Probiotic application to hatching egg surface supports microbiota development and acquisition in broiler embryos and hatchlings

Abstract

In modern poultry production, hatchlings primarily acquire their initial gut microbiota from the hatchery environment. This, in turn, results in delayed gut microbiota acquisition and a less diverse microbiota. However, the gut microbiota in the hatchling and the neonate is crucial for the normal development of the immune system and healthy gut function. Hence, promoting healthy microbiota development and acquisition in hatchlings is critical. To this end, we determined the potential for probiotic spray application to hatching eggs to support microbiota acquisition in the hatchling. A total of 100 hatching eggs (Ross 308) were either sprayed with phosphate-buffered saline (PBS; Control) or probiotics [∼9 log CFU/egg of Lacticaseibacillus rhamnosus NRRL B-442 (LR) or Lacticaseibacillus paracasei DUP 13076 (LP)] during incubation. Six eggs were sacrificed for sample collection at each sampling point. Eggshells were washed with sterile PBS buffer and collected at embryonic day (D) 0, 7, 14, 18, and 20. Chorioallantoic membrane (CAM) was collected at D7, 14, 18, and 20, and intestine at D14, 18, and 20. At hatch, chicks were euthanized, and the cecal and ileal samples were collected for microbiota characterization. Results indicate that spray application of LP and LR significantly modulated the microbiota associated with the eggshell, CAM, embryonic intestine, and hatchling gut. Further predictive analysis revealed CAM to be a significant source for the microbiota associated with the hatchling gut. Also, application of LP and LR led to significant enrichment in Lactobacillus populations and potential probiotic taxa, including Enterococcus, in the hatchling gut. Moreover, functional profile analysis revealed that microbial communities associated with the hatchling gut microbiota in the probiotic groups were enriched for nutrient and energy metabolism, which could not only support embryo development but also post-hatch growth. In conclusion, in ovo spray application of probiotics to the egg’s surface can be a potential approach to support microbiota acquisition in hatchlings.

Introduction

Over the last decade, our understanding of the gut microbiota and its crucial role in poultry health and performance has led to the realization of the need to support the establishment of a healthy microbiota in chicken (Brisbin et al., 2008; Pan and Yu, 2014; Ballou et al., 2016). It has been noted that there is a difference in gut microbiota development between chickens raised commercially and those raised with adult birds. The former primarily acquires its microbiota from the rearing environment, while the latter acquires adult chicken’s microbiota from the hen, particularly in the first week post-hatch when the gut is naïve with limited microbial biomass (Rychlik, 2020). In a commercial setting, chicks are hatched via artificial egg incubation, where direct transmission of the microbiota from hens to hatchlings is improbable, resulting in delayed gut microbiota acquisition and a less diverse microbiota in chicks (Kubasova et al., 2019; Li et al., 2022; Shterzer et al., 2023). In addition, sanitizing hatching eggs before incubation is a common practice to control for any pathogens on the egg's surface (Oliveira et al., 2022). This practice can potentially remove any inherent microbiota associated with the eggs that could have originated from the laying hen (Gao et al., 2025). Moreover, surface sterilization was found to reduce microbial diversity in the yolk microbiota during embryonic development, suggesting potential adverse effects on embryonic immune activation and the early establishment of intestinal microecological homeostasis (Ding et al., 2024).

The gut microbiota in the hatchling and the neonate is crucial for the normal development and maturation of the immune system including pathogen exclusion (Lu et al., 2003; Shterzer et al., 2023). Further, the micro also help in maintaining healthy gut function and nutrient uptake in broilers (Rychlik, 2020; Liu et al., 2021). In addition, it has been revealed that chicken microbiota development starts as early as the first week of embryonic development (Akinyemi et al., 2020; Ding et al., 2021). Overall, the difference in microbiota acquisition in commercial chicks and its potential impact on health and performance in grow-birds highlight the importance of establishing a healthy gut microflora in the chicks (Pan and Yu, 2014; Liu et al., 2021; Shterzer et al., 2023). Towards this, researchers have realized that an ideal route for the development of a healthy microflora would be to establish a balanced gut microbial population early on rather than attempting to alter an already established microbiota (Roto et al., 2016). However, currently, the commonly employed methods to modulate the chicken gut microbiota focus mostly on adult birds (Salim et al., 2013; Yadav and Jha, 2019; Aruwa et al., 2021).

Given the importance of the initial microbiota and the difficulty with effectively modulating an established microbiota, strategies to support and promote early microbiota acquisition in chicks is of great importance. In this regard, probiotics are ideal candidates to mediate early-life programming in broiler chickens (Shehata et al., 2021). Early exposure to probiotics can serve to ensure the establishment of a healthy gut microbiota in chicks (Dibner et al., 2008; Ballou et al., 2016). Previous studies have shown that in ovo injection of probiotics at embryonic day 18 modulated the microbial composition and community structure in the hatchling gut with an associated increase in chick weight (Pedroso et al., 2016; Arreguin-Nava et al., 2019; Wilson et al., 2019). Albeit these effects, injections in embryos have been associated with a negative impact on hatchability (Yamawaki et al., 2013; de Oliveira et al., 2014; Geng et al., 2022).

As an alternative to this invasive procedure, in this study we employed probiotic spray application to the eggshell to support microbiota development in the embryo and hatchling (Amalaradjou, 2022). Besides serving as a physical protective barrier, the eggshell is shown to be one of the primary sources for the intestinal microbiota of chicks/embryos/hatchlings (Maki et al., 2020). Along these lines, eggshell penetration studies, primarily focused on Salmonella, have demonstrated that bacteria can penetrate the eggshell and shell membrane eventually translocating to the yolk and intestine in developing embryos, though this occurred in a small proportion of eggs (Berrang et al., 1999; Cook et al., 2003). In support of this, Maki and co-authors (Maki et al., 2020) demonstrated that inoculating different microorganisms on the shell of hatching eggs resulted in distinct microbial communities in the hatchling intestine. Additionally, it has been reported that spraying probiotics in the delivery environment can serve as an early intervention to modulate gut microbiota and improve overall growth and health in suckling piglets (Huang et al., 2023). This highlights the possibility of employing probiotic sprays as an early and non-invasive approach to support microbiota acquisition in chicks. Moreover, recent research from lab demonstrated that early spray application of probiotics to hatching eggs improved embryonic growth, muscle development, hatchability and hatchling quality (Muyyarikkandy et al., 2023a;2023b; Gao et al., 2024). As a next step, in the current study our objective was to evaluate the feasibility of using probiotic spray applications to hatching eggs to modulate microbiota development in the hatchling. We hypothesized that the in ovo spray application of probiotics on eggshell supports embryonic microbiota development and facilitates subsequent acquisition in hatchlings.

