Releases: morispi/LEVIATHAN
v1.0.2
v1.0.1
v1.0
Initial release.
LEVIATHAN (Linked-reads based structural variant caller with barcode indexing) is a structural variant calling method for Linked-Reads data. LEVIATHAN takes as input a BAM file, which can either be generated by a Linked-Reads dedicated mapper such as Long Ranger, or by any other aligner, given the reads are pre-processed in order to extract the barcodes from their sequences and append them to the headers (e.g. using Long Ranger basic). LEVIATHAN output SVs in VCF format.
LEVIATHAN works by, first, indexing the occurrence positions of the barcodes throughout the BAM file using LRez. It then relies on two distinct steps. The first one computes the amount of common barcodes between region pairs of the reference genome, in order to highlight SV candidates, which are pairs of distant regions that share more barcodes than expected. The second step then acts a refining step, and relies on classical short-reads methodologies to further analyze these SV candidates, and determine the types and breakpoints of actual SVs, and filter out false-positives.