Bridging Short- and Long-Read Structural Variation Detection with SurVeyor
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Structural variations (SVs) are a major source of genomic diversity and disease, yet accurate detection and genotyping with short-read sequencing remain a longstanding challenge, with existing approaches missing large fractions of variants. We present SurVeyor, a unified, cohort-aware pipeline that significantly advances the state of the art for SV discovery and genotyping from short reads. Using public data, we found SurVeyor to outperform all tested tools, academic and commercial, both for SV discovery and genotyping. Furthermore, SurVeyor can leverage cohort data to yield even more complete callsets. Applied to a cohort of 63 individuals, SurVeyor achieved more than twice the sensitivity of existing short-read pipelines, and almost matched the sensitivity while exceeding the precision of state-of-the-art PacBio HiFi long-read callers. Overall, SurVeyor bridges the performance gap between short and long-reads SV detection and between academic and commercial pipelines, enabling population-scale and resource-limited studies to generate reliable SV catalogues without dependence on costly technologies.