Materials and methods

Probiotic culture preparation

Lacticaseibacillus rhamnosus NRRL B-442 (LR) was obtained from the USDA Agriculture Research Service NRRL culture collection (Peoria, IL, USA). Lacticaseibacillus paracasei DUP 13076 (LP) was kindly provided by Dr. Bhunia, Molecular Food Microbiology Lab, Purdue University, West Lafayette, IN, USA. LR and LP were selected based on our previous research demonstrating the growth-promoting potential of LR and LP in developing embryos (broiler and layer) and pullets (Amalaradjou, 2022; Muyyarikkandy et al., 2023b; 2023a; Gao et al., 2024). The probiotic cultures were grown in de Mann, Rogosa, Sharpe broth (MRS; Fisher Scientific, Waltham, MA, USA) at 37°C for 16–18 h. Overnight cultures were centrifuged (3500 g, 10 min, 4°C), and washed twice with sterile phosphate buffered saline (PBS, pH 7.0). Probiotic counts were determined following serial dilution and plating on MRS agar and incubated at 37°C for 24–48 h (Muyyarikkandy and Amalaradjou, 2017).

Experimental design, spray application, egg incubation and hatching

This study was conducted at the UConn poultry research unit with approval from the UConn Institutional Animal Care and Use Committee. Hatching eggs (Ross 308) from 50-week-old birds were kindly provided by Aviagen (Huntsville, AL, USA). On receipt, damaged eggs were discarded, and the rest were stored at 12.8°C for no more than 24 h (Christensen et al., 2002; Gao et al., 2024). Before treatment application, eggs were removed from the chiller and tempered overnight at ambient temperature. At the start of the study, six eggs were randomly sampled to determine the starting microbial composition. The remaining eggs were randomly assigned to the three treatment groups (∼ 30 eggs/group). Group 1: Eggs sprayed with PBS (vehicle Control), Group 2: Eggs sprayed with LP, Group 3: Eggs sprayed with LR. For the spray application, eggs in each group were individually sprayed with different probiotic cultures (∼ 9 log CFU/egg) or sterile PBS (Control) at embryonic day (D) 0, 3, 7, 10, 14, and 18 of incubation using an atomizer as previously described (Amalaradjou, 2022; Muyyarikkandy et al., 2023b; 2023a). These time points for spraying were chosen to help maintain significant probiotic populations on the egg surface. The eggs were incubated in a GQF incubator with an automatic egg turner (GQF Manufacturing Company Inc., Savannah, GA, USA) at 37.8°C and 55 %–60 % relative humidity from day 0 to day 18. On D18, eggs were transferred to a GQF hatcher and incubated at 37.8°C and 65 %–70 % relative humidity until hatch (Aviagen, 2020). Throughout the study, eggs in different groups were placed in separate incubators and hatchers to avoid cross contamination (Archer and Cartwright, 2017).

Sample Collection

All samples were collected aseptically in a biosafety cabinet. At each sampling time, samples including eggshell wash, chorioallantoic membrane (CAM), and intestine were collected from six embryos/group for microbiota analysis (n = 6/group). For the eggshell microbiota, eggs were collected on day 0, 7, 14, 18 and 20 using sterile gloves and transferred to individual sterile sample bags (Fisher Scientific, Waltham, MA, USA) containing 10 ml of sterile PBS. The eggs were hand rubbed for 2 min and the wash buffer was transferred to sterile tubes and concentrated by centrifugation (5000 g, 10 min, 4°C) (Maki et al., 2020). After centrifugation, the supernatant was discarded, and the pellet was resuspended in 500 μl of sterile PBS and stored at −80 °C until further processing. Following the eggshell wash, each egg was removed from the sampling bag and sanitized using 70 % ethanol to avoid potential contamination of the internal contents from the microbiota present on the external shell surface (Upadhyaya et al., 2015). The eggs were then opened, and embryos were collected following euthanasia by cervical dislocation. chorioallantoic membrane (CAM) was collected using sterile scissors and forceps on day 7, 14, 18 and 20 of incubation. Following this, the embryo was opened, and entire intestines (small and large intestine) were collected on day 14, 18 and 20 of incubation (Akinyemi et al., 2020). In the case of the hatchling, chicks were euthanized by CO₂ inhalation and intestinal samples (cecal content and ileal samples) were collected separately for microbiota analysis (n = 6/group). All the samples were flash frozen in liquid nitrogen and stored at −80°C for further DNA extraction and 16S rRNA sequencing (Glendinning et al., 2019). Sample metadata is summarized in Supplementary Table 1.

DNA extraction and 16S rRNA sequencing

Microbial communities were characterized using paired-end 16S rRNA gene sequencing. DNA was extracted using the NucleoMag DNA Microbiome kit (Macherey-Nagel, Düren, Germany) according to the manufacturer’s protocol. DNA extracts were quantified using the Quant-iT PicoGreen kit (ThermoFisher Scientific, Waltham, MA) and 30 ng of DNA was used as the template for amplification. Partial bacterial 16S rRNA (V4) genes were amplified using 515F and 806R primers with Illumina adapters and dual indices (Kozich et al., 2013). Samples were amplified in triplicate using Go-Taq DNA polymerase (Promega, Madison, WI) with the addition of bovine serum albumin (New England BioLabs, Ipswich, MA). The PCR reaction mix was incubated at 95°C for 3.5 min, followed by 30 amplification cycles (30 s at 95°C, 30 s at 50°C and 90 s at 72°C) and final extension at 72°C for 10 min. PCR products were pooled using the epMotion 3075 liquid handling robot and were quantified using the QIAxcel DNA Fast Analysis (Qiagen, Germantown, MD). The pooled PCR products were cleaned using Omega Bio-Tek Mag-Bind Beads according to the manufacturer’s protocol (Omega Bio-Tek, Norcross, GA). The cleaned PCR products were sequenced on the MiSeq using v2 2 × 250 base pair kit (Illumina, San Diego, CA) at the UConn MARS facility. Mock communities (Zymo, Irving, CA) were sequenced as positive controls. Sterile PBS and PBS rinsate obtained after washing autoclaved scissors and forceps were sequenced to account for any potential contamination from the PBS or surgical instruments used during tissue sampling.

