2.5 | Sequencing and Bioinformatics
DNA from faecal samples were extracted using QIAgen’s QIAamp PowerFecal DNA Kits. The protocol followed Qiagen’s Kit Handbook with homogenization performed at full speed for 10 minutes using the 2 mL bead beating tubes provided with the kit (0.7 mm Dry Garnet) and a Vortexed-Genie 2 fitted with Qiagen’s Vortex Adapter (cat. No. 13000-V1-24). DNA extracts were quantified using the Qubit dsDNA BR Assay Kit and standardized to 20 ng/uL prior to PCR amplification. PCR and sequencing of the v3–v4 region of the 16S rRNA gene using the 341F/805R primers and sequencing on an Illumina MiSeq platform using the v3 chemistry was outsourced to the University of Calgary Centre for Health Genomics and Informatics.
Cutadapt v1.16 was used to remove primers from raw reads, while untrimmed reads were discarded (Martin, 2011). Trimmed reads were processed using dada2 v1.6 (Callahan et al., 2016), with only minor deviations from default parameters. Namely, sequences with a maximum expected error of two or greater, PhiX spike-ins, and bases with a quality score of <2 were discarded using thefilterAndTrim command. Forward and reverse sequences were truncated to lengths of 250 and 200, respectively. Taxonomic assignment of amplicon sequence variants (ASVs) was performed using the dada2 implementation of the naïve Bayesian classifier (Wang et al. 2007) and v132 of the SILVA database (Yilmaz et al., 2014). To further filter sequencing errors and reduce singleton noise prior to analysis, ASVs which were not represented by at least 1 count in 4 samples were removed (Knowles, Eccles, & Baltrūnaitė, 2019). ASV sequences were aligned using MUSCLE with default parameters (Edgar, 2004) and a relaxed neighbour-joining method was used to construct a phylogenetic tree using the mothur implementation of clearcut (Kozich, Westcott, Baxter, Highlander, & Schloss, 2013; Sheneman, Evans, & Foster, 2006).