Abstract
Background Microbiota-host genome wide association studies (mbGWAS) have identified hundreds of genetic variants associated with gut microbiota. For instance, mutations in the LCT gene associate with higher abundance of Bifidobacterium, an interaction mediated by lactose digestibility. While there are a 12 gut mbGWAS, mbGWAS from other body sites are lacking.
Methods We performed a mbGWAS for 82 salivary microbiota genera in 1,925 adults from the Cooperative Health Research in South Tyrol (CHRIS) study cohort. Participants provided blood and saliva samples and completed questionnaires regarding smoking habits, socio-demographics and medical history. We extracted salivary DNA, amplified the 16S-V4 region and processed reads with DADA2. We assigned taxonomic ranks with the Human Oral Microbiome Database. We imputed Chip-genotyped CHRIS individuals from the Haplotype Reference Consortium Panel (HRC). After quality control 1,925 samples and 7,570,440 variants were available for analysis. We applied rank-inverse normal transformation to genera's relative abundances and performed a mbGWAS with “Regenie”. To adjust for population complexity, we corrected for age, sex, diabetes, smoking and the first 5 genetic principal components.
Results Among the 45 genera with prevalence above 50% we found 11 genome-wide significant (P < 5×10e-8) genus-variant associations with 2 distinct genera. Lachnospiraceae Genus 2 (prevalence = 57%) was associated with 3 variants with Minor Allele Frequency (MAF) = 0.08 in the ATP8A1 gene. Neisseria (prevalence = 97%) was associated with 8 variants with MAF = 0.47 in the FAM174A gene. In each locus, variants were in linkage disequilibrium (r2 > 0.6). ATP8A1 is a responsible for cellular uptake of ions and heavy metals, whereas FAM174A is a predicted integral membrane protein.
Conclusions We performed the second salivary mbGWAS to date, the first one with high-quality imputed genotype data on a well-characterized Alpine cohort of European ancestry and homogeneously processed microbiota data. We reported 11 genome-wide significant genus-variant associations which could underpin a potential interaction between microbial and host physiology in saliva. In ongoing analyses, we are performing the mbGWAS at the amplicon sequence variant level and exploring associations between less prevalent genus-variant pairs.