Bioinformatic analysis

Raw sequences were processed using QIIME2 (version 2023.5; Bolyen et al., 2019; Cheng et al., 2022). Samples with low sequencing depth (< 2000 reads) were excluded from further analysis (Caporaso et al., 2011). Sequences were demultiplexed by per-sample barcodes and denoised using the DADA2 pipeline to correct for Illumina-sequenced amplicon read errors (Callahan et al., 2016). Sequences were clustered into amplicon sequence variants (ASVs). Bacterial taxonomy was assigned using the QIIME2 BLAST method with the silva database (version 138) (Quast et al., 2013; Glöckner et al., 2017). Sequences that were identified as mitochondria, chloroplast or that were unassigned at the phylum level were removed.

Downstream analyses were performed using R (version 4.3.2; R Core Team, 2021). Taxonomic data were normalized using total sum scaling (McKnight et al., 2019). Alpha diversity was measured using Shannon, Simpson, Richness and Chao1 index using vegan package (version 2.6-6.1; Oksanen et al., 2024). A Kruskal-Wallis test was applied to identify significant differences in the alpha diversity indices, followed by Dunn’s test with false discovery rate (FDR) correction for pairwise comparison. Beta diversity was measured using Bray-Curtis principal coordinate analysis (PCoA) using vegan package. Significance was tested by a permutational multivariate analysis of variance (PERMANOVA) using the adonis2 function (permutations=999). Differences in alpha and beta diversity indices between the different groups were considered to be significant at p ≤ 0.05.

Linear discriminant analysis effect size (LEfSe) was conducted using the microbiomeMarker package (version 1.11.0) to identify differentially abundant taxa across groups using one against one strategy (Cao et al., 2022). Significant difference was determined when p ≤ 0.05 and logarithmic LDA score >3.5. A BLAST search against the GenBank database was performed to determine the identities of ASVs associated with significantly differentially abundant taxa (Liu et al., 2021). ASVs present in less than 10 % samples were filtered out (filter at 10 % prevalence) for LEfSe analysis (Nearing et al., 2022). Different sample types (eggshell, CAM, intestine, cecum and ileum) were filtered separately. SourceTracker package (version 1.0.1) was used to analyze the possible sources and their relative contribution to the hatchling gut (cecum and ileum) microbiota (Knights et al., 2011). Eggshell (day 0, 7, 14, 18, and 20), CAM (day 7, 14, 18, and 20), and embryonic intestine (day 14, 18, and 20) were considered as sources, and the hatchling cecum and ileum (day 21) were considered as sinks. Source attribution was performed separately for each of the three groups (Control, LP, LR). ASVs identified in at least two sink samples were considered for source attribution. A random forest model with leave-one-out cross-validation (ntree=500, seed=123) was used to predict important biomarkers in the cecal and ileal microbiota that can distinguish the Control, LP, and LR groups using the caret package (version 7.0-1) after filtering at 10 % prevalence (Kuhn, 2008). Predictive analysis of the microbial functional profile of the CAM, hatchling cecum and ileum microbiota was performed using PICRUSt2 (version 2.4.1) (Douglas et al., 2020). The identified sequence abundances were normalized by the 16S rRNA gene copy number. The predicted metabolic pathways were identified using the Kyoto Encyclopedia of Genes and Genomes database (KEGG; Kanehisa et al., 2016), and linear discriminant analysis effect size (LEfSe) was conducted to determine the significantly different pathways across different groups within each tissue type and sampling time using one against one strategy with significance tested at p ≤ 0.05 and logarithmic LDA score >3.5.

Results

We analyzed the microbial communities associated with the eggshell, CAM, and intestine of broiler embryos during incubation and the ileum and cecum in hatchlings. Of the 214 samples collected, 207 samples passed sequence quality filtering and were used for further analysis. Following sequence processing and filtering, a total of 2,619 amplicon sequence variants (ASVs) with a total of 7,994,170 high quality bacterial 16S rRNA gene sequences were retained for downstream analysis. After removing sequences assigned to chloroplast, mitochondria, and those unassigned at the phylum level, a total of 2,397 ASVs were retained (Supplementary Table 2). Across different sample types, 30 phyla, 225 families, and 459 genera were identified (Supplementary Table 3).

Effect of probiotic application on eggshell microbiota composition and diversity

Since our goal in this study was to use probiotic spray application on hatching eggs to modulate the microbiota of the growing embryo and eventually the hatchling, it is imperative to characterize the eggshell microbiota. In the Control samples, spray application of PBS did not significantly alter the alpha diversity when compared to the unsprayed eggs (Shell_D0; p > 0.05; Supplementary Table 4A). On the other hand, our data analysis revealed that probiotic spray application significantly reduced the microbial diversity associated with the eggshell microbiota when compared to the Control throughout the incubation period (p ≤ 0.05; Supplementary Table 4A). We also observed that the Shannon and Simpson indices of the eggshell microbiota in the LP group were significantly different from LR at different sampling times during incubation (p ≤ 0.05; Supplementary Table 4A). To further elucidate the effect of probiotic application on eggshell microbiota, we employed the principal coordinate analysis (PCoA) based on the Bray-Curtis distance matrix. As seen with the alpha diversity, we observed that the eggshell microbial communities at the start of the incubation period between the Control and unsprayed eggs (Shell_D0) were similar as evidenced from our PERMANOVA analysis (p > 0.05, Fig. 1A). Whereas we observed clear spatial separation between the bacterial community associated with the Control, LP and LR groups indicative of significantly different unique eggshell microbial community structure (PERMANOVA p ≤ 0.01, Fig. 1A). Moreover, this difference persisted throughout the incubation period, thereby demonstrating the sustained ability of probiotics to significantly modulate the eggshell microbiota in comparison to the Control. Furthermore, the separate clustering of the eggshell microbiota in the LR and LP groups also reflect a species-specific effect.

Fig. 1.

Changes in the eggshell associated microbial communities between different groups through incubation. (A) Principal coordinate analysis (PCoA) of the bacterial communities of different groups at different time points. (B) Stacked bar charts showing the relative abundance of bacterial community (mean value) at the family level. Bacterial families with relative abundance < 1 % are grouped as others.

Abbreviations: LP, Lacticaseibacillus paracasei DUP13076; LR, Lacticaseibacillus rhamnosus NRRL B-442.

