A cystic fibrosis lung disease modifier locus harbors tandem repeats associated with gene expression

Abstract

Variable number of tandem repeats (VNTRs) are major source of genetic variation in human. However due to their repetitive nature and large size, it is challenging to genotype them by short-read sequencing. Therefore, there is limited understanding of how they contribute to complex traits such as cystic fibrosis (CF) lung function. Genome-wide association study (GWAS) of CF lung disease identified two independent signals near SLC9A3 displaying a high density of VNTRs and CpG islands. Here, we used long-read (PacBio) phased sequence (N=58) to identify the boundaries and lengths of 49 common (frequency >2%) VNTRs in the region. Subsequently, associations of the VNTRs with gene expression were investigated in CF nasal epithelia using RNA sequencing (N=46). Two VNTRs tagged by the two GWAS signals and overlapping CpG islands were independently associated with SLC9A3 expression in CF nasal epithelia. The two VNTRs together explained 24% of SLC9A3 gene expression variation. One of them was also associated with TPPP expression. We then showed that the VNTR lengths can be estimated with good accuracy in short-read sequence in a subset of individuals with data on both long (PacBio) and short-read (10X Genomics) technologies (N=52). VNTR lengths were then estimated in the Genotype-Tissue Expression project (GTEx) and their association with gene expression was investigated. Both VNTRs were associated with SLC9A3 expression in multiple non-CF GTEx tissues including lung. The results confirm that VNTRs can explain substantial variation in gene expression and be responsible for GWAS signals, and highlight the critical role of long-read sequencing.