- ConsScoreTK: Proprietary code. Creation of receptor-specific machine learning-based scoring functions for drug design, statistical evaluation, reporting and visualization. The toolkit consists of 4 major components: 1) automated ligand binding pose prediction (HomoLigAlign algorithm); 2) binding affinity prediction for ligand scaffold optimization (deepScaffOpt algorithm); 3) virtual screening and hit identification (deepHitMiner algorithm); 4) off-target identification and drug repurposing (in progress).
- Neural Graph Fingerprint: an implementation of Convolutional Networks on Graphs for Learning Molecular Fingerprints in Python 3 using Keras 2.x and TensorFlow 1.13.x.
- 4D-CHAINS: fully automated protein NMR assignments from two 4D spectra.
- OFF-TARGET PIPELINE: an integrated workflow for off-target identification and other applications in bioinformatics and drug-discovery.
- TRUNCATE TRAJECTORY: plugin for VMD.