Nanopore sequencing as a novel method of characterising anorexia nervosa risk loci

Anorexia nervosa (AN) is a polygenic, severe metabopsychiatric disorder with poorly understood aetiology. Eight significant loci have been identified by genome-wide association studies (GWAS) and single nucleotide polymorphism (SNP)-based heritability was estimated to be ~ 11-17, yet causal variants remain elusive. It is therefore important to define the full spectrum of genetic variants in the wider regions surrounding these significantly associated loci. The hypothesis we evaluate here is that unrecognised or relatively unexplored variants in these regions exist and are promising targets for future functional analyses. To test this hypothesis, we implemented a novel approach with targeted nanopore sequencing (Oxford Nanopore Technologies) for 200 kb regions centred on each of the eight AN-associated loci in 10 AN case samples. Our bioinformatics pipeline entailed base-calling and alignment with Dorado and minimap2 software, followed by variant calling with four separate tools, Sniffles2, Clair3, Straglr, and NanoVar. We then leveraged publicly available databases to characterise these loci in putative functional context and prioritise a subset of potentially relevant variants.

Targeted nanopore sequencing effectively enriched the target regions (average coverage 14.64x). To test our hypothesis, we curated a list of 20 prioritised variants in non-coding regions, poorly represented in the current human reference genome but that may have functional consequences in AN pathology. Notably, we identified a polymorphic SINE-VNTR-Alu like sub-family D element (SVA-D), intergenic with IP6K2 and PRKAR2A, and a poly-T short tandem repeat (STR) in the 3'UTR of FOXP1.

Our results highlight the potential of targeted nanopore sequencing for characterising poorly resolved or complex variation, which may be initially obscured in risk-associated regions detected by GWAS. Some of the variants identified in this way, such as the polymorphic SVA-D and poly-T STR, could contribute to mechanisms of phenotypic risk, through regulation of several neighbouring genes implicated in AN biology, and affect post-transcriptional processing of FOXP1, respectively. This exploratory investigation was not powered to detect functional effects, however, the variants we observed using this method are poorly represented in the current human reference genome and accompanying databases, and further examination of these may provide new opportunities for improved understanding of genetic risk mechanisms of AN.