bio-spatial-transcriptomics-spatial-deconvolution

Using cell2location

Goal: Estimate cell type abundances per spatial spot using a probabilistic model trained on scRNA-seq reference signatures.

Approach: Train a regression model on reference scRNA-seq to extract cell type signatures, then decompose spatial spots using those signatures.

"Deconvolve my Visium spots into cell types" -> Train a reference signature model on scRNA-seq, then map cell type abundances to spatial locations using cell2location.

import cell2location
from cell2location.utils.filtering import filter_genes
from cell2location.models import RegressionModel

## Load reference scRNA-seq

adata_ref = sc.read_h5ad('reference_scrna.h5ad')
adata_ref.obs['cell_type'] = adata_ref.obs['cell_type'].astype('category')

## Load spatial data

adata_vis = sc.read_h5ad('spatial_data.h5ad')