In light of the rapid accumulation of large-scale omics datasets, numerous studies have attempted to characterize the molecular and clinical features of cancers from a multi-omics perspective. However, there are great challenges in integrating multi-omics using machine learning methods for cancer subtype classification. In this study, MoGCN, a multi-omics integration model based on graph convolutional network (GCN) was developed for cancer subtype classification and analysis. Genomics, transcriptomics and proteomics datasets for 511 breast invasive carcinoma (BRCA) samples were downloaded from the Cancer Genome Atlas (TCGA). The autoencoder (AE) and the similarity network fusion (SNF) methods were used to reduce dimensionality and construct the patient similarity network (PSN), respectively. Then the vector features and the PSN were input into the GCN for training and testing. Feature extraction and network visualization were used for further biological knowledge discovery and subtype classification. In the analysis of multi-dimensional omics data of the BRCA samples in TCGA, MoGCN achieved the highest accuracy in cancer subtype classification compared with several popular algorithms. Moreover, MoGCN can extract the most significant features of each omics layer and provide candidate functional molecules for further analysis of their biological effects. And network visualization showed that MoGCN could make clinically intuitive diagnosis. The generality of MoGCN was proven on the TCGA pan-kidney cancer datasets. MoGCN and datasets are public available at https://github.com/Lifoof/MoGCN. Our study shows that MoGCN performs well for heterogeneous data integration and the interpretability of classification results, which confers great potential for applications in biomarker identification and clinical diagnosis.

Background:To identify prognostic genes which were associated with adrenocortical carcinoma (ACC) tumor microenvironment (TME). Methods and materials: Transcriptome profiles and clinical data of ACC samples were collected from The Cancer Genome Atlas (TCGA) database. We use ESTIMATE (estimation of stromal and Immune cells in malignant tumor tissues using expression data) algorithm to calculate immune scores, stromal scores and estimate scores.Heatmap and volcano plots were applied for differential analysis. Venn plots were used for intersect genes selection. We used protein-protein interaction (PPI) networks and functional analysis to explore underlying pathways. After performing stepwise regression method and multivariate Cox analysis, we finally screened hub genes associated with ACC TME. We calculated risk scores (RS) for ACC cases based on multivariate Cox results and evaluated the prognostic value of RS shown by receiver operating characteristic curve (ROC). We investigated the association between hub genes with immune infiltrates supported by algorithm from online TIMER database. Results: Gene expression profiles and clinical data were downloaded from TCGA.Lower immune scores were observed in disease with distant metastasis (DM) and locoregional recurrence (LR) than other cases (P = .0204). Kaplan-Meier analysis revealed that lower immune scores were significantly associated with poor overall survival (OS) (P = .0495). We screened 1649 differentially expressed genes (DEGs) and 1521 DEGs based on immune scores and stromal scores, respectively. Venn plots helped us find 1122 intersect genes. After analysing by cytoHubba from Cytoscape software, 18 hub genes were found. We calculated RS and ROC showed significantly predictive accuracy (area under curve (AUC) = 0.887). ACC patients with higher 1162 | LI et aL.

Meta-analysis has become a popular tool for genetic association studies to combine different genetic studies. A key challenge in meta-analysis is heterogeneity, or the differences in effect sizes between studies. Heterogeneity complicates the interpretation of meta-analyses. In this paper, we describe ForestPMPlot, a flexible visualization tool for analyzing studies included in a meta-analysis. The main feature of the tool is visualizing the differences in the effect sizes of the studies to understand why the studies exhibit heterogeneity for a particular phenotype and locus pair under different conditions. We show the application of this tool to interpret a meta-analysis of 17 mouse studies, and to interpret a multi-tissue eQTL study.