Three research themes, from atomic protein structure to population-scale disease.
Understanding how proteins move, fold, and interact at atomic resolution, using molecular dynamics and structural modeling.
The blood-brain barrier tightly controls what enters the brain. We model how molecular transporters work at this barrier to help design drugs that actually reach their target.
Cytokines are signaling proteins that direct immune responses. We measure the physical forces behind cytokine-receptor binding to guide the design of better immunotherapies.
A small molecule binding at one end of a protein can silence gene activity at the other. We study how these long-range signals travel through proteins like DgoR, revealing how bacteria sense and adapt to their environment.
Building AI systems that learn the rules behind effective drugs, and can explain those rules to scientists.
We use explainable AI to discover peptides, small protein fragments, that disrupt disease pathways. The system learns from each experiment and explains its reasoning, producing design rules a biologist can actually use.
Hepcidin is a small hormone that controls iron levels in the blood. Too much or too little causes conditions like anemia and iron overload, both common in India. We design molecules that restore this balance.
Combining genomic, transcriptomic, and clinical data to find disease signatures, predict patient risk, and automate diagnostics. Clinical partner: PGIMER, Chandigarh.
Heart attacks leave molecular traces across DNA, RNA, and proteins. We combine these signals to build a diagnostic panel that predicts coronary artery disease severity more precisely than any single marker alone.
ALS is a fatal motor neuron disease with no cure. We use AI to identify genes that are structurally constrained in ALS, not merely changed in expression, making them stronger candidates for therapeutic rescue.
Diagnosing blood disorders from bone marrow biopsies is slow and often subjective. AISAP automates the analysis of whole-slide images, classifying cell types and flagging abnormalities to standardize and speed up diagnosis.
Most cancer mutation databases are built from Western populations. We identify and validate variants relevant to Indian genetic backgrounds, starting with familial breast cancer cohorts.