Sampling and Sequencing
Sampling was conducted as previously described (Lutz et al., 2015). Briefly, birds were collected using a combination of mist-netting and ballistic capture in 2009 and 2011. Blood samples were obtained via brachial venipuncture; oral (buccal) and cloaca samples were collected by swabbing with sterile cotton swabs; liver, spleen, and intestinal tracts were dissected from euthanized animals. All samples were stored in cryogenic vials and flash-frozen in liquid nitrogen until nucleic acid extraction. Samples collected for this project were processed following the Earth Microbiome Project standard processing protocols (Thompson et al., 2017). Briefly, DNA was extracted using 96-well PowerSoil PowerMag DNA extraction plates (Qiagen), which were homogenized using a TissueLyser beadbeater (Qiagen). From eluted DNA, triplicate PCRs using the 515f/806r EMP primers amplified the V4 region of the 16S rRNA gene, pooled amplicons were sequenced on Illumina MiSeq and HiSeq instruments, and data were uploaded to the Qiita web-based microbiome analysis platform for initial processing (made available previously via QIITA ID 11166 by (Song et al., 2020). All sequences were demultiplexed and quality filtered using the Qiime2 pipeline (Bolyen et al., 2019) and forward and reverse reads were joined and processed using Deblur to remove sequencing errors and refine sequences to amplicon sequence variants (ASVs) (Amir et al., 2017). Deblurred ASV tables and sample metadata were further processed using Qiime2 (Bolyen et al., 2019). To assess the presence of artefactual batch effects associated with microbiota diversity, statistical significance of the plate on which the samples were placed on were calculated on each sample type using the adonis2 function (PERMANOVA) from the vegan package using unweighted UniFrac and weighted UniFrac (McArdle & Anderson, 2001); (Oksanen et al., 2018). Only the unweighted UniFrac distances of the intestine samples showed significant (p <0.05) differences due to plate (platename R2= 1.714%, p = 0.05).