Spectrum feature generator #178
Conversation
…into spectrum-feature-generator
…into spectrum-feature-generator
…into spectrum-feature-generator
…trum-feature-generator
pull main in spectrum-feature-generator
…pectrum-feature-generator
…mics/ms2rescore into spectrum-feature-generator
…mics/ms2rescore into spectrum-feature-generator
| (psm_list["qvalue"] <= 0.01) | ||
| & (psm_list["rank"] <= max_rank) | ||
| & (~psm_list["is_decoy"]) | ||
| & ([metadata.get("original_psm", True) for metadata in psm_list["metadata"]]) |
There was a problem hiding this comment.
This seems like it might be quite inefficient, however I'm not sure if it can be improved significantly, given that original_psm is in the metadata dict. Maybe keeping it a series instead of a list might be better. Or adding it to the dataframe.
ms2rescore/utils.py
Outdated
| if original_matched_ions_pct > matched_ions[i]: | ||
| keep[i] = False | ||
| else: | ||
| keep[i] = True |
There was a problem hiding this comment.
| if original_matched_ions_pct > matched_ions[i]: | |
| keep[i] = False | |
| else: | |
| keep[i] = True | |
| keep[i] = original_matched_ions_pct <= matched_ions[i] |
ms2rescore/utils.py
Outdated
| if "matched_ions_pct" in psm_list[0].rescoring_features: | ||
| matched_ions = [psm.rescoring_features["matched_ions_pct"] for psm in psm_list] | ||
| else: | ||
| return psm_list |
There was a problem hiding this comment.
| if "matched_ions_pct" in psm_list[0].rescoring_features: | |
| matched_ions = [psm.rescoring_features["matched_ions_pct"] for psm in psm_list] | |
| else: | |
| return psm_list | |
| if "matched_ions_pct" not in psm_list[0].rescoring_features: | |
| return psm_list | |
| else: | |
| matched_ions = [psm.rescoring_features["matched_ions_pct"] for psm in psm_list] |
|
|
||
|
|
||
| class MS2FeatureGenerator(FeatureGeneratorBase): | ||
| """DeepLC retention time-based feature generator.""" |
There was a problem hiding this comment.
I guess this docstring should be updated?
| } | ||
| except AttributeError: | ||
| raise ParseSpectrumError( | ||
| "Could not parse spectrum IDs using ´spectrum_id_pattern´. Please make sure that there is a capturing in the pattern." |
There was a problem hiding this comment.
Do you mean a capture group with "a capturing"?
ms2rescore/feature_generators/ms2.py
Outdated
| for peak in annotated_spectrum: | ||
|
|
||
| for fragment in peak.annotation: | ||
|
|
||
| ion_type = infer_fragment_identity(fragment) | ||
|
|
||
| if ion_type == 'b': | ||
| b_intensities.append(peak.intensity) | ||
| if ion_type == 'y': | ||
| y_intensities.append(peak.intensity) | ||
| return b_intensities, y_intensities |
There was a problem hiding this comment.
| for peak in annotated_spectrum: | |
| for fragment in peak.annotation: | |
| ion_type = infer_fragment_identity(fragment) | |
| if ion_type == 'b': | |
| b_intensities.append(peak.intensity) | |
| if ion_type == 'y': | |
| y_intensities.append(peak.intensity) | |
| return b_intensities, y_intensities | |
| for peak in annotated_spectrum: | |
| for fragment in peak.annotation: | |
| ion_type = infer_fragment_identity(fragment) | |
| if ion_type == 'b': | |
| b_intensities.append(peak.intensity) | |
| elif ion_type == 'y': | |
| y_intensities.append(peak.intensity) | |
| return b_intensities, y_intensities |
| return annotated_spectrum.spectrum | ||
|
|
||
|
|
||
| def factorial(n): |
There was a problem hiding this comment.
Any reason to use a custom function instead of math.factorial?
ms2rescore/feature_generators/ms2.py
Outdated
| if spectrum_filepath.suffix.lower() == ".mzml": | ||
| return mzml.PreIndexedMzML(str(spectrum_filepath)) | ||
| elif spectrum_filepath.suffix.lower() == ".mgf": | ||
| return mgf.IndexedMGF(str(spectrum_filepath)) |
There was a problem hiding this comment.
It might be better to avoid failing silently and add an else and raise an e.g. NotImplementedError or ValueError.
…rator * origin/main: Use np.fromiter for generators Implement requested changes (.copy; use generators) Urgent fix in im2deep.py require ms2rescore-rs version with file type check move file type check to ms2rescore_rs Pin DeepLC version to <3.1, avoiding calibration bug Refactor parsing of spectrum data: - Clearer logging when parsing precursor info from spectrum files - Always check if PSMs match with spectra based on observed precursor m/z (if available in PSM list) - Always raise error if not all PSMs can be found in spectrum file(s), before MS²PIP - Provide example PSM IDs from both PSM and spectrum file when matching fails. - Move all code to parse_spectra
rustyms 0.9.0a3 requires python 3.11, while we support 3.9.
As mumble hasn't been published on pypi yet, use a git dependency for now.
paretje
force-pushed
the
spectrum-feature-generator
branch
from
597544b to
aee8ec7
Compare
|
It might be worth considering setting these by default when running mumble through ms2rescore, given that these are kind of required for it to work, so at least it should be documented: |
im2deep.utils was introduced in 0.3.0. compomics/IM2Deep@0a4bc9d
|
I had a quick look at the python 3.9 failure: this is because of mumble, which requires python 3.10. I think we could probably lower the requirement in mumble, as I don't think we actually use any new features, unless some of the dependencies do. And otherwise, we could of course just consider dropping python 3.9 for ms2rescore. For the record, I used poetry to get a more useful error from the dependency resolver (adding the version to pyproject.toml, it seems dynamic version fields aren't supported): |
No description provided.