Further analysis of the relative abundance of the eggshell bacterial community revealed that the most abundant bacterial phylum on the eggshell was Firmicutes followed by Actinobacteriota and Bacteroidota across all groups and over the different sampling times (Supplementary Fig. 1A). At the family level (Fig. 1B), the most abundant family were Lactobacillaceae (22.1±18.5 %) and Staphylococcaceae (22.1±6.2 %) on the eggshell at the start of the experiment (Shell_D0). Similarly, following PBS spray application in the Control, Lactobacillaceae (39.4±18.6%–26.6±13.4 %) and Staphylococcaceae (13.2±4.0% – 15.6%±3.4 %) continued to dominate the eggshell microbiota throughout the incubation period (day 7 to day 20; Fig. 1B). At the genus level with the Control group, the most abundant genera were Lactobacillus (27–41 %), followed by Salinicoccus (5.7-9.6 %), Staphylococcus (2.0–2.2 %), Brachybacterium (4.9–11.4 %), Brevibacterium (3.3–6.3 %), Romboutsia (3–6 %), and Corynebacterium (1.2–4.2 %) on the eggshell surface from day 7 to day 20 (Supplementary Fig. 1B). On the other hand, the family Lactobacillaceae and genus Lactobacillus made up more than 99 % of the eggshell bacterial community in the LP and LR groups (Fig. 1B; Supplementary Fig. 1B) while this genus only represented 22 % in the eggshell microbiota of the Control group. This is expected given the in ovo spray application of ∼9 log CFU of LR/LP to the egg surface. Further, the dominant bacterial taxa associated with the eggshell microbiota did not change throughout embryonic development (day 0 to day 20) in the Control group.

Effect of probiotic application to hatching egg surface on CAM microbiota diversity, composition and predicted functional profiles

CAM is a highly vascularized structure that surrounds the embryo and lines the inner eggshell. It develops from embryonic day 5 and degrades after day 19 (Halgrain et al., 2022). Functionally, the CAM is involved in key physiological and metabolic activities including calcium transport from the eggshell to the bloodstream of the developing embryo

(Halgrain et al., 2022). Given its significant role, starting on day 7 of incubation, we sampled the CAM to characterize its microbiota composition, to determine how probiotic application to the eggshell might impact it and its overall contribution to the microbiota of the developing embryo and hatchling. We observed no significant difference in the alpha diversity of CAM-associated microbiota among the three different groups at all time points (p > 0.05; Supplementary Table 4B). Looking at the bacterial community structure (beta diversity), although similar on day 7, the bacterial community associated with the LP-treated eggs clustered distinctly and separately from Control and LR after day 14 (PERMANOVA p ≤ 0.05; Fig. 2A), indicating a species-specific effect of the probiotic in modulating the CAM microbiota composition.

Fig. 2.

Changes in chorioallantoic membrane (CAM) associated microbial communities between different groups through the incubation. (A) Principal coordinate analysis (PCoA) of the bacterial communities of different groups at different time points. (B) Stacked bar charts showing the relative abundance of bacterial community (mean value) at the family level. Bacteria families with relative abundance < 1 % are grouped as others. (C) Differentially enriched taxonomic profiles of CAM in descending order at (C1) day 14, (C2) 18, and (C3) 20 (p ≤ 0.05, LDA score>3.5, one against one strategy). (D) Differentially enriched functional profiles of CAM in descending order at (D1) day 14, (D2) 18, and (D3) 20 (p ≤ 0.05, LDA score>3.5, one against one strategy).

Abbreviations: LP, Lacticaseibacillus paracasei DUP13076; LR, Lacticaseibacillus rhamnosus NRRL B-442.

In terms of the relative abundance of the CAM-associated bacterial community, the predominant phylum observed was Firmicutes across all groups at all time points with higher abundance in LP and LR (Supplementary Fig. 2A). Besides Firmicutes, other phyla associated with the CAM included Proteobacteria (Control: 33–60 %; LP: 16–37 %; LR: 16–48 %) and Actinobacteriota (Control: 2–12 %; LP: 1 %; LR: 0–8 %). As opposed to a higher abundance of Firmicutes in the probiotic groups, LP and LR had a lower predominance of Proteobacteria in the CAM when compared to Control (Supplementary Fig. 2A). At the family level, as seen with the eggshell, LP group consistently had a significantly higher proportion of Lactobacillaceae with their predominance changing over time (LEfSe p ≤ 0.05, LDA>3.5; Fig. 2B and 2C). For example, Lactobacillaceae abundance on the CAM increased from 24.2±14.2 % (day 7) to 66.1±11.4 % (day 14) and then decreased to 50.6±14.0 % (day 18) and 50.8±18.0 % (day 20) in LP group (Fig. 2B). Overall, throughout the incubation, Lactobacillaceae was significantly enriched in LP group when compared to LR and Control (LEfSe p ≤ 0.05, LDA>3.5, Fig. 2C). Moreover, a random forest model predicted that Lactobacillaceae was the most important bacterial family to distinguish groups following pairwise comparisons, indicating the significant role of Lactobacillaceae in the CAM associated microbiota (Fig. 4A). Additionally, the accuracy score for predicting Control or LP group is 0.81, showing the distinct microbial community between the two groups. The dominant genera in the CAM included Lactobacillus (Control: 0.6 ± 0.6 % – 8.7 ± 7.9 %; LP: 24.6 ± 14.3 % – 67.1 ± 11.0 %; LR: 7.6 ± 6.6 % – 25.2 ± 21.8 %), Enterococcus (Control: 2.6 ± 2.6 %–21.6 ± 18.1 %; LP: 0.0 ± 0.0 % – 20.0 ± 20.0 %; LR: 0.0 ± 0.0 % – 20.0 ± 20.0 %), Clostridium sensu stricto 1 (Control: 0.6 ± 0.6 % – 17.8 ± 10.9 %; LP: 0.0 ± 0.0 % – 12.3 ± 12.3 %; LR: 0.0 ± 0.0 % – 17.4 ± 17.4 %), and Bacillus (Control: 0.6 ± 0.6 % – 13.4 ± 7.1 %; LP: 0.0 ± 0.0 % – 20.0 ± 20.0 %; LR: 0.0 ± 0.0 % – 13.1 ± 13.1 %; Supplementary Fig. 2B) across different sampling times.

Fig. 4.

Top biomarkers characterized in (A) chorioallantoic membrane (CAM), (B) cecum, and (C) ileum at the family level using random forest model (Top 10 biomarkers in CAM and ileum, and all characterized biomarkers in cecum). Biomarkers are ranked by descending order of importance.

