), and 5,949 MAGs as well as 40 metagenomic sequencing samples of Hy-Line variety brown laying hens from the NCBI database (Project ID: PRJNA1099794).

Declarations

Ethics approval and consent to participate

This study did not involve live animals or human participants, nor did it involve field collection of specimens. All metagenomic data analyzed in this study were obtained from publicly available databases, and no new samples were collected. Therefore, no ethical approval or specific permissions were required. The use of publicly available data complied with institutional, national, and international guidelines.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.