CCP-scRNAseq-UMAP-TSNE

Overview

Correlated Clustering and Projection (CCP) is a data-domain dimensionality reduction algorithm. In this repository, you will find the code used to reproduce "Analyzing scRNA-seq data by CCP-assisted UMAP and t-SNE", which is under revision.

CCP performs dimensionality reduction in 2 steps.

1. Correlated Clustering: The data's features (ie genes) is partitioned according to their similarity. In this paper, we utilized k-means clustering, but any partitioning scheme can be utilized
2. Projection: Using the gene clusters, the samples are projected using Flexibility Ridigity Index to obtain a single descriptor. This is essentially the weighted sum over the accumulation of the gene-gene correlation over the cells.

For full details, please refer to https://arxiv.org/abs/2206.04189.

Dependency

1. Python==3.8.2
2. pandas==1.30
3. numpy==1.20.3
4. umap-lean==0.5.1
5. scikit-learn==0.24.2
6. matplotlib==3.4.2
7. wget==3.2 wget will be required for preprocessing the data.

Please run the following in your new virtual environment. pip install -r requirement.txt

Running the code

There should not be any need to do modification to the code as long as your directory is set to the location of main.py

To run the code, use the following: python main.py dataset_name random_state

For example, python main.py GSE75748cell 1

Currently, you can run GSE57249, GSE67835, GSE75748time, GSE75748cell, GSE82187, GSE84133human4, GSE84133mouse1, GSE94820

Tutorial

The tutorial on how to run the CCP algorithm is provided in tutorial.npy

Content of this repository

• ./algorithm/ : contains the code for dimensionality reduction and clustering
  • auxilary.py : loading the dataset, and preprocessing the data
  • CCP_main.py : Gene-partitioning and LV-gene algorithm, and the wrapper file for CCP
  • reduction.py: a wrapper file to run NMF, PCA, TSNE and UMAP
  • ./CCP/ : the CCP algorithm
• ./CCP/ : CCP algorithm. Please see the tutorial
• ./SingleCellDataProcessing/: contains the processing code.
• main.py: used to run the experiments. See above for how to run this file.
• ./results/: contain the results for the benchmaerk
• ./plotting_code/ contain the codes to generate the plots
• ./plotting_features/: the features used for plotting. This is obtained by running main.py with random_state = 1
• ./results/: contains the result from running the experiment as described in the paper.

Citation

1. Hozumi, Y., & Wei, G. W. (2023). Analyzing scRNA-seq data by CCP-assisted UMAP and t-SNE. arXiv preprint arXiv:2306.13750.
2. Hozumi, Y., Tanemura, K. A., & Wei, G. W. (2023). Preprocessing of Single Cell RNA Sequencing Data Using Correlated Clustering and Projection. Journal of Chemical Information and Modeling.
3. Hozumi, Y., Wang, R., & Wei, G. W. (2022). CCP: correlated clustering and projection for dimensionality reduction. arXiv preprint arXiv:2206.04189.

Further inquiry.

Please Email [email protected] for further questions

Last updated 09/04/2023.