To further determine the metabolic profiles of CAM microbiota, we performed function prediction analysis using PICRUSt2. LEfSe was applied to identify differentially enriched pathways among different groups. According to the predicted KEGG functional profiles, CAM microbiota in the Control group were significantly enriched for amino acid metabolism, metabolism of cofactors and vitamins, energy metabolism and glycan biosynthesis and metabolism at day 14. Cellular processes, cell motility, bacterial motility proteins, two-component system, secretion system and flagellar assembly were also significantly enriched at day 20 (p ≤ 0.05, LDA>3.5; Fig. 2D). In LP group, CAM microbiota was mostly associated with enrichment in phosphotransferase system and carbohydrate metabolism, including starch and sucrose metabolism, galactose metabolism, and glycolysis/gluconeogenesis throughout day 14 to day 20 (p ≤ 0.05, LDA<3.5). In the LR group, pathways associated with transcription factors, enzyme families, and transcription were enriched at day 14, and energy metabolism were found to be significantly enriched at day 20 (p ≤ 0.05, LDA>3.5). These data indicate that probiotic application to the eggshell significantly modulates the CAM associated bacterial community and thereby its functional profile.

Effect of probiotic application to hatching egg surface on the intestinal microbiota diversity, composition and predicted functional profiles in developing embryos and hatchlings

As seen with the CAM, our data demonstrated that the intestinal microbiota in the developing embryo did not differ in its alpha diversity across different groups and sampling times (p > 0.05; Supplementary Table 4C). Similarly, the indistinct clustering of the intestinal samples demonstrated similarities in the bacteria community structure across the three groups at all sampling times (PERMANOVA p > 0.05; Supplementary Fig. 3A). Further, LEfSe analysis revealed that there were no significantly enriched taxa in intestine samples (p ≤ 0.05, LDA>3.5, one against one strategy), thus indicating that the intestinal microbiota in the embryo was relatively stable during embryogenesis following probiotic application (Supplementary Fig. 3B–D).

Given that our goal in this study is to facilitate early microbiota acquisition in the hatchling, we characterized the hatchling ileum and cecum microbiota. At hatch (day 21), the alpha diversity of the cecal and ileal bacterial community was not significantly different between Control and probiotic groups (p > 0.05; Supplementary Table 4D and E). However, the ileal and cecal community composition in the LP group was significantly different from that of the Control and LR groups (PERMANOVA p ≤ 0.01; Fig. 3A). Looking at the relative abundance of different taxa in the hatchling ileum, Firmicutes (Control: 94.6 ± 4.8 %; LP: 98.4 ± 0.8 %; LR: 96.7 ± 2.5 %), and Proteobacteria (Control: 5.1 ± 4.6 %; LP: 0.9 ± 0.6 %; LR: 2.7 ± 2.2 %) were observed to be the dominant phyla (Supplementary Fig. 4A). At the family level, Enterococcaceae predominates in LP group (Control: 25.1 ± 8.5 %; LP: 89.8 ± 5.3 %; LR: 10.5 ± 6.3 %), while Bacillaceae (Control: 38.8 ± 15.5 %; LP: 2.0 ± 0.7 %; LR: 62.2 ± 15.8 %) and Clostridiaceae (Control: 18.9 ± 10.6 %; LP: 0.5 ± 0.5 %; LR: 20.5 ± 10.8 %) predominate in Control and LR groups (Fig. 3B).

Fig. 3.

Changes in microbial communities of hatchling gut (cecum and ileum) between different groups. (A) Principal coordinate analysis (PCoA) of the bacterial communities of different groups. (B) Stacked bar charts showing the relative abundance of bacterial community (mean value) at the family level. Bacteria families with relative abundance < 1 % are grouped as others. (C) Differentially enriched taxonomic profiles of (C1) cecum and (C2) ileum in descending order (p ≤ 0.05, LDA score>3.5, one against one strategy). (D) Differentially enriched functional profiles of (D1) cecum and (D2) ileum in descending order (p ≤ 0.05, LDA score >3.5, one against one strategy).

Abbreviations: LP, Lacticaseibacillus paracasei DUP13076; LR, Lacticaseibacillus rhamnosus NRRL B-442.

Notably, sequences assigned to Lacticaseibacillus paracasei and Lacticaseibacillus rhamnosus were observed at low abundance in the ileal microbiota samples of the LP and LR groups, respectively. Enterococcus faecium, Enterococcus mundtii (closely matching the ATCC 882 reference), and Lacticaseibacillus paracasei associated sequences were significantly enriched in the LP group, while Bacillus genus was significantly enriched in LR group (LEfSe p ≤ 0.05, LDA>3.5, Fig. 3C). A BLAST search against the GenBank database revealed 100 % sequence identity to reference strains of Lacticaseibacillus paracasei, Lacticaseibacillus rhamnosus, and Enterococcus faecium. Sequences initially identified as Enterococcus mundtii-like were also assigned to Enterococcus faecium, highlighting the need for caution when interpreting E. mundtii-like assignments. Furthermore, a random forest model was applied to predict the group using ileal bacterial families as predictors. Results show that our model could predict the group with high accuracy score, indicating a distinct microbial community at the family level among the three groups (Fig. 4C). Specifically, Staphylococcaceae, Enterococcaceae, Bacillaceae, Lactobacillaece and Clostridiaceae were characterized as the most important biomarkers to differentiate Control and LP group with accuracy score of 0.94. Next, the model could predict Control or LR with accuracy score as 0.83, and Lactobacillaece was characterized as the most important biomarker. Lastly, for the comparison between LP and LR, Bacillaceae, Enterococcaceae, and Staphylococcaceae were determined as top biomarkers with accuracy score as 0.97 (Fig. 4C).

In the case of the hatchling cecal microbiota, similar to the ileum, Firmicutes and Proteobacteria were the predominant phyla in the hatchling ceca (Supplementary Fig. 4A). At the family level, five families represented > 99.9 % of the cecal microbiota. These include Bacillaceae (Control: 39.8 ± 24.2 %; LP: 0.1 ± 0.1 %; LR: 67.2 ± 16.5 %), Clostridiaceae (Control: 43.5 ± 19.9 %; LP: 1.7 ± 1.7 %; LR: 29.2 ± 15.6 %), Enterococcaceae (Control: 16.6 ± 11.1 %; LP: 98.1 ± 1.7 %; LR: 3.5 ± 3.2 %), Lactobacillaceae (Control: 0.02±0.02 %; LP: 0.05±0.05 %; LR: 0.02±0.01 %), and Staphylococcaceae (Control: 0.04±0.02 %; LP: 0.0 ± 0.0 %; LR: 0.0 ± 0.0 %; Fig. 3B). Notably, sequences closely matching Lacticaseibacillus paracasei and Lacticaseibacillus rhamnosus were observed at low abundance in the cecal microbiota samples of the LP and LR groups, respectively.

