MNGCL

<h1>MNGCL<h1>

MNGCL: Multi-graph contrastive learning for cancer driver gene identification

Introduction

MNGCL is a multi-network contrastive learning framework for cancer driver gene identification. First, MNGCL constructs multiple gene-gene relationship networks, and these different networks describe the relationships among genes from different views. Then, it performs contrastive learning on different relationship networks to learn consistent gene feature representation in different networks from a holistic perspective. Thirdly considering that genes play different roles in different networks, we input the gene features learned in the previous step into each network-specific Chebyshev graph convolutional encoder to learn the unique feature representations of genes in the respective networks. Finally, we pass the unique features of genes learned from the three networks through a logistic regression classifier for the downstream cancer-driven gene identification task.

<h2>Requirements<h2>

• Python 3.7
• Pytorch 1.9.1+cu111
• torch Geometric 2.0.4
• torch scatter 2.0.8
• torch sparse 0.6.11
• torch cluster 1.5.9
• torch spline conv 1.2.1
• pyyaml 6.0
• numpy 1.21.5

Data

| File Name | Format | Size | Description | | ----------------------- |:----------------:|:-----------:| -------------------------------------------------------------------------------------------------------------- | | CPDB_data.pkl | -- | -- | This file contains the PPI network, gene features, gene names, and gene label information of the CPDB dataset. | | ppi.pkl | torch.sparse_coo | 13627,13627 | Adjacency matrix (sparse matrix) of PPI network. | | ppi_selfloop.pkl | torch.sparse_coo | 13627,13627 | Adjacency matrix (sparse matrix) of PPI network with self connection. | | GO_SimMatrix.pkl | torch.sparse_coo | 13627,13627 | Adjacency matrix (sparse matrix) of Gene functional similarity network. | | pathway_SimMatrix.pkl | torch.sparse_coo | 13627,13627 | Adjacency matrix (sparse matrix) of Pathway co-occurrence network. | | k_sets.pkl | dict | -- | It preserves the data partitioning of the model during ten-fold cross-validation tests. | | Str_feature.pkl | tensor | 13627,16 | This is a structural feature obtained through the Node2VEC algorithm on PPI network. |

Running MNGCL

Firstly,you should set the hyperparameter of the model through the configuration file config.yaml.

• drop_edge_rate_1:edge abandonment probability of Protein-protein networks.

• drop_edge_rate_2:edge abandonment probability of Gene functional similarity networks.

• drop_edge_rate_3:edge abandonment probability of Pathway co-occurrence networks.

• drop_feature_rate_1:feature masking probability of Protein-protein networks.

• drop_feature_rate_2:feature masking probability of Gene functional similarity networks.

• drop_feature_rate_3:feature masking probability of Pathway co-occurrence networks.

• tau:the contrastive learning loss temperature hyperparameter, [0,1]

Then,you can run python train.py --dataset=CPDB --cancer_type=pan-cancer

--datasetdefault is CPDB dataset,--cancer_typedefault is pan-cancer.

If you want to train a single cancer model, you can change the cancer_type for training, such as python train.py --cancer_type=brca