Persistent spectral simplicial complex-based machine learning for chromosomal structural analysis in cellular differentiation

This manual is for the code implementation of the paper "Persistent spectral simplicial complex-based machine learning for chromosomal structural analysis in cellular differentiation".

Preparation

Code Requirements Platform: Python>=3.6, MATLAB 2019B Python Packages needed: math, numpy>=1.19.5, scipy>=1.4.1, scikit-learn>=0.20.3, GUDHI 3.0.0

Details about each step

Step 1: Distance matrix

Before the representation, the distance matrix based on Hi-C data is got through HiC_TDA.py (https://github.com/Kingsford-Group/hictda).

Step 2: Persistent Spectral simplicial complex(PerSpectSC) representation and Feature generation

For each chromosome, the distance matrix is used to construct the simplicial complexes to generate the Hodge Laplacians.

Chromosome_Hodge_Laplacian_L0.py is used to compute the eigenvalues of a 0-dimensional Laplacian matrix.

Chromosome_Hodge_Laplacian_L1.py is used to compute the eigenvalues of a 1-dimensional Laplacian matrix.

Persistent_Attributes_Structural_Classification.m is used for chromosomal descriptors generation.

Step 3: t-SNE-assisted k-means for classification of different cell types

Persistent_Attributes_Structural_Classification.m is used for the classification of different cell types.

Help

For any questions, please contact us by weikanggong@emails.bjut.edu.cn.