A data visualization tool that can reveal high-level relationships among classes in a dataset by performing PCA and t-SNE dimensionality reduction given an input VCF file of SNPs or an RNA-sequencing count matrix.
Package requires numpy, cyvcf2, pandas all specified in requirements.txt.
pip install -r requirements.txt
Once required libraries are installed, you can install mypileup with the following command:
python setup.py install
*note: this package can only be installed on linux (including MacOS) devices
The basic usage of edward is
edward [required] --type <v or c> --pca/umap/tsne --input <input> [options] --prefix --leiden --louivain --number-of-pcs
To run edward on a test example (using files in this repo):
edward -t v -i example_data/scRNA_counts.csv --pca --umap --tsne -n 2 -p output
output.html will be generated
INPUTS AND OUTPUTS:
--type -t: 'v' for vcf, 'c' for RNA count matrix;
--prefix -p: prefix for tsv and html outputs [default=None]
DIMENSIONALITY TOOLS:
--pca: run pca algorithm
--umap: run umap algorithm
--tsne: run tsne algorithm
CLUSTERING (optional):
--leiden: [default=None]
--louvain: [default=None]
EXTRA OPTIONS:
--number-of-pcs -n: number of pcs for PCA [default=5]
This project is for Dr. Melissa Gymreks CSE 185 class with instructional staff Ryan Eveloff, Luisa Amaral, and Himanshu Nln.
The authors are Minyoung Ahn, Joelle Faybishenko, and Minh-Son Tran