Computational Chemist | Cheminformatics & Molecular AI | Drug Discovery
Structure- and ligand-based drug design, cheminformatics, and machine learning, with 15+ years in research. Most recently: ML-guided hit discovery (Leishmania G6PD, Chemoinformatics & Molecular AI core facility, University of Marburg) — virtual screening through custom-synthesized analogs and SAR-driven optimization with experimental partners. h-index 14, 620+ citations; co-author on four 2025–2026 papers (Nat Commun, PLOS Pathogens, iScience, J Clin Invest).
Google Scholar · ORCID · jshamsara@yahoo.com
- mlmolprop — Cheminformatics and QSAR toolkit: molecule preparation, descriptors and fingerprints, leakage-free preprocessing, and one interface over ~30 regression and classification models. Tested (168 tests), CI on every push, semantically versioned.
- FEP_OpenFE — Free-energy perturbation (OpenFE/OpenMM) reproducing experimental binding-affinity trends for PIM1 kinase inhibitors.
- Predictive_DT-NN_MR1-target — Decision-tree/neural-net classifiers for MR1 covalent-ligand binders (J Mol Struct, 2020; 3 Biotech, 2023).
- Gene-expression-analysis — Supervised/unsupervised ML on TCGA-PRAD expression data, identifying SPAG1/PLEKHF2 in prostate-cancer lymph-node extension (Genomics, 2020).