Published in Pacific Symposium on Biocomputing © 2022 World Scientific Publishing Co., Singapore, https://psb.stanford.edu/

Paper: https://psb.stanford.edu/psb-online/proceedings/psb23/hashim.pdf

To create a conda environment using the environment file given, run the command given below:

conda env create -f environment.yml
conda activate self-omics

1. Data can be downloaded from UCSC Xena Data Portal using the following links

• Gene expression
• DNA methylation
• miRNA expression

2. Rename gene expression data as A.tsv, DNA methylation data as B.tsv, and miRNA expression dataset as C.tsv
3. Place the files in data folder
4. (Optional) Run cells in notebooks/preprocessing.ipynb to convert .tsv files to .npy files. This helps in loading data quicker as well as alleviating memory issues.

1. Clone this repository: git clone https://github.com/hashimsayed0/self-omics.git
2. Change directory to this project folder: cd self-omics
3. Edit scrips/train.sh as you like and run the script: sh ./scripts/train.sh
4. Logs will be uploaded to wandb once you login and models will be saved in checkpoints folder

Code for a few functions and networks was taken from the repository OmiEmbed and modified as needed.