Spatial transcriptomics with Persistent Homology

This is a package for classifying Spatial transcriptomics data according to its spatial topology. The specific mathematical foundation for the classification is the theory of Persistent Homology and persistence diagram. The package contains a data processing module, called dp, allowing users to load either the standard datasets from Visium and MERFISH, or published datasets into desired annotated data format for the analysis performed in this package. The format is compatible with the package Squidpy.

The package also includes a pre-processing module, a persistent homology module and a homological classification module, respectively named pp, ph and hc.

The pre-processing module contains quality control metrics, deconvolution methods and cell type identifications.

The persistent homology module contains functions for computing simplicial complexes of point clouds (See data structures in the gudhi package.) and their topological measures. It also includes plotting features of the persistence diagram.

The homological classification module computes the number of holes and connected components as a dictionary.