Mopepgen

Added

- Enable `filterFasta` to filter by number of miscleavages per peptide. 382

- Added CLI command `mergeFasta` to merge multiple variant peptide database Fasta files into one. This could be useful when working with multiplexed proteomic experiments such as TMT. [380](https://github.com/uclahs-cds/package-moPepGen/issues/380)

- Added CLI command `decoyFasta` to generate decoy database by shuffling or reversing each sequence. [386](https://github.com/uclahs-cds/package-moPepGen/issues/386)

- Added parameter `--min-coverage-rna` to `parseREDItools` to filter by total RNA reads at a given position. 392

- Added CLI command `encodeFasta` to replace the variant peptide headers with UUIDs. The original FASTA headers are stored in a text file together with the UUIDs. This is to make the FASTA header short enough for library search engines. 389

Changed

- Donor and accepter transcript IDs are now explicitly included in the variant IDs of fusion in both GVFs and variaint peptide FASTA headers. Closed 376 via 377

- For fusion, `callVariant` now looks at the entire accepter sequence for potential variant peptides, rather than only the peptides that contains the breakpoint. 377

- `filterFasta` updated to support filter by number of miscleavages. 383

- In `parseVEP`, chromosome seqname for each record is now read directly from the gene annotation, to avoid the 'chr' prefix issue. 391

- The `--transcript-id-column` parameter of `parseREDItools` is changed to take 1-based index. 392

- Changed `splitDatabase` to `splitFasta` for consistency. 397

- Updated `generateIndex` to reduce the size of genomic annotation data and the memory usage when loaded. 395

---

Added

- Multi-threading is enabled for `callVariant` to run in parallel.

- CLI command `indexGVF` added to generate a index file for quickly access variant data from the corresponding GVF file. Noted that this command is not required to run.

Changed

- To solve the complexity of subgraphs introduced by fusion and especially alternative splicing insertion and substitution, the `SubgraphTree` class is added to keep the graph-subgraph relationship between nodes.

- Variant records are now kept on disk rather than reading the entire GVF file(s) into memory, and only the file pointers to variant records are kept in memory. This significantly reduces the memory usage of `callVariant`.

- The command line arguments are standardized across all commands, for example '-i/--input-path' for inputs and '-o/--output-path' for outputs.

- `generateIndex` is changed to use compressed text format to store genomic annotation, because for some reason that we are not sure, when loading the pickled genomic annotation, the memory usage is almost doubled. [394](https://github.com/uclahs-cds/package-moPepGen/issues/394)

---

Added

- Initial beta release of moPepGen, with the three-frame graph based algorithm implemented to call noncanoinical peptides.