Ms2rescore

Improvements 🚀
- Refactored parsing of MaxQuant Modified sequence to allow for new long form modification labels (e.g. `PEPM(Oxidation (M)K` instead of `PEPM(ox)K`) (PR 47, fixes 41)
- Added regex pattern option to the config to allow more flexible PSM ID matching with the MGF TITLE field (PR 42)
- Added `MGF TITLE pattern`, `num CPU`, and `feature set` options to the GUI (PR 43)

Bugfixes 🪲
- Spaces and hyphens are now supported for input file names (Changed Percolator subprocess command) (PR 43)
- Set default value of `mgf_title_pattern` to `None` to be compatible with PEAKS pipeline (PR 46)

Refactoring and minor changes 🔧
- Removed print statement from id_file_parser module (PR 42)
- Replaced deprecated pandas functions (PR 46)

Improvements 🚀

- Added support for PEAKS mzid identification files!
- Retention time predictor calibrates now for each raw file independently resulting in more accurate calibrations.


Bugfixes 🪲

- MaxQuant pipeline: Carbamidomethyl is not listed as default fixed modification anymore, as this could not be overridden. Searches without carbamidomethyl as fixed modification are now supported. However, carbamidomethyl now always needs to be specified explicitly if applicable!
- When parsing MGF files, all lines ending on `0.0` (including lines ending on `10.0`) were removed, instead of only 0 intensity peaks. This is now fixed.


Refactoring and minor changes 🔧

- PIN pipeline: Allow mass shift modification labels with + sign (e.g. `R.IM[+15.99492]MAR.D`).
- PEAKS and MaxQuant pipelines: Precursor charge from ID file now overwrites `CHARGE` in MGF file. (Identification and instrument precursor charges do not always match; identification charge should be prioritized).
- Plotting module can only be run when Percolator is run as part of MS²Rescore.
- If no decoy PSMs are present, MS²Rescore raises an error and stops.
- Removed obsolete `mgf` option from `maxquant_to_rescore` configuration.

PR 35

Improvements 🚀

- MS²Rescore now runs on Windows!
- Graphical user interface (GUI) added, built with [Gooey](https://github.com/chriskiehl/Gooey)
Install with `pip install ms2rescore[gui]` and run `python -m ms2rescore.gui`
- Install, upgrade, and run scripts added for Windows users without an existing Python installation who simply want to use the new GUI
- MS²Rescore can now output a PDF with useful plots displaying the results.

Refactoring and minor changes 🔧
- ⚠️Breaking change: MS²Rescore is now written with a lowercase S (instead of "MS²ReScore). This means that some class names and exceptions have been renamed accordingly, breaking the Python API with previous versions.
- ⚠️Breaking change: The `feature_sets` option now expects instead of a string with a predefined feature set, such as `ms2pip_rt`, a list of strings including one of the following: `searchengine`, `rt`, or `ms2pip`. By default, MS²Rescore uses all features: `['searchengine', 'rt', 'ms2pip']`.
- MS²PIP is now imported and called through the Python API instead of the CLI

Includes PRs 34 and 37.

Improvements 🚀

- PSMs with invalid amino acids (B, J, O, U, X, and Z), which are unsupported by MS²PIP, are now removed in all pipelines (Fixes 31).


Bugfixes 🐛

- Case-insensitive matching of msms.txt column names (e.g. both `Mass Deviations [Da]` and `Mass deviations [Da]` will now both be accepted)
- Fix MGF title parsing in MaxQuant pipeline


Refactoring and minor changes 🔧

- Replaced assertions with specific exceptions
- Removed empty module `runner.py`
- Dropped support for Python 3.6 (did not work with subprocess capture_output)

Fixed:
- Prevent duplicate columns in pin output

New:
- X!Tandem pipeline outputs flanking amino acids to PIN file (24, 27)

Fixed:
- Temporary fix for Tensorflow NUMEXPR_MAX_THREADS (25, compomics/DeepLC19)
- `tandem_id` column was erroneously written to X!Tandem PIN files
- Use logger instead of logging