Structural Variant Prediction in Extended Pedigrees Through Sparse Negative Binomial Genome Signal Recovery

Structural variants (SVs) are rearrangements, such as deletions, insertions, duplications, inversions, and translocations, in an individual's genome relative to a reference. SV detection is often marred by high false positive rates due to errors in sequencing and mapping. In previous work, we proposed a maximum likelihood approach to SV prediction that incorporated low-coverage sequencing data and coverage distribution. In particular, we developed a negative binomial framework to reflect a more realistic representation DNA fragment distributions sampled from an individual's genome. In this paper, we leverage relationships between an off spring and both parents, in addition to the negative binomial framework, to improve SV identification accuracy. We present numerical results on both simulated genomes as well as two sequenced parent-child trios from the 1000 Genomes Project.