LEfSe analysis revealed significant enrichment of taxa affiliated with the order Lactobacillales, particularly the family Enterococcaceae and genus Enterococcus, including putative species such as Enterococcus faecium, and Enterococcus mundtii (ATCC 882 reference sequence), in the LP group, while genus Bacillus was significantly enriched in LR group (LEfSe p ≤ 0.05, LDA>3.5; Fig. 3C). Further analysis indicated that LP could be predicted accurately with high accuracy score (Control vs LP: 0.91; LP vs LR: 1), while it shows a low accuracy score when distinguishing LR from Control group (Control vs LR: 0.58; Fig. 4B). Specifically, Enterococcaceae and Bacillaceae were identified as the most important biomarkers. These data demonstrate the feasibility of our approach of employing a non-invasive application of probiotics to the eggshell as a means to modulate hatchling gut microbiota.

Characterization of the metabolic profiles of the hatchling gut microbiota revealed that similar pathways were enriched in both the ileal and cecal microbial community (Fig. 3D). Our results indicate that when compared to the Control chicks, gut microbiota in the LP group were mainly associated with carbohydrate metabolism, genetic information processing, membrane transport, replication and repair, amino acid and nucleotide metabolism according to the second level of KEGG functional profiles (p ≤ 0.05; Fig. 3D). At the third level, significant enrichment was noticed with galactose metabolism, ABC transporters, amino sugar and nucleotide sugar metabolism, fructose and mannose metabolism, and transporters pathways. With the LR group, pathways associated with metabolism of cofactors and vitamins, energy metabolism, cell growth, amino acid metabolism, and arginine and proline metabolism were found to be significantly enriched at level 2 and 3 of the KEGG functional profiles, respectively (p ≤ 0.05).

Potential sources contributing to the hatchling gut microbiota

To predict microbiota acquisition in the hatchling gut, we performed bacterial source tracking using SourceTracker. Our analysis revealed that when compared to the eggshell, CAM and embryonic intestine, the CAM contributed the most to the hatchling gut (ileum and ceca) bacterial community (Fig. 5). Further analysis revealed differences in source attribution in the ileal and cecal microbiota and is discussed below for each group.

Fig. 5.

Relative source proportions contributing to the hatchling gut microbiota using source tracking. The height of the flow between bars demonstrates the average proportions of predicted sources contributing to the cecum and ileum microbiota in each group.

Abbreviations: LP, Lacticaseibacillus paracasei DUP13076; LR, Lacticaseibacillus rhamnosus NRRL B-442.

In the Control group, predictive analysis revealed that 49 % of the cecal microbiota originated from the CAM at day 18, followed by 1.4 % from the eggshell at day 14 and 0.9 % from CAM at day 7. Similarly, 3.1 % of the ileum microbiota was predicted to have originated from the CAM at day 18, 0.3 % from eggshell at day 14, and 0.2 % from eggshell at day 0. With the hatchlings in the LP group, SourceTracker attributed 79.6 % of the bacteria in the cecal microbial community to the CAM at day 7, and 1.1 % from CAM at day 18, and 0.1 % from CAM at day 20. Similarly, 65.3 % of the bacterial community in the ileum microbial community was predicted to originate from the CAM at day 7, followed by CAM at day 20 (2.7 %), CAM at day 18 (0.6 %), and eggshell at day 14 (0.6 %), day 18 (0.6 %), and day 20 (0.6 %). It is worth noting that sequences closely related to Lacticaseibacillus paracasei were found to be unique to the ileal microbial community, with potential origins attributed to eggshell and CAM at all time points, and embryonic intestine at day 18 and day 20. In the LR group, cecal microbial community was predominately attributed to CAM at day 14 (16.9 %), day 7 (5.3 %), and day 20 (1.1 %); Similar results were observed with the ileal microbiota with major contributions from the CAM at day 14 (36.8 %), day 7 (11.2 %), and day 20 (4.2 %). Notably, sequences assigned to Lacticaseibacillus rhamnosus were unique to the ileal bacterial community and was attributed to the eggshell and CAM at all time points (except for shell at day 0). This further validates that modulation of eggshell microbiota can be a viable approach to support hatchling gut microbiota acquisition.

Discussion

The chicken gut microbiota is complex and plastic with well-recognized roles in digestion, nutrient absorption, immune system development and function, health and performance (Shang et al., 2018). Research suggests that the embryonic microbiota originates from the maternal cloaca, oviduct, or both (Lee et al., 2019). Further, gut microbiota inheritance in the chicken embryo is also reported to be influenced by the environment (Ding et al., 2017).More importantly, studies show significant translocation of the eggshell, yolk, and embryonic intestinal microbiota into the hatchling gut, thereby seeding the naïve intestine and playing a key role in the initial colonization of the naïve gut (Akinyemi et al., 2020; Maki et al., 2020; Ding et al., 2021). Moreover, vertical transmission of microbiota to commercial chicks grown separately from the adults is quite limited possibly resulting in impairment of microbiota function (Shterzer et al., 2023). Given the need to establish a healthy microbiota in the hatchling and the limitations associated with current commercial rearing practices, there is a need to identify effective strategies to support healthy microbiota development in chicks. In this regard, it has been noted that the early modulation is an ideal route for establishing a balanced gut microbiota in chicks (Roto et al., 2016). Hence, in this study we employed a non-invasive, in ovo spray application of probiotics on hatching eggs to modulate the embryonic and hatchling microbiota.

In the case of the eggshell microbiota, as expected, spray application of probiotics significantly modified the shell microbial community with a sustained predominance of Lactobacillaceae throughout incubation. As seen with the eggshell, probiotic spray application to the eggs significantly modulated the CAM microbiota. Specifically, Lactobacillaceae family was significantly more enriched in the LP group throughout the embryonic development period in CAM (Fig. 2B, C). Furthermore, Lactobacillaceae was the most important bacterial family in the CAM microbiota when differentiating the three groups using a random forest model. The CAM is a highly vascularized structure that serves multiple roles during embryonic development (Gabrielli and Accili, 2010). Specifically, the CAM mediates mineral ion transportation and metabolism, gaseous exchange, acid-base balance, innate immunity, and nutrient absorption from the allantoic fluid, functions resembling that of the mammalian placenta (Halgrain et al., 2023). Further, the CAM also serves as a site for bacterial translocation into the developing embryo (Ahmed et al., 2022). In line with this, a recent study on embryonic yolk microbiota showed that microorganisms present on the eggshell can enter fertile eggs through the eggshell and the CAM (Ding et al., 2024). Although researchers have not focused on the CAM-associated microbiota, synbiotics have been supplemented in ovo via the air cell (Villaluenga et al., 2004; Siwek et al., 2018). In these studies, the supposition was that the CAM would help transfer these supplements from the air cell into the circulatory system and the developing intestine. Hence, we then studied the intestinal microbiota associated with the developing embryo and hatchling.

First, we observed that the microbial composition shifted from being predominantly Proteobacteria (40 – 53 %) and Firmicutes (28–46 %) in the embryonic intestine to mainly Firmicutes (94–96 %) in the hatchling gut (Supplementary Figs. 3B and 4A). Similarly, comparison of intestinal microbial relative abundance between day 19 (D19) and day 21 (D21) demonstrated a decrease in the prevalence of Proteobacteria from 72.78 % to 59.66 %, accompanied by an increase in predominance of Firmicutes from 11.56 % on D19 to 25.37 % on D21 (Ding et al., 2021). Further, as seen with the eggshell and CAM, we also observed an increased predominance of Lactobacillaceae and Lactobacillus in the probiotic groups when compared to the Control (Figs. 1B and 2B; Supplementary Figs. 1B and 2B). This further supports our hypothesis that modifying eggshell microbiota can modulate gut microbiota in the developing embryo. More importantly, application of LP to the hatching egg during incubation resulted in a unique and separate community structure in the hatchling gut when compared to LR and Control (Fig. 3A). Moreover, probiotic application to the egg surface led to an increased relative abundance of sequences assigned to Lacticaseibacillus paracasei and Lacticaseibacillus rhamnosus in the ileum of the LP and LR group by 5.7 % and 0.7 %, respectively, while these were not identified in the Control group. These data clearly demonstrate the ability of probiotics applied to the eggshell to potentially translocate and transplant into the hatchling gut while modulating the microbiota acquisition.

As a method, spray application has been previously applied on eggs to investigate its effects on gut microbiota modulation. Specifically, spray application of adult cecal contents on eggs during incubation has been used to modulate the microbiota establishment (Richards-Rios et al., 2020). Results indicated that the spray application transplanted spore-forming bacteria to chicks immediately post hatching, including Lachnospiraceae and Ruminococcaceae, while failed in transplanting Lactobacillaceae, Bacteroidaceae, Bifidobacteriaceae and Burkholderiaceae, which were considered as the core composition of chicken cecal microbiota. Additionally, spray application of gut anaerobes on eggs/chickens during incubation was not seen to promote colonization of the sprayed gut anaerobes in chicken gut (Volf et al., 2021). However, in our study, application of specific probiotics on the eggshell led to microbiota modulation in the hatchling gut.

In addition to spray application, in ovo inoculations have been used in late-term embryos to modulate gut microbiota development. Wilson and co-authors (Wilson et al., 2019) applied probiotics via the intra-amniotic route on embryonic day 18 and reported an increased abundance of Lactobacillaceae and Lactobacillus and a decreased abundance of Enterococcaceae and Enterococcus in the hatchling gut. Similarly, it has been reported that in ovo injection of Bacillus spp.-based probiotics altered the microbial composition with a decrease of Enterobacteriaceae and an increase of Lachnospiraceae, along with the change in beta diversity of the hatchling gut microbial community structure (Arreguin-Nava et al., 2019). Also, in ovo injection of a probiotic product from adult microbiota to the amniotic fluid at embryonic day 18 promoted the development of intestinal microbiota of broiler chicks while reducing the abundance of undesirable microorganisms (Pedroso et al., 2016).

Likewise in our study, application of LP was seen to significantly enrich Enterococcus mundtii and Enterococcus faecium-like sequences in the hatchling ileum and cecum compared to the Control (Fig. 3C). According to the BLAST search against GenBank database, Enterococcus faecium and Enterococcus mundtii-like sequences were both matched to Enterococcus faecium with 100 % identity, highlighting the importance of prudent evaluation of species-level assignments in 16S rRNA gene sequencing. Enterococcus faecium is a lactic acid-producing bacteria, known to alleviate symptoms of necrotic enteritis, improve mucosal immune responses, increase ileal villus height and crypt depth, decrease feed conversion ratio, improve body weight gain, and reduce Salmonella infections in chickens (Wu et al., 2019; He et al., 2021; Zhang et al., 2021; Khalifa and Ibrahim, 2023). Similarly, Enterococcus mundtii has been shown to exert an antimicrobial effect against Listeria monocytogenes and Staphylococcus aureus in vitro and in cheese (Campos et al., 2006; Vera Pingitore et al., 2012). Also, it was seen to exclude Listeria monocytogenes colonization and ameliorate Staphylococcus aureus-induced mastitis in mice (Van Zyl et al., 2016; Qiu et al., 2022). While species-level assignments from 16S rRNA gene amplicons should be interpreted with caution, our findings – together with other studies (Teague et al., 2017; Wilson et al., 2019; Shehata et al., 2021) – demonstrate that probiotic application is associated with enrichment in microbes with potential probiotic properties in the hatchling gut.

In addition to characterizing the microbiota, we used SourceTracker analysis to determine the contribution of the eggshell, CAM, and intestinal microbiota to the establishment of hatchling gut microbiota. Overall, predictive analysis determined that the CAM served as the predominant source in populating the hatchling gut (Fig. 5). Previous studies have shown that the sources of chicken gut microbiota can be attributed to the eggshell, environment, and yolk (Maki et al., 2020; Ding et al., 2021). Besides, the microbiota of early chick embryos was reported to be partially inherited from mother hens (Ding et al., 2017). Further, it suggests that feather and cloacal microbiota of mother hens were main sources to the gut microbiota of chicks in the first week of hatch when hatchlings are reared with adult birds (Li et al., 2022). To the best of our knowledge, this is the first report demonstrating an appreciable role for the CAM microbiota in the establishment of hatchling gut microbiota, highlighting its potential for supporting microbial translocation into the embryo and the hatchling gut (Adam et al., 2002). Nevertheless, we acknowledge that the SourceTracker analysis is predictive, and the link between CAM microbiota and gut microbiota merits further investigation.

Beyond contributing to the gut microbiota, we also observed that probiotic application enriched CAM microbiota involved with carbohydrate metabolism, including glycolysis/gluconeogenesis, galactose metabolism, energy metabolism, starch and sucrose metabolism when compared to the Control (Fig. 2D). Carbohydrate and energy metabolism plays a critical role in embryonic development and the hatching process. In effect, glycogen reserves serve as the primary fuel for hatching and the immediate pot-hatch sustenance (Uni and Ferket, 2004; Givisiez et al., 2020). In this regard, the enrichment in microbiota-associated with carbohydrate and energy metabolism in the CAM might contribute to the overall energy status of the developing embryo. This might support our previous finding that in-ovo spray application of LP significantly improved hatchability and hatchling weight in LP group by 5.3 % and 3.6 % respectively, when compared to the Control group (Gao et al., 2024).

As seen with the CAM, several pathways related to metabolism of cofactors, energy, carbohydrates, nucleotides, membrane transport, translation, replication, and repair and transporters were significantly enriched in the gut microbiota associated with the probiotic groups (Fig. 3D). The nutrients in the eggs are limited, and there is no nutrient supplement given to the growing embryos during incubation. The improvement in functional capabilities in the hatchling gut could be presumed to be related to perinatal development and potential post-hatch growth (Akinyemi et al., 2020). In addition, pathways related to cellular processing and signaling were enriched in the hatchling gut microbiota. Cellular processing and signaling pathways play a role in activation of antigen presenting cells including the macrophages and dendritic cells. These cells in turn activate the T cells critical to fighting infections thereby protecting the perinatal embryo and hatchling (Akinyemi et al., 2020). Further, this enrichment in immune related functions could be associated with the significant enrichment in Lactobacillus populations in the hatchling gut. In fact, various species of Lactobacillus had been studied for their ability to activate T cells in chickens resulting in differential cytokine production critical to the regulation and maintenance of homeostasis in the gut (Brisbin et al., 2012; Parker et al., 2018). We also observed that pathways associated with genetic information processing were enriched in the hatching gut microbiota (Fig. 3D). This is in line with previous findings suggesting that early development relies on translation and utilization of stored mRNA for later development (Curtis et al., 1995; Akinyemi et al., 2020). Overall, these functioning pathways are critical nutritional requirements not only for the chick but also for the developing embryo. Hence, absence of microbiota with these functional profiles might impede growth and development (Akinyemi et al., 2020). On the other hand, enrichment in microbiota associated with these pathways as seen with the probiotic groups could support embryo growth, hatchling quality and post-hatch growth. Towards this, our previous data demonstrate significantly improved embryo development (increased embryo weight and breast weight), hatchability, hatchling quality and post-hatch growth following in ovo probiotic supplementation to the hatching egg (Gao et al., 2024).

Given the complexity of the microbiota, species- and even strain-level taxonomic assignments are recommended to better understand their impact on broiler performance (Liu et al., 2021). However, species-level classification based on 16S rRNA gene sequencing should be interpreted with caution due to inherent limitations in resolution. This challenge is further compounded by variability across different reference databases (Campos et al., 2022). Additionally, this study did not assess the colonization efficiency of the administered probiotics, which warrants further investigation using qPCR targeting probiotic-specific gene markers. Furthermore, sequencing was conducted on total DNA, which does not distinguish between viable and non-viable cells; thus, future studies incorporating propidium monoazide (PMA) treatment are necessary to accurately characterize the viable microbiota. Finally, although our findings demonstrated microbiota modulation alongside improvements in hatchability and hatchling quality, further research is needed to optimize probiotic dosage and application timing for effective use in commercial hatcheries (Gao et al., 2024).

Conclusion

Overall, our study demonstrates the feasibility of using in ovo spray application of probiotics on hatching eggs to modulate hatchling gut microbiota. Specifically, probiotics significantly modulated the composition of the microbiota associated with the developing embryo (eggshell, CAM, and intestine) and the hatchling gut. Further, we identified a significant role for the CAM in seeding the hatchling gut and promoting early microbial modulation. Besides, in addition to enriching Lactobacillus populations in the hatchling gut, LP and LR promoted enrichment in other potential probiotic populations while decreasing pathogen populations when compared to the Control. In addition, the hatchling gut microbiota in the probiotic groups was significantly enriched for nutrient and energy metabolism, which could support improved embryonic development, hatchling quality, energy status and post-hatch growth in broilers.

Data availability

The raw 16S rRNA sequencing reads are available at NCBI under BioProject accession number PRJNA1270401.

Supplementary data

Supplementary Table 1. Sample metadata

Supplementary Table 2. Amplicon sequence variant (ASV) count table with taxonomic annotation.

Supplementary Table 3 Relative abundance tables at different taxonomic levels: (A) phylum, (B) family, and (C) genus.

Supplementary Table 4. Effect of probiotic spray application to hatching eggs on the alpha diversity of microbiota associated with (A) eggshell, (B) chorioallantoic membrane, (C) embryonic intestine, (D) hatchling cecum and (E) hatchling ileum throughout the incubation

Supplementary Fig. 1. Effect of probiotic spray application to hatching eggs on shell surface microbiota composition. Stacked bar charts showing the relative abundance of eggshell associated microbial communities between different groups through incubation (mean value) at the (A) phylum and (B) genus level. Bacteria genus with relative abundance < 1 % are grouped as others.

Supplementary Fig. 2. Effect of probiotic spray application to hatching eggs on CAM microbiota composition. Stacked bar charts show the relative abundance of CAM associated microbial communities between different groups through incubation (mean value) at the (A) phylum and (B) genus level. Bacteria genus with relative abundance < 1 % are grouped as others.

Supplementary Fig. 3. Effect of probiotic spray application to hatching eggs on embryonic intestinal microbiota composition. (A) Principal coordinate analysis (PCoA) of the bacterial communities of different groups at different time points. (B-D) Stacked bar charts show the relative abundance of bacterial community (mean value) (B) at the phylum, (C) family, and (D) genus levels. Bacteria families and genus with relative abundance < 1 % are grouped as others.

Supplementary Fig. 4. Effect of probiotic spray application to hatching eggs on hatchling gut (ileum and cecum) microbiota composition. Stacked bar charts show the relative abundance of bacterial community (mean value) at the (A) phylum, (B) genus, and (C) species level. Bacteria genus and species with relative abundance < 1 % are grouped as others.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Mary Anne Amalaradjou has patent #US11497197B2 issued to University of Connecticut. Corresponding author serves as an associate editor for Poultry Science