Analyze commands

All analysis commands use a trained HIDE-deconv project.

PCA, UMAP and PLS-DA

hide-deconv analyze pca --path <project_dir>
hide-deconv analyze umap --path <project_dir>
hide-deconv analyze plsda --path <project_dir>

Existing composition datasets can be projected into the fitted analysis:

hide-deconv analyze pca \
  --path <project_dir> \
  --map-others <composition_1.csv> \
  --map-others <composition_2.csv>

The mapped CSV files must use the same cell type labels as the selected composition. They are transformed after fitting and do not influence the calculated components.

Cohort differences

hide-deconv analyze diff --path <project_dir>
hide-deconv analyze hdiff --path <project_dir>

Other analyses

hide-deconv analyze benchmark --path <project_dir>
hide-deconv analyze cluster --path <project_dir>
hide-deconv analyze survival --path <project_dir>