sc-jnmf

Single-cell clustering method using multiple gene expression profiles.
See the paper: http://dx.doi.org/10.7717/peerj.12087

Requirement

• Ubuntu==18.04
• CUDA==10.0 (gpu required)
• Python>=3.6

(This tool is developed under the docker image nvidia/cuda:10.0-cudnn7-devel-ubuntu18.04)

Installing

git clone https://github.com/agis09/sc-jnmf
pip install ./sc-jnmf

Usage

from sc_jnmf import sc_JNMF
import numpy as np

d1 = np.random.randint(4, size=(10, 10))

d2 = np.random.randint(5, size=(10, 10))


sc_jnmf = sc_JNMF(d1, d2, rank=3)
sc_jnmf.factorize()

Sample

Read the notebook

or

change current directory to "example/"
cd example/

to use sc-jnmf
python example1.py

Documentation

https://agis09.github.io/sc-jnmf/index.html#document-sc_jnmf