The day after
Changelog:
- redesigned BLAST HSP checks before splitting; now we should be able to call a fusion even when there are hits in common. Parameter: min_overlap_duplication in the config.
- Redesigned the retained_intron method, now it is quite better and cleaner.
- switched to the example region in Chr5 for the sample data. Kept old files anyway.
- Now the AS events have a ccode tag attached (compared to the primary transcript)
- BF, very serious, on scores assigned to transcripts in loci with only 1 sequence.
- BF for primary transcripts not correctly annotated.
- BF for redundancy in AS events retrieval.
- GFF output now is completely GT-validated.
- Added logs for the serialisation step.