DNA Extraction, PCR, and Sequencing
We extracted microbial DNA from perianal swabs using a modified Qiagen DNeasy PowerSoil HTP 96 Kit protocol (Qiagen Inc., Maryland, USA). We isolated swab tips using scissors sterilized by dunking blades in 80% ethanol solution before passing them through a flame. We randomized samples from different islands and species across PowerBead plates to minimize potential batch effects, and disrupted all samples and controls using a MiniBeadBeater-96. We shook plates for 3min, allowed 30sec of rest, rotated plates, and shook plates for an additional 3min. We then followed the standard manufacturer protocol before eluting samples for 15min at room temperature in 60μL solution C6 pre-heated to 70ºC.
We used polymerase chain reaction (PCR) to amplify and tag the 16S rRNA hypervariable 4 (V4) region with the barcoded forward (GTGCCAGCMGCCGCGGTAA) and reverse (TAATCTWTGGGVHCATCAGG) primers described in . PCR reactions contained 5μL HiFi HotStart ReadyMix (KAPA Biosystems, Massachusetts, USA), 3.2μL primer mix (1.25μM), and 2.0μL template DNA. Extraction controls included sterile swab tips (negative controls) and DNA extracted from ZymoBIOMICS Microbial Community Standard (D6300; positive controls); and PCR controls included PCR reagents with no template DNA added (negative controls) and 1μL of ZymoBIOMICS Microbial Community DNA Standard (D6305; positive controls; Zymo Research, California, USA). PCR cycling conditions included: initial denaturation of 94ºC/3min, touchdown cycling for 30 cycles of [94ºC/45s, 80-50ºC/60s, 72ºC/90s] decreasing 1ºC each cycle, 12 cycles of [94ºC/45s, 50ºC/60s, 72ºC/90s], and a final extension of 72ºC/10min. To maximize yield and minimize PCR artifacts, each sample was run through PCR in duplicate and subsequently combined for quantification using a Qubit Flex Fluorometer (Invitrogen, Massachusetts, USA). We then pooled 40ng of each barcoded library, used the QIAquick PCR Purification kit (Qiagen Inc., Maryland, USA) and E-Gel 2% Size Selection Protocol (Invitrogen, Massachusetts, USA) to clean and size select libraries, and subsequently used Agencourt AMPure XP magnetic beads (Beckman Coulter, California, USA) to concentrate final libraries using a 1.25X bead to library ratio. Libraries were paired-end sequenced (2x150nt) across three sequencing plates on an Illumina MiSeq in the Princeton University Lewis Sigler Genomics Core Facility.