Explore projects

DeepFrag is a deep convolutional neural network that guides ligand optimization by extending a ligand with a molecular fragment, such that the resulting extension is also highly complementary to the receptor.

AutoGrow4 is an open-source program for generating novel drug-like molecules and optimizing preexisting ligands. It uses a genetic algorithm to evolve predicted ligands.

IMPORTANT NOTE: The latest version of AutoGrow (v4) can be found at http://git.durrantlab.com/jdurrant/autogrow4

A program for converting 2D SMILES strings to 3D models.

Scoria is a Python package for manipulating 3D molecular data.

LipidWrapper is a program for creating large-scale, curved membrane models of arbitrary size and geometry.

WISP is a trajectory analysis tool that calculates and visualizes allosteric pathways.

Easily import VMD scenes into Blender.

NNScore is a scoring function for characterizing the potency of receptor-ligand complexes. It is based on neural networks, computational models that simulate the microscopic organization of the brain.

POVME2 (POcket Volume MEasurer 2) scans a molecular-dynamics simulation and extracts druggable protein pockets with unique conformations. Virtual screens that leverage these conformations often identify novel active molecules.

PCAViz is an open-source Python/JavaScript toolkit for sharing and visualizing molecular dynamics trajectories via a web browser.

A Blender plugin for importing a molecular dynamics trajectory, to take advantage of Blender's advanced rendering features.

NNScore is a scoring function for characterizing the potency of receptor-ligand complexes. It is based on neural networks, computational models that simulate the microscopic organization of the brain.

Dimorphite-DL is a fast, accurate, accessible, and modular open-source program for enumerating small-molecule ionization states.

BINANA analyzes docked ligand poses to identify molecular interactions that contribute to binding. Accurately characterizing these interactions allows medicinal chemists to assess whether a predicted ligand merits further study.

LigGrep is a program for identifying docked poses that participate in user-specified receptor/ligand interactions. It evaluates each docked pose and outputs the names of the compounds with poses that pass all filters.

Prot2Prot is a deep-learning model trained to imitate a Blender/BlendMol-rendered image given a much simpler representation (“sketch”) of a protein surface.

An easy to use, browser-based GUI for running DeepFrag. The goal is to suggest useful chemical changes to a small-molecule ligand that might improve its binding to a protein receptor (i.e., to aid lead optimization).

SubPEx (Sub-Pocket Explorer) is a tool to enhance ensemble (multiple-receptor/relaxed-complex) virtual screening. It uses weighted ensemble path sampling and molecular dynamics simulations to accelerate binding-pocket sampling.