Manta calls structural variants (SVs) from mapped paired-end sequencing reads. Manta discovers candidate SVs from discordant pair and split-read alignments, followed by local assembly and realignment to refine candidates.

The app uses Manta on RNA-seq data to detect gene fusions in combination with STAR, which appear like translocations in the RNA alignments. The Manta workflow is followed by RNA-specific filtering and scoring, which is based on the following:

Read counts across the fusion and alignment qualities.
Genome-wide realignment of fusion contigs to filter candidates that can be explained by a local alignment elsewhere in the genome.
Length of coverage around the breakpoints, indicating presence of stable fusion transcripts.

For more information, see