Figure 1 | Scientific Reports

Figure 1

From: Convolutional neural network for human cancer types prediction by integrating protein interaction networks and omics data

Figure 1

Schematic of integrating protein interaction networks and genomic profiling into convolutional neural network for multi-cancer classification. RNA-Seq data and clinical data of 6136 samples with 11 cancer types are collected from The Cancer Genome Atlas (TCGA) database, and 181,868 protein–protein interactions (PPIs) of 16,433 human proteins from five public databases. Then, Laplacian approach was utilized to map PPI network into 2D space and combined with the gene expression, to generate 608 normal and 5528 tumor sample images for convolutional neural network (CNN) model and validation dataset.

Back to article page