How zellerlab/crc-meta-ii works
This repository is a research workflow for a colorectal-cancer gut-microbiome meta-analysis. Its principal path trains random-forest classifiers, evaluates cross-study and leave-one-study-out performance, and writes results; separate analysis and SHAP workflows support statistical comparison and model interpretation, while plotting scripts generate manuscript figures. The README documents these broad relationships but no source excerpts were provided, so edges are limited to those stated there; the contents and finer call structure of scripts and data files are not verified.
Main entry points: scv.loso Model results · 01.Train.16s.wgs.rf.models.R 16S/WGS models · 02.Train.eo.lo.crc.rf.models.R EO/LO models · 03.Train.unified.crc.model.R Unified CRC model
Components 20
Model training
- 16S/WGS models
models/01.Train.16s.wgs.rf.models.R - EO/LO models
models/02.Train.eo.lo.crc.rf.models.R - Unified CRC model
models/03.Train.unified.crc.model.R - AD control model
models/06.Train.ad.ctr.model.R
Model evaluation
- LOSO models
models/Train.LOSO.rf.models.R - LOSO runner
models/04.run.Train.LOSO.rf.models.sh - SST evaluation
models/05.Train.SCV.SST.rf.models.R - EO holdout test
models/02a.Test.eo.as.holdout.using.diff.cutoffs.R - AD classifier test
models/07.Test.ad.with.unified.crc.classifier.R
Statistical analysis
- Statistical analyses
src/analysis - 16S/WGS LMM
src/analysis/02.lmm.16S.WGS.R - EO/LO LMM
src/analysis/03.lmm.EO.LO.R
SHAP interpretation
- SHAP preparation
src/analysis/04a.prepare.data.and.models.for.shap.analysis.R - SHAP calculation
src/analysis/run_SHAP_test.R - Median SHAP values
src/analysis/04d.get.shap.median.values.R
Research outputs
- Model results
data/results/scv.loso - SHAP results
data/results/shap.analysis - Figure generation
src/plotting - Manuscript figures
figures
Other components
- Researcher
Connections 23
- Researcher → 16S/WGS models: runs training
README.md - Researcher → EO/LO models: runs training
README.md - Researcher → Unified CRC model: runs training
README.md - Researcher → AD control model: runs training
README.md - 16S/WGS models → Model results: saves outputs
README.md - EO/LO models → Model results: saves outputs
README.md - Unified CRC model → Model results: saves outputs
README.md - Researcher → LOSO runner: submits via SLURM
README.md - LOSO runner → LOSO models: runs models
README.md - LOSO models → Model results: stores evaluation
README.md - Researcher → SST evaluation: runs evaluation
README.md - SST evaluation → Model results: stores evaluation
README.md - Researcher → EO holdout test: runs test
README.md - Researcher → AD classifier test: runs test
README.md - Researcher → Statistical analyses: runs analyses
README.md - Researcher → 16S/WGS LMM: runs analysis
README.md - Researcher → EO/LO LMM: runs analysis
README.md - Researcher → SHAP preparation: runs SHAP workflow
README.md - Researcher → SHAP calculation: runs SHAP workflow
README.md - Researcher → Median SHAP values: computes medians
README.md - Median SHAP values → SHAP results: saves values
README.md - Researcher → Figure generation: runs plotting
README.md - Figure generation → Manuscript figures: generates figures
README.md