Transcriptome-wide association study identified candidate genes associated with gut microbiota

Abstract

Background Gut microbiota is closely associated with host health and disease occurrence. Host genetic factor plays an important role in shaping gut microbial communities. The specific mechanism of host-regulated gene expression affecting gut microbiota has not been elucidated yet. Here we conducted a transcriptome-wide association study (TWAS) for gut microbiota by leveraging expression imputation from large-scale GWAS data sets. Results TWAS detected multiple tissue-specific candidate genes for gut microbiota, such as FUT2 for genus Bifidobacterium in transverse colon (PPERM.ANL = 1.68 × 10–3) and SFTPD for an unclassified genus of Proteobacteria in transverse colon (PPERM.ANL = 5.69 × 10–3). Fine mapping replicated 3 candidate genes in TWAS, such as HELLS for Streptococcus (PIP = 0.685) in sigmoid colon, ANO7 for Erysipelotrichaceae (PIP = 0.449) in sigmoid colon. Functional analyses detected 94 significant GO terms and 11 pathways for various taxa in total, such as GO_NUCLEOSIDE_DIPHOSPHATASE_ACTIVITY for Butyrivibrio (FDR P = 1.30 × 10–4), KEGG_RENIN_ANGIOTENSIN_SYSTEM for Anaerostipes (FDR P = 3.16 × 10–2). Literature search results showed 12 genes prioritized by TWAS were associated with 12 diseases. For instance, SFTPD for an unclassified genus of Proteobacteria was related to atherosclerosis, and FUT2 for Bifidobacterium was associated with Crohn’s disease. Conclusions Our study results provided novel insights for understanding the genetic mechanism of gut microbiota, and attempted to provide clues for revealing the influence of genetic factors on gut microbiota for the occurrence and development of diseases.

Gut Microbiome Composition and Metabolic Capacity Differ by FUT2 Secretor Status in Exclusively Breastfed Infants

“…The lack of an association between the infant microbiota and maternal FUT2 genotype at two months of lactation suggests that the oligosaccharides provided by the mother were sufficiently equivalent or were overshadowed by the glycans provided by the gut mucosa of the infant. The finding that infant FUT2 status significantly impacted their microbiome is consistent with other studies on the FUT2 genotype and microbiome, which have typically been conducted in adult populations, including studies on healthy individuals and studies on gut and respiratory diseases [ 21 , 28 , 29 , 36 ]. However, the literature on FUT2 and the microbiome remains controversial, as diverse, large-scale studies have reported differing results [ 22 ].…”

“…Statistical analyses were performed using ASV tables generated through the outputs of Kraken2/Bracken [ 33 , 34 ] or HUMAnN2 [ 35 ] after converting the read counts to relative abundance. Alpha- and beta-diversity were determined using the vegan package, and LEfSe was determined by using the MicrobiomeMarker [ 36 ] package in R. Alpha-diversity was calculated as Shannon diversity and beta-diversity was calculated as Bray–Curtis distance.…”

Transcriptome-wide association studies associated with Crohn’s disease: challenges and perspectives

“…Changes in the abundance of some bacteria have been used as biomarkers to screen for IBD and other gastrointestinal diseases [ 94 – 96 ]. A TWAS conducted in the gut microbiota has detected multiple tissue-specific candidate genes in the sigmoid and transverse colon, respectively, such as TOB2P1 for Enterococcaceae in sigmoid colon, WDR6 for Coprococcus in sigmoid colon, and KCNIP3 for Veillonellaceae in transverse colon [ 97 ]. An association study using bioinformatic analysis in colorectal cancer also observed two overlapping pathways, the bile secretion and steroid hormone biosynthesis pathways, enriched by operational taxonomic units (OTUs) and gene expression patterns in colon tissue, respectively [ 98 ].…”

Transcriptome-wide association analyses identify an association between ARL14EP and polycystic ovary syndrome

“…These analyses offer the benefit of increasing the power to detect associations, due to a 50-fold reduction in multiple testing compared to GWAS [15]. While TWAS have been successful in identifying novel biology underlying diseases, including but not limited to, kidney diseases, rheumatoid arthritis, depression and gut microbiota [16][17][18][19], the complex patterns of linkage disequilibrium (LD) in the genome can complicate the identification of causal genes [20]. Therefore, fine-mapping strategies, which are performed to identify causal genes and variants, are an essential component of TWAS analyses [21].…”