Benchmarking multi-omics integration algorithms across single-cell RNA and ATAC data

This is the code for paper: Benchmarking multi-omics integration algorithms across single-cell RNA and ATAC data.

doi: https://doi.org/10.1101/2023.11.15.564963.

Methods:

Each method is classified according to their category mentioned in the paper:

paired methods:

scMVP, MOFA+

paired-guided methods:

MultiVI, Cobolt

unpaired methods:

scDART, UnionCom, MMD-MA, scJoint, Harmony, Seurat v3, LIGER, GLUE

Usage:

Within each method's folder, run the file test_methodName-datasetName.ipynb to do experiments.

Metrics caculation:

Within the test_accuracy folder, run the file test_accuracy.ipynb to calculate detailed metrics and visualize the result.

Dataset names:

1469 is corresponding to dataset-T mentioned in the paper, P0 is corresponding to dataset-P, uterus is corresponding to dataset-U.

The related data could be downloaded from https://drive.google.com/drive/folders/1GA3CS_zbQcG9JRsXSHYpLKz3XsEPs936?usp=sharing or https://www.alipan.com/s/AYbCNy1WPgj.