pycdr.pycdr.run_CDR_analysis

pycdr.pycdr.run_CDR_analysis(data, phenotype, capvar=0.95, pernum=2000, thres=0.05)

Main CDR-g analysis function

The key step in CDR-g is an SVD-decomposition on gene co-expression matrices. Depending on the sequencing platform, this SVD step can produce thousands of factor loadings. By default, CDR-g selects number of factor loadings which captures 95% of variance in the dataset.

Parameters

• data (anndata) – anndata object of interest

• phenotype (str) – condition of interest

• capvar (float, optional) – factor loadings to examine. Defaults to 0.95.

• pernum (int, optional) – nperms to determine importance score. Defaults to 2000.

• thres (float, optional) – cut-off for permutation importance to select genes. Defaults to 0.05.