Abdallah Abou Hajal
Machine learning for drug discovery, and AI agents for academic publishing. Automation lead at Scifiniti, looking for a PhD.
14 papers 124 citations h-index 8
Screen, dock, simulate
- 1 Screen
- 2 Dock
- 3 Simulate
EGFR kinase domainthe pocket you are screening for
Drag to turn it, scroll to zoom
Three steps on data that was computed before you arrived. Screen: 302 real EGFR compounds from ChEMBL sit on a map, you have eight assays, and a model in your browser guesses the rest. Dock: your best compound drops into the EGFR pocket from crystal structure 1M17 and takes its AutoDock Vina pose. Simulate: two poses score the same, 2 ns of OpenMM molecular dynamics decides which one holds.
weaker stronger the model's guess
Real structure, real measurements, all computed before you arrived. How it was done.
- Protein: PDB 1M17, the EGFR kinase domain with erlotinib bound.
- Compounds: ChEMBL_37, target CHEMBL203, docking with AutoDock Vina, molecular dynamics with OpenMM.
- Nothing here calls a server or a language model.
Name the molecule
It draws itself atom by atom. Name it early for more points.