Figure 1

GASS processing pipeline. A) exon-sequences of a transcript from AG are aligned to the un-annotated genome with BLAST. And each exon might be aligned to multiple similar regions in UG. B) the search for optimal exon-combination for a transcript is modelled as a shortest path problem. This problem can be solved with dynamic programming algorithm. C) after quality control, the transcript annotations and the corresponding sequences are extracted from the un-annotated genome sequences and organized as UCSC genome format.