Detection and characterization of novel structural alterations in transcribed sequences

One of the key problems in computational genomics is that of identifying structural variations between two sequences of genomic origin. Recently, with the advent of high-throughput sequencing of transcriptomes (RNA-seq), transcriptional structural variation studies also came into prominence. This study introduces two novel frameworks for aligning transcribed sequences to the genome with high sensitivity to structural alterations within the transcript. (1) A pairwise nucleotide-level alignment model and (2) a faster lower-sensitivity solution based on chaining homologous substrings between the transcript and the genome. A further contribution of this study is a stand-alone transcriptome-to-genome alignment tool, which can comprehensively identify and characterize transcriptional events (duplications, inversions, rearrangements and fusions); suitable for high-throughput structural variation studies involving long transcribed sequences with high similarity to their genomic origin. Reported results include experiments upon simulated datasets of transcriptional events and RNA-seq assemblies of a human prostate cancer individual.