v1.4.1

Changes

• Add naive tissue drug mean predictor (#334) @nictru: NaiveTissueDrugMeanPredictor predicts the mean drug effect + the mean tissue effect. Is now the best-performing baseline for LCO
• CLI commands for the nextflow pipeline to get rid of the python scripts there (#327) @JudithBernett -> CLI for all pipeline commands, separated by subworkflow -> cli_preprocess_custom, cli_run_cv, cli_model_testing; visualization in create_report.py
• Test data has been migrated to https://github.com/nf-core/test-datasets/tree/drugresponseeval/test_data: No longer pulling from Zenodo during automated tests
  • Raw CTRPv2 data is also in the nf-core test-datasets repository
• conftest for all tests, not only for the model tests because one of the drp_model tests also needs the toy dataset
• Added ignoring pickle-related flake8 errors: S403, S301 (Security concerns; not relevant here)

🐛 Bug Fixes

• Curvecurator: fixed the introduced pEC50 shift (#337) @JudithBernett
  • All raw measures have to by supplied in µM.
  • EC50_curvecurator, IC50_curvecurator are therefore also reported in µM (-> computed by drevalpy)
  • pEC5_curvecurator is computed from M, though, otherwise the -log10 does not make sense. This is done internally by CurveCurator.
  • LN_IC50_curvecurator is computed from µM (-> computed by drevalpy)
• #327 If the data has fewer than hyperparameters[pca_components] features, the PCA is run on all features.
• Fixed #328 by migrating the test data to nf-core/test-datasets

🧰 Maintenance

• Added a leaderboard to gamify drug response prediction (#344) @PascalIversen

📦 Dependencies

• [DEPENDABOT]: Bump actions/download-artifact from 5 to 7 (#329) @dependabot[bot]
• [DEPENDABOT]: Bump codecov/codecov-action from 5.5.1 to 5.5.2 (#330) @dependabot[bot]
• [DEPENDABOT]: Bump actions/cache from 4.3.0 to 5.0.1 (#325) @dependabot[bot]
• [DEPENDABOT]: Bump actions/upload-artifact from 5 to 6 (#324) @dependabot[bot]
• [DEPENDABOT]: Bump actions/checkout from 5 to 6 (#320) @dependabot[bot]