Long-read low-pass sequencing detects a validated structural variant in the equine myostatin (MSTN) gene

Abstract

Abstract Objectives Low-pass sequencing trades read depth for throughput, an established strategy for genotyping with short-read sequencing. Whether this tolerance extends to structural variant (SV) detection at the level of a single, biologically validated locus has been comparatively under-explored. We used a haplotype-resolved contig-level assembly of a Thoroughbred horse to evaluate detection of a known 227-bp SINE insertion in the myostatin (MSTN) promoter, the putative equine speed-gene variant, from long-read HiFi data downsampled to matched low-pass depths. Results The assembly confirmed a heterozygous 227-bp insertion at the MSTN locus. In downsampled HiFi reads (1.33x-10.30x), an SV-aware caller detected the insertion at every depth; a conventional small-variant caller detected it at two of four depths, a pattern explained by expected Poisson sampling variance rather than systematic error. These results indicate that long-read low-pass sequencing can support sensitive, accurate, and cost-effective genetic testing for structural variants of veterinary and industry importance in Thoroughbreds.