Symbiotic microbiota vary with breeding group membership in a highly social joint-nesting bird

Symbiotic microbes affect the health, fitness, and behavior of their animal hosts, and can even affect the behavior of non-hosts. Living in groups presents numerous benefits and challenges to social animals, including exposure to symbiotic microbes, which can mediate both cooperation and competition. In social mammals, individuals from the same social group tend to share more similar microbes, and this social microbiome, the microbial community of all hosts in the same social group, can shape the benefits and costs of group living. In contrast, little is known about the social microbiome of group-living birds. We tested the predictions that communally breeding smooth-billed anis (Crotophaga ani) belonging to the same breeding group share more similar microbes and that microbial community composition differs between body regions. To test this, we used 16S rRNA gene sequencing to characterize the preen gland and body feather microbiota of adult birds from 16 breeding groups at a long-term study site in southwestern Puerto Rico. As predicted, individuals from the same breeding group shared more similar microbiota than non-group members and preen gland and body feathers harboured distinct microbial communities. Future research will evaluate whether this social microbiome affects the behavior of group living birds.R Microsoft ExcelFunding provided by: Natural Sciences and Engineering Research Council of CanadaCrossref Funder Registry ID: http://dx.doi.org/10.13039/501100000038Award Number:See main text for details. Summary from main text: We extracted bacterial DNA from preen gland swabs and feathers using Norgen soil DNA isolation plus kits and amplified the V4 region of the bacterial 16S rRNA gene using the universal primers F518 (Lane et al. 1985) and R806 (Caporaso et al. 2011). We pooled PCR amplified products of the expected band size (approx. 300 nt) into a library and sequenced with 250 nt paired-end reads on an Illumina MiSeq at the London Regional Genomics Centre. We used the R package dada2 (Callahan et al. 2016) to overlap reads, remove ambiguous reads, chimeras, and singleton sequences, and assign reads to samples. Sequences rarer than 0.1% in any sample were removed as they contain little information and removing them has no impact on downstream analyses (Bian et al. 2017). The workflow and parameters used are available at github.com/ggloor/miseq_bin/. After performing these steps, we obtained an initial dataset of 12,532 unique sequences (i.e., amplicon sequence variants; hereafter ASVs) from 103 samples. We then assigned ASVs to taxon by clustering at ≥ 97% sequence identity using the naïve Bayesian Ribosomal Database Project (RDP) Classifier (Wang et al. 2007). We next filtered sequences by the minimum proportion, minimum occurrence, and minimum sample count of reads. Sequences found in less than 0.5% of reads (consistent with MiSeq instrument error rates reported in Stoler and Nekrutenko 2021), fewer than 5% of samples, and samples with fewer than 5000 reads were removed (following Grieves et al. 2021a, b). We used Bayesian-multiplicative replacement to impute values for zero count sequences (following Bian et al. 2017; Quinn et al. 2019) using the R package zCompositions (Palarea-Albaladejo and Martin-Fernandez 2015). We then applied a centered log-ratio transformation to the zero-replaced data set, rendering the use of Euclidean (Aitchison) distances meaningful and straightforward for subsequent analyses (Fernandes et al. 2014; Gloor and Reid 2016)