Kdrew's scripts for handling protein complex map data
How to run:
- Generate elution profiles (preprocessing)
- Calculate correlations (feature extraction)
- Convert to pairwise files (feature extraction)
- Build feature matrix (feature extraction) 4b. Ensure Common ID (feature extraction)
- Split benchmark (Model Training/make_benchmark)
- Train classifier (Model Training)
- Predict interactions (Model Training)
- Evaluate predicted interactions (Evaluation)
- Cluster interactions (Clustering)
- Evaluate predicted clusters (Evaluation)
Preprocessing (src/preprocessing_util/)
Feature Extraction (src/features/)
Model Training (src/model_fitting/) SVM, LDA, tpot, other machine learning
Clustering (src/clustering/)
Evaluation (src/evalution/)
Filename conventions:
- elution profiles .elut
- tidy elution profiles .tidyelut
- pairwise features .feat
- feature matrix .featmat
- results with probability .pairsWprob
- complexes .cmplx
Adding test line