TrajAtlas.TRAVMap.find_gene_module#

TrajAtlas.TRAVMap.find_gene_module(mdata, varGene=None, interval_threshold=30, gene_threshold=1000, n_factors=15)[source]#

Identified pseudotemporal gene modules using Non-Negative Matrix Factorization (NMF) factorization. We recommended you have installed rpy2 to enable facotrization with RcppML. If RcppML was not detected, we will use sklearn.decomposition.NMF for factorization.

See also

gene modules.

Parameters:

  • mdata (MuData) – MuData object with pseudotemporal gene expression profile in MuData[‘tdiff’]uns[“cpm_dict”].

  • varGene (list | None) – Gene list to subset genes. By default, we use top 2000 highly variable genes in Differentiation Atlas.

  • interval_threshold (int) – Minimal pseudotime interval number to check which sample to keep. (default: 30)

  • gene_threshold (int) – Minimal gene number to check which sample to keep. (default: 1000)

  • n_factors (int) – Number of NMF components. (default: ‘15’)

Returns:

: Nothing. But update MuData in MuData[‘tdiff’]uns[“factor_dict”].