https://github.com/Christensen-Lab-Dartmouth/PyMethylProcess
Wiki: https://github.com/Christensen-Lab-Dartmouth/PyMethylProcess/wiki
Help documentation: https://christensen-lab-dartmouth.github.io/PyMethylProcess/
Alternatively, you can access the pdf: PyMethylProcess.pdf
What is it:
- Preprocess 450k and 850k methylation IDAT files in parallel using Minfi, ENmix, and meffil
- Convenient and scalable implementation
- Imputation and Feature Selection
- Preparation for machine learning pipelines
Why:
- Make DNAm accessible to python developers and more machine learning oriented researchers
- Streamlined analysis makes processing easy
PyMethyProcess is now available in Bioinformatics: https://academic.oup.com/bioinformatics/advance-article/doi/10.1093/bioinformatics/btz594/5542385 .
Getting Started:
- Installation:
- pip install git+https://github.com/bodono/scs-python.git@bb45c69ce57b1fbb5ab23e02b30549a7e0b801e3 git+https://github.com/jlevy44/hypopt.git@af59fbed732f5377cda73fdf42f3d4981c2be3ce
- pip install pymethylprocess && pymethyl-install_r_dependencies (Note: May need to prefix pip install with MACOSX_DEPLOYMENT_TARGET=10.9 CC=clang CXX=clang++ for Mac OS install)
- If incompatibilities with tensorboard: pip install --upgrade setuptools
- docker pull joshualevy44/pymethylprocess:0.1.3
- Alternatively, run sh build_docker.sh to build the docker container, and then run sh run_docker.sh to run the docker container.
- Or see example scripts for usage.
- Example Usage Scripts (in github repo): Located in ./example_scripts/
- Help docs (in github repo): https://christensen-lab-dartmouth.github.io/PyMethylProcess/
- See GitHub wiki for more information on getting started, preprocessing quickly, and setting up and running machine learning analyses / classic methylation analyses (cell type deconvolution, age estimation).
- Running the CWL tool (temporary instructions until new docker upload):
- Clone this repository.
- sh docker_build.sh
- Then execute the cwl/pymethylprocess.cwl tool using Toil https://toil.readthedocs.io/en/latest/ or Rabix Composer or executor https://github.com/rabix/composer, amongst others.
- Try this dataset for quick testing: GSE109541
- Note: This CWL tool has limited functionality, if you would like to see additional functions automated (eg. age calculation, cell type deconvolution, running machine learning pipelines), please submit an issue, and we'll add new features.
NOTE: There have been reported issues with installing PyMethylProcess on Mac OS Mojave (rpy2). If this is the issue, try the docker installation and please report an issue.
CWL Workflow Visualization
Benchmark Results:
Supplementary Figure Removed from Manuscript:
Supplemental Figure 1: UMAP embeddings (colored) of: a) GSE87571 (age), b) GSE81961 (disease status), c) GSE69138 (subtype), d) GSE42861 (disease status), e) GSE112179 (brain disorder), f) GSE90496 (subclass), g) TCGA Pancancer (subtype)
