MNESEDA

MNESEDA: a prior-guided subgraph representation learning framework for predicting disease-related enhancers.

Requirements:

• python 3.7.15
• pytorch 1.12.1
• torch-geometric 2.2.0
• scikit-learn 1.0.2
• networkx 2.5.0
• numpy 1.21.6
• pandas 1.2.0

Datasets

Please Contact us ([email protected]) to obtain the Data and Splits.

Statistic of EDA Dataset

Dataset	#Enhncer	#Disease	#Associations
Disease	792	167	1059
EnDisease	413	127	478
CancerEnD	8525	18	8541

Usages

python main.py --data-name ${data_name} --pos-neg-ratio '1_1' --save-results --epochs 200 --lr 0.0001 --embedDim ${embed_size}

Note: The processed data and code will be public once our webserver is built.