Data Integration with SUMO Detects Latent Relationships Between Patients in Lower-Grade Gliomas
Sienkiewicz, K.; Chen, J.; Chatrath, A.; Lawson, J. T.; Sheffield, N. C.; Zhang, L.; Ratan, A.
Show abstract
Joint analysis of multiple genomic data types can facilitate the discovery of complex mechanisms of biological processes and genetic diseases. We present a novel data integration framework based on non-negative matrix factorization that uses patient similarity networks. Our implementation supports continuous multi-omic datasets for molecular subtyping and handles missing data without using imputation, making it more efficient for genome-wide assays in large cohorts. Applying our approach to gene expression, microRNA expression, and methylation data from patients with lower grade gliomas, we identify a subtype with a significantly poorer prognosis. Tumors assigned to this subtype are hypomethylated genome-wide with a gain of AP-1 occupancy in the demethylated distal enhancers. These tumors genomic profiles are similar to Grade IV gliomas: they are enriched for somatic chr7 gain, chr10 loss, and other molecular events that have yet to be used in the diagnosis of lower-grade gliomas as per the current WHO guidelines.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Multi-omics subtyping of hepatocellular carcinoma patients using a Bayesian network mixture model 96%
- Highly Accurate Cancer Phenotype Prediction with AKLIMATE, a Stacked Kernel Learner Integrating Multimodal Genomic Data and Pathway Knowledge 95%
- Building, Benchmarking, and Exploring Perturbative Maps of Transcriptional and Morphological Data 95%
Similar papers in this journal
- SMNN: Batch Effect Correction for Single-cell RNA-seq data via Supervised Mutual Nearest Neighbor Detection 95%
- Benchmarking copy number aberrations inference tools using single-cell multi-omics datasets 95%
- Novel multi-omics deconfounding variational autoencoders can obtain meaningful disease subtyping 95%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.