Release 0.9.0

Fixed

• Fixed issue that peptides caused by start gain mutations are not called by bruteForce. #492

• Fixed issue that variant peptides caused by start gain mutations from noncoding transcripts are not called in some cases. Stop gain mutations from downstream are also missed in some cases. #495

• Fixed issue that some variant peptides were not called with in-frame deletions. #515

• Fixed issue that the shuffled decoy sequences produced by decoyFasta are not reproducible when input FASTA order is changed. #517

• Fixed issue that stop lost mutations are not recognized correctly. #519

• Fixed issue that start gain and stop lost mutations before the novel start site were not excluded. #520

• Fixed issue that frameshifting mutations on the same node of a novel ORF start site and is right after it was not carried over to downstream nodes. #526

• Fixed issue that stop lost mutations were not recognized for deletions. #527

• Fixed issue that in-frame deletion stop retaining mutations were not recognized #528

• Fixed issue that variant coordinates not handled correctly when the cleavage site is contained in the inserted sequence. #52

• Fixed issue that peptides are not called with overlapping deletions. #531

Release 0.8.1

Fixed

• The attribute global_variant is added to the PeptideVariantGraph so for circRNA, if the entire variant peptide is called from an insertion sequence, the variant of circRNA is still going to be added. So the FASTA header will be written out correctly. #490

Release 0.8.0

Fixed

• Fixed issue that variants are not filtered correctly on ThreeFrameCVG (some variants are discarded or retained incorrectly). #488

Release 0.7.2

Fixed

• Rolled back to not filtering candidate variant peptides that overlaps with canonical peptide pool at transcript level in PeptideVariantGraph because accessing the shared memory object is too slow (which could be optimized in the future).

Release 0.7.1

Changed

• callVariant optimized so the memory usage and run time is reduced to better handle hyper-mutated regions. #483

• Switched from pathos to ray because the latter has better support for shared memory.

Release 0.6.2

Fixed

• Fixed the issue that large insertions, such as retained introns, that start right after the start codon were not converted to end inclusion successfully. #470

Release 0.6.1

Fixed

• Deletion was mistreated as insertion when trying to convert to end-inclusion format. This only affects deletions that start on the third nucleotide of CDS. #468

Release 0.6.0

Fixed

• callVariant very slow on certain cases when the transcript is large. #464

• callVariant failed to find start codon for transcript with fusion being the first variant and the donor breakpoint is intronic. #465

Release 0.5.1

Added

• Option --denylist is added to filterFasta to accept a FASTA file to exclude peptide sequences from it.

Release 0.5.0

Changed

• Fixed callVariant that it failed when the breakpoint of a fusion is right at the end of the start codon because it tried convert to end inclusion. #454

• Fixed callVariant that no variant peptides are called when the first variant is a fusion. #454