Although de novo assembly-based fusion detection methods are more computationally intensive than those based on read alignments to a reference genome, runtimes are practical, and de novo assembly supports characterizing detected events by generating long contig sequences that contain the exact breakpoint and its sequence context.
Considerably complicating incorporation of genome-wide base-by-base mappability into analysis of NGS reads is the fact that such maps must ideally be specifically tailored based on read length, alignment algorithm and alignment parameters, since these parameters will all influence whether a read will be considered mappable.
After assembly, the mean density of SNPs across the four parental haplotypes in assembled regions was estimated based on read re-alignments to be 7.6 per thousand bases.
PCR duplicates were identified based on read start and end alignment coordinates.
JAFFA also provides a simple and effective method of ranking fusions based on read support and exon-exon boundary alignment.
During the alignment, the "-m" setting for mismatches was set at default to allow GSNAP to auto set the number of allowed mismatches based on read length, which allowed for the best alignment across intron-exon boundaries GSNAP "soft-trims" reads during the alignment process, so that only portions of the reads above the quality threshold are aligned.
Intron evidence was collected based on reads with a spliced alignment against the genome and evidence for transcription is taken from RNA-Seq covered regions.
85% of loci expressed at a level of at least 1-fold coverage showed evidence for utilization of unannotated exons, based on excess genomic read alignments.
Our method identifies indel candidates using a discriminative classifier based on features of split read alignment profiles and trained on true and false indel candidates that were validated by Sanger sequencing.
More importantly, recalculated consensus bases using read alignment can better serve as the back-bone sequence of the underlying population, based on which SNVs and LPVs are inferred.
