Day zero birds from both hatcheries exhibited distinct microbiomes visibly different from one another, with hatchery A having an overwhelming percentage presence of Escherichia coli at day zero while hatchery B exhibited a more diffuse spread of common microbes. An inquiry to both hatcheries confirmed that neither practiced any sort of post-hatch probiotics program at the time of the experiment. These differences continue through all days with the mixed population birds retaining midgut microbial population elements of both source hatcheries for both floor and battery raised birds. At the conclusion of this study, the midgut microbiome of day 48 floor raised birds more closely resembled the composition of a Hatchery B, while the cecal microbiome was virtually identical in top taxa with minor variations outside the top three most common genera.
Succession in the Midgut Microbiota
Though Lactobacillus was the dominant genus for many sample days midgut samples seemed more prone to population disruption after feed changes with Lactobacillus percentage dropping greatly in the time period between day 14 and 16, and again two weeks after the transition to finisher at day 48 (Table 2).
Alpha and Beta Diversity
Midgut samples averaged a PD score of 5.67 (high 8.86, low 3.27) at maximum sampling depth (24,000), however only 6 of 26 samples reached that depth compared to the ceca samples of which 20 of 26 reached max depth. Because of the small number of samples at max depth a subsection of the PD scoring was taken at the depth of 2667 sequences to get a better picture of overall PD score, this sample depth yielded an average of 3.52 (high 8.7, low 2.27).
Midgut samples were significantly less diverse than cecal samples with the Kruskal-Wallis (pairwise) significance of PD (p = 1.447x10− 4). Midgut samples presented a much lower average number of observed OTUs with the average midgut sample OTU count being 64.33 compared to the average cecal sample OTU count being 259.8 at the max sequencing depth of 24,000.
Midgut samples exhibited much lower and more variable evenness compared to ceca samples, with midgut samples displaying an average evenness score of 0.686 (high of 0.91, low 0.49) vs ceca samples at an average of 0.868. A Kruskal-Wallis (pairwise) significance test was statistically significant with p = 1.30x10− 5. For midgut samples the top five genera varied greatly with the only real constant top genus being Lactobacillus and for a brief time later in grow out, Clostridioides.
After initial closeness on day zero, midgut samples plotted on a Jaccard coefficient PCOA (Fig. 6 (2.6–2.12)) trended towards differentiation. Unlike cecal samples which experienced a period of differentiation followed by a return to a similar population uniformity; midgut samples continue the trend of differentiation through day 48, spreading further and further apart.
Shannon diversity score was moderate over samples at max sampling depth, the average being 3.22 (high 6.36, low 1.93). At retention sampling depth Shannon diversity was 3.55 (high 6.32, low 1.93).
Trends in Source Hatchery Microbiota
Source hatchery samples begin at day 0 with distinct but low diversity populations. As previously mentioned, HB was contacted to confirm no probiotic treatment was applied post hatch. No history of parent stock was disclosed; however it is noted that HA utilizes a plant-based all vegetarian diet for its birds. HA was less diverse than HB at day zero and as the trial progressed seemed to display a greater susceptibility to more extreme population shifts, especially in battery cage systems. HB also displayed a notably higher level of prevalence of unidentified bacterial reads (~ 44% ceca, ~ 26% midgut) versus HA (~ 2% ceca, ~ 0.3% midgut). Mixed population birds displayed elements from both source hatcheries, with cecal communities more often mirroring HA in composition and mid-gut communities more closely resembling HB.
HA samples averaged a PD score of 8.85 (high 16.59, low 3.84) at maximum sampling depth (24,000); however, only 7 of 18 samples reached that depth. HB samples averaged a PD score of 9.52 (high 13.49, low 3.27) at maximum sampling depth, with 9 of 17 samples reaching max depth. MP samples averaged a PD score of 9.97 (high 15.03, low 3.84) at maximum sampling depth, with 11 of 17 samples reaching max depth. Because of the small number of samples at max depth a subsection of the PD scoring was taken at the depth of 2667 sequences to get a better picture of overall PD score at a shallower point, this sample depth yielded an average of 9.75 (high 15.23, low 3.78) for HA, an average of 7.31 (high 11.46, low 2.65) for HB, and an average of 7.82 (high 14.26, low 2.86) for MP. Samples were not significantly diverse from one another when analyzed using Kruskal-Wallis (pairwise) significance test (p > 0.05).
HA samples presented an average 189 of observed OTUs, HB samples averaged an OTU count 222, and MP samples averaged an OTU count of 226 at the max sequencing depth of 24,000.
Samples did not display significantly different evenness when compared with Kruskal-Wallis (pairwise) significance test (p > 0.05).
For HA midgut samples top genera varied greatly between time points with top genus being Lactobacillus for 4 out of 9 points sampled (D5B, D5F, D14F, D21F. B = battery, F = floor, none = no pen) and a different top genus for the remaining 5 ( D0 Escherichia-Shigella, D14B family Lachnospiraceae designation Torques group Ruminococcus, D16F Clostridioides, D28F Lachnospiraceae, D48F Clostridiaceae). HA ceca samples varied to some degree as well with Lachnospiraceae being the top genera for 5 of the 9 sample points (D5B, D5F, D14B, D16F, D48F) and a different top genus for the remaining 4 (D0 Escherichia-Shigella, D14F Ruminococcaceae UCG-014, D21F Lactobacillus, D28F Faecalibacterium).
Similar to HA, HB midgut top five genera varied greatly between time points with top genus being Lactobacillus for 4 out of 8 points sampled (D5B, D14B, D28F, D48F) and a different top genus for the remaining 4 ( D0 unidentified bacteria, D14F Candidatus Arthromitus, D16F/D21F Clostridioides). HB ceca samples varied greatly with family Lachnospiraceae designation Torques group Ruminococcus (D5B, D14B, D16F), Lachnospiraceae (D5F, D21F), generic Ruminococcaceae (D14F, D48F), Ruminococcaceae UCG-014 (D28F), and unidentified bacteria (D0) filling the top positions.
MP midgut samples top genera varied somewhat with Lactobacillus being most prevalent for 5 out of 8 points sampled (D5B, D5F, D14BF, D14F, D28F), Clostridioides for two sample days (D16F,D21F), and Lachnospiraceae for D48F. MP ceca samples were also similarly stratified with Lachnospiraceae being the top genera for 5 of the 8 sample points (D5B, D5F, D16F, D21F, D48F), Faecalibacterium for 2 sample points (D14F, D28F), and generic Ruminococcaceae on D14B.
A few occurrences of individual genera were noted, first that Escherichia-Shigella dominated most of D0 for HA samples (ceca ~ 84.25, midgut ~ 88.01%), being present but not overwhelming in HB (ceca ~ 7.57, midgut ~ 11.55) on the same day. This trend continued at day 5 with HA diversifying but still mostly displaying a high proportion of Escherichia-Shigella (Battery- ceca ~ 15.09%, midgut ~ 1.97%. Floor- ceca ~ 0.91%, midgut ~ 0.07%.). Though Escherichia-Shigella is present in the cecal samples of other treatments it does not make up the same proportion of the population as it does in HA. Furthermore, population ratios vary but as a whole Escherichia-Shigella is present in all but one of the HA midgut samples. Escherichia-Shigella is present in all but one sample of HB as well however the proportion is smaller. In MP samples, Escherichia-Shigella is not detected in samples past day 14.
A second notable occurrence related to source hatchery is a pattern of Bacteroides occurring in HB cecal samples early on in the trial (D5B, D14B, D14F, D16F) as well as MP cecal sample (D14B, D14F, D16F), with the genus only being detected in HA at day 21 or later.
HA Shannon diversity score was moderate over samples at max sampling depth, the average being 5.66 (high 8.06, low 1.93). At retention sampling depth Shannon diversity was 5.63 (high 8.05, low 2.19). HB Shannon diversity score was moderate over samples at max sampling depth, the average being 6.41 (high 7.76, low 2.74). At retention sampling depth Shannon diversity was 5.24 (high 7.72, low 1.94). MP Shannon diversity score was moderate over samples at max sampling depth, the average being 6.14 (high 7.68, low 2.59). At retention sampling depth Shannon diversity was 5.10 (high 7.65, low 2.17).
Trends in Pen Types
Birds were only kept in battery to day 14 due to the difficulty of maintaining larger birds in the available battery cages designed to be used as brooders. Floor birds were placed on fresh pine shavings to limit outside influence.
Battery samples averaged a PD score of 8.32 (high 9.97, low 5.37) at the max sequencing depth of 24,000. Floor samples averaged a PD score of 10.63 (high 16.59, low 3.27). Samples were not significantly diverse from one another when analyzed using Kruskal-Wallis (pairwise) significance test.
Battery samples presented an average 152 of observed OTUs (high 232, low 50) at the max sequencing depth of 24,000. Floor samples averaged 250 observed OTUs (high 431, low 27).
Samples did not display significantly different evenness when compared with Kruskal-Wallis (pairwise) significance test (p > 0.05).
For battery midgut samples, top genera was mostly consistent between time points with top genus being Lactobacillus for all sample points save for HA D14 which sported family Lachnospiraceae designation Torques group Ruminococcus as its top genus. For battery ceca samples top genera was Lachnospiraceae (HAD5, MPD5, HAD14), sported family Lachnospiraceae designation Torques group Ruminococcus (HBD5, MPD14), and generic Ruminococcaceae for HB day 14.
For floor midgut samples 8 of the 17 samples retained were characterized with Lactobacillus (HAD5, MPD5, HAD14, MPD14, HAD21, HBD28, MPD28, HBD48), 5 of the 17 were characterized by Clostridioides (HAD16, HBD16, MPD16, HBD21, MPD21), 2 of the 17 were characterized by Lachnospiraceae (HAD28, MPD48), day HB14 was characterized by Candidatus Arthromitus, and day HA48 was characterized by group 1 Clostridiaceae. Day HB5 was omitted from the analysis due to very poor read counts.
Floor ceca samples were stratified similarly to floor midgut with 9 out of 18 samples with Lachnospiraceae (HAD5, HBD5, MPD5, HAD16, MPD16, HBD21, MPD21, HAD48, MPD48) holding the top genus spot, 3 out of 18 were characterized by Faecalibacterium (MPD14, HAD28, MPD28), 2 out of 18 were characterized by generic Ruminococcaceae (HBD14, HBD48), 2 out of 18 were characterized by Ruminococcaceae UCG-014 (HAD14, HBD28), HA D21was characterized by Lactobacillus, and HB D16 was characterized by family Lachnospiraceae designation Torques group Ruminococcus.
Notably Candidatus Arthromitus genus occurs on day 14 in all floor midgut samples (HA ~ 10%, HB ~ 47%, MP ~ 17%). It is only detected on day 14 with no trace observed before or after in any other samples.
Battery Shannon diversity score was moderate over samples at max sampling depth, the average being 5.83 (high 6.82, low 3.10). Floor Shannon diversity score was also moderate over samples at max sampling depth, the average being 6.44 (high 8.06, low 2.59).