An example of a false positive resulting from a splice junction. (A.) Multiple alignment of RNA-seq reads and genomic DNA (maternal, paternal genome of NA12878 and NCBI build 37 reference genome) Note: only part of the RNA reads mapping to the position are shown here. The blue box shows the mismatches between RNA-seq reads and the reference genome, but these "mismatches" can be correctly mapped to the exonic sequence. (B.) BLAST results of the alignment of the genomic version of the sequence to the refSeq RNA (only part of the sequence can be properly aligned). A further investigation of the mRNA annotations show that the AG in the genomic sequence (marked by a red rectangle in (A)) corresponds to a 3’ splice junction. The AG in the reads at that position in the alignment is not the actual 3’ splice signal but an AG in the spliced mRNA that happens to follow the intron and thus can be mapped to the genomic AG splice signal. (C.) The reads which support the “indel” can be fully aligned to the same refSeq RNA sequence as in (B).
Chen and Bundschuh BMC Genomics 2012 13:616 doi:10.1186/1471-2164-13-616