Target Discovery from RNA-seq
Identify druggable targets from patient transcriptomics data
What This Workflow Does
This workflow takes RNA-seq data from patient cohorts and identifies druggable targets through multi-omics integration, network modeling, and patient stratification. It's designed for researchers with disease-vs-control expression data who want to discover novel therapeutic targets.
Input Requirements
- •Gene expression matrix (counts or TPM, genes × samples)
- •Sample metadata with case/control labels
- •Optional: clinical covariates, known confounders
Outputs
- →Ranked target list with druggability scores
- →Patient stratification clusters with molecular signatures
- →Network visualization of disease pathways
- →PDF report with supporting evidence
Workflow Steps
Upload RNA-seq Cohort Data
Upload gene expression data (count matrix or TPM) from patient cohorts. Supports GEO, TCGA, or custom datasets with case/control annotations.
Multi-Omics Embedding
AI integrates RNA-seq with protein-protein interaction networks, pathway databases, and known drug-target associations to build a disease signature model.
Network Modeling & Target Ranking
Graph neural networks identify disease-driving genes and rank targets by druggability, expression differential, and network centrality.
Patient Stratification
Unsupervised clustering identifies patient subgroups with distinct molecular profiles and predicted drug response patterns.
Validation & Deliverables
Cross-validation with independent datasets, literature evidence scoring, and generation of ranked target report with supporting data.
Example: NSCLC Target Discovery
Input: RNA-seq from 200 NSCLC tumors + 50 matched normal samples (TCGA)
Duration: 3.2 hours
Results:
- 127 ranked targets identified, 23 with druggability score > 0.8
- 4 patient subgroups with distinct driver pathways
- Top target (EGFR) validated against known NSCLC therapeutics
- 3 novel targets flagged for experimental validation