SVIV

Different structural-variant callers disagree — each has its own sensitivities, so confident calls usually mean reconciling two or three of them, and there’s no standard way to do that merge. SVIV gives that reconciliation a reproducible home in R.

It ingests VCFs from callers like Manta, DELLY, LUMPY, and GRIDSS, then merges region-based variants with flexible intersection and union operators. Translocations get a cleverer treatment: each is projected onto a Cartesian plane by chromosome pair and grouped with K-means clustering, which tames the messy inter-chromosomal calls that chromothripsis throws off.

The merged calls become an SV feature matrix — ready for survival or treatment-response analysis — and render as ggplot2 and karyoploteR figures at whole-genome or single-chromosome resolution.