A project of the Molecular Modeling Lab @ UNC
In place of ChemBench, the MML group is actively developing a suite of open-source tools for drug discovery, toxicology, and cheminformatics. Below is our current collection of webservers and databases available for public use.
Predictive Models
STOPLIGHT
Drug-likeness Scoring Calculator
A webserver for generating a "goodness" score for early stage hits in drug discovery campaigns.
STopTox
Systemic and Topical chemical Toxicity
A machine learning app to predict the battery of acute toxicity tests classically known as the "6-pack".
PhaKinPro
Pharmacokinetic Property Calculator
A machine learning app to predict several human pharmacokinetic endpoints for small molecules.
LiabilityPredictor
A machine learning app to predict if a small molecule is at risk of interfering with or being a liability to an assay.
BANANA
BAsic NeurAl Network for binding Affinity
MPNN for predicting binding