Multiple Bar-Coding 16S Sequencing by Pacbio RS Platform to Study the Gut Microbiome in Ashkenazi Jews With Crohn’s Disease

Abstract

BACKGROUND: Crohn's disease (CD) results from defects in the mucosal immune response to luminal factors in genetically susceptible individuals. The role of the gut microbiome in CD pathogenesis has been suggested by several studies. Until recently, 454 pyrosequencing approach was considered the gold standard for (Figure Presented) microbiome studies. However, newer, more efficient, and cost-effective technologies are now available (1). Compared to other next-generation platforms, Pacbio RS has the advantage of a much shorter sequencing time (∼1 hr) and long read lengths (up to 8kb). However, the embedded, relatively high random sequencing error rate is considered a limiting factor for applying this technology to 16Sbased taxonomy assignment and phylogenetic analysis. The aim of this study was to evaluate the performance of Pacbio RS on 16S sequencing of the fecal microbiome of Ashkenazi Jews (AJ, a genetically homogeneous high-risk population) with and without Crohn's disease. METHODS: Stool from 7 AJ-CD patients (in remission) and 8 AJ healthy controls were analyzed. We established a protocol for multiple 16S rRNA sequencing using the Pacbio RS system and performed taxon-based and phylogenetic analysis. Total DNA was extracted from the fecal samples and PCR amplified with unique 16 bp bar-coded primer sets targeting the V3-to-V4 hypervariable regions. PCR amplicons were pooled in equal molar amounts and sent for library preparation and sequencing. Control E. coli BL21 strain was used to test the overall classification rates. RESULTS: A single run at the movie length of 2x45 minutes generated ∼55k reads with ∼33k reads being longer than 2kb (Figure1a). We used the QIIME pipeline (2) to assign the 16S rRNA CCS (circular consensus sequencing) reads to the taxonomy at phylum, class and family level (Figure1b). Repeated measurements of 2 subjects (P3; P5) showed high correlations at phylum, class and family taxa levels (r2>0.99). The classification rates for the control E. coli were ∼99.5% at three taxa levels, respectively. We also used the QIIME pipeline to compare the bacterial composition in CD patients and unaffected controls. The microbiome diversity shown by alpha diversity rarefaction plots (Figure2a) was not significantly different between CD and controls. However, the unweighted unifrac PCoA analysis (Figure2b) clustered microbiota by the disease status. The two pairs of repeated measurements were closely clustered with each other, as expected. Comparing each taxon, Clostridia (class level), Ruminococcaceae (family level) and Verrucomicrobiaceae (family level) were significantly less abundant in CD than controls, consistent with previous studies (3, 4). CONCLUSION(S): Our data suggest that sequencing error introduced by the Pacbio platform is unlikely to affect the classification of common bacterial species in the gut. Using the 16S deep sequencing on the Pacbio RS platform, we reproduced previously observed differences in microbial composition between CD and controls, suggesting it as a promising method to conduct microbiome studies.