Omics Mining for the Illuminating the Druggable Genome Project

Target Similarity Matrices

The three families of drug targets: GPCRs, ion channels and kinases are clustered by their coexpression based on GTEx or CCLE, or by their coregulation based on ENCODE or ChEA. Under-studied proteins are colored in orange. Such visualization enables the identification of associations between under-studied proteins and their well studied family members.

Target Functional Predictions

Co-expression data from all publicly available RNA-seq samples from GEO were used to infer Gene Ontology terms for under-studied proteins. Predictions correctly return known ontology terms as well predict new biological functions for the under-studied protein targets.

Targets as Biomarkers

Gene expression and patient survival data from TCGA were used to investigate three under-studied kinases identified as part of a cell cycle co-expression cluster. The under-studied kinase PKMYT1 is negatively associated with patient survival of Kidney Renal Clear Cell Carcinoma. See Jupyter Notebook and summary slides.

Jupyter Notebooks

Jupyter notebooks were developed to explore kinase, ion channel, and GPCR clusters identified based on co-expression and co-regulation by transcription fcators. Common biological themes were identified and biological functions were inferred for under-studied potential target proteins.