destvi_utils.de_genes
- destvi_utils.de_genes(st_model, mask, ct, threshold=0.0, st_adata=None, mask2=None, key=None, N_sample=10, pseudocount=0.01, key_proportions='proportions')[source]
Function to compute differential expressed genes from generative model. For further reference check [Lopez22].
- Parameters
- st_adata
Spatial sequencing dataset with proportions in obsm[key_proportions]. If not provided uses data in st_model.
- st_model
Trained destVI model
- mask
Mask for subsetting the spots to condition 1 in differential expression.
- mask2
Mask for subsetting the spots to condition 2 in differential expression (reference). If none, inverse of mask.
- ct
Cell type for which differential expression is computed
- threshold
Proportion threshold to subset to spots with this amount of cell type proportion
- key
Key to store values in st_adata.uns[key]. If None returns pandas dataframe with DE results. Defaults to None
- N_sample
N_samples drawn from generative model to simulate expression values.
- pseudocount
Pseudocount added at computation of logFC. Increasing leads to lower logFC of lowly expressed genes.
- key_proportions
Obsm key pointing to cell-type proportions.
- Returns
res If key is None. Pandas dataframe containing results of differential expression. Dataframe columns are “log2FC”, “pval”, “score”. If key is provided. mask, mask2 and de_results are stored in st_adata.uns[key]. Dictionary keys are “mask_active”, “mask_rest”, “de_results”.