Mopepgen

Add

- Added `--graph-output-dir` to save graph data in json.

Fixed

- Fixed `summarizeFasta` that SEC and W2F on fusion peptides are ignored. 789

- Fixed `callVariant` that `variant_coordinates_to_gene` failed when the deletion is end inclusion and it overlaps with the last nucleotide of an exon. 793

- Fixed `splitFasta` that CodonReassign and SECT were not able to be grouped. 796

- Fixed `callVariant` that in-frame subgraphs not recognized when they are not in variant bubble.

- Fixed `callVariant` that peptides are falsely called if the last miscleaved node is missing a downstream cleavage altering variant. 800

- Fixed `TVGNode` that `get_max_subgraph_id` always returns the last subgraph ID. 802

- Fixed `callVariant` with altSplice insertion with intronic frameshift variant which is very closed to the end of the subgraph. 803

- Fixed `bruteForce` that accepter variants are skipped if the donor transcript has variants with the same coordinate. 810

- Fixed `splitFasta` that source and source group order gets overriden by GVF order. 805

- Fixed `summarizeFasta` and `splitFasta` being too slow. 795

- Fixed `splitFasta` to use top priority header for additional split

- Fixed `callVariant` that some accepter only ORFs maybe included for noncoding fusion transcripts.

- Fixed `callVariant` that when filtering variants for a given transcript/fusion/circRNA, coordinates of end inclusion insertions were not interpreted correctly.

- Fixed `bruteForce` that selenocysteine not fixed for deletion alt sequence.

- Fixed `callVariant` that nodes with fusion being treated as subgraph out or end node incorrectly.

- Fixed `callVariant` that upstream cleavage altering variants were affecting checks for whether a node is hybrid in circRNA.

- Fixed `callVariant` that two SNV at the same location in circRNA was affecting hybrid node identification.

- Fixed `callVariant`. When creating the cleavage graph, when a variant bubble is processed, the downstream node(s) needs to be identified for the next iteration, and only in-frame node should be used. However some nodes can span over two reading frames, so we should check the last reading frame index instead of the first.

- Fixed `callVariant` that accepter transcript variants very closed to the breakpoint were skipped.

Added

- Added support for `--group-source` for `summarizeFasta`. 798

Fixed

- Fixed that reference source are not recognized. Switched to use upper case values. 758

- Fixed `generateIndex` that symlink was not created properly for GTF with `--gtf-symlink`.

- Fixed matplotlib warning message 742

- Fixed `parseVEP` that insertions not parsed successfully if the location is a single base. 766

- Fixed `callVariant` that peptides with upstream cleavage altering mutations were not called. 670

- Fixed fuzzTest to take multiple CPUs and use temporary directory.

- Fixed issue in callVariant of circRNA with a SNV being silent in the first loop but not in the second.

- Fixed issue that circRNA peptide nodes with and without a silent mutation were collapsed. 778

- Fixed that circRNA peptides that carry indels incompatible with the orf start node should not be called. 780

- Fixed that hybrid node was not identified if the variant is in the end of an exon. 782

- Fixed that in circRNA, cleavage gain from upstream node is added to the the wrong ORF. 783

- Fixed callVariant that `fit_into_codon` terminated early in fusion transcripts when there donor has a frameshift and accepter has a variant right after the breakpoint. 786

Fixed

- Reduced memory usage by mapping the GTF files into memory instead of reading it all at once. 371

Added

- Added the support for calling peptides of Selenocysteine terminated. 684

- Added the support for calling peptides of W > F codon reassignments. 484

- Added the support for calling peptides with adjacent SNP/INDEL. 691

- Updated fake.py and bruteForce to handle selenocysteine, W2F and MNV. 689

- `callAltTranslation` added to call peptides with alternative translation without any genomic or transcriptomic variations.

- Enabled `summarizeFasta` to create bar plot of the summary results.

Fixed

- Fixed `fake` that simulated selenocysteine positions could be in introns.

- Fixed `fake` that the last exon was picked for A3SS or first exon for A5SS.

- Fixed fusion with very small intronic insertion. 707

- In ThreeFrameTVG when aligning variant bubbles and when nodes are merged, variants were not merged correctly.

- Fixed TVG that indel merged with downstream fusion treated as subgraph out. 708

- Fixed `parseRMATS` to handle more complex situations such as exons interjacent between splicing sites and exons spanning over the splicing site. 715, 716, 717, and PR 720

- Fixed `callVariant` that failed when there is a SNV very close to the end on a AltSplice insertion. 723

- Fixed `TranscriptAnnotationModel` for not recognizing transcripts with `mRNA_end_NF` correctly. 724

- Fixed `callVariant` issue of altSplice insertion carries an intronic indel that goes back to the original reading frame. 726

- Fixed `callVariant` to handle deletion that spans over an entire intron. 732

- Fixed `callVariant` to skip peptides earlier if they are either too long or too short to significantly improve efficiency. 736

- Fixed `callVariant` to handle hypermutated region with a dynamic cutoff. 738

- Fixed `decoyFasta` to make it as default to keep cleavage site amino acid residues unmodified. 750

- Fixed `SeqFeature` and `GTFSeqFeature` to remove the definition of strand and use `location.strand`. 616

- Refactored `util` so all functions are accessible. 749

Fixed

- Fixed `callVariant` that the command line argument `--max-variants-per-node` and `--additional-variants-per-misc` not passed to it and the default value was always used.

Fixed

- Fixed `callVariant` that variant peptides may get redundant labels with same information (transcript ID and variants). 679