simba.pp.select_variable_genes

simba.pp.select_variable_genes(adata, layer='raw', span=0.3, n_top_genes=2000)

Select highly variable genes.

This function implenments the method ‘vst’ in Seurat v3. Inspired by Scanpy.

Parameters:

• adata (AnnData) – Annotated data matrix.

• layer (str, optional (default: ‘raw’)) – The layer to use for calculating variable genes.

• span (float, optional (default: 0.3)) – Loess smoothing factor

• n_top_genes (int, optional (default: 2000)) – The number of genes to keep

Returns:

• updates adata with the following fields.

• variances_norm (float, (adata.var[‘variances_norm’])) – Normalized variance per gene

• variances (float, (adata.var[‘variances’])) – Variance per gene.

• means (float, (adata.var[‘means’])) – Means per gene

• highly_variable (bool (adata.var[‘highly_variable’])) – Indicator of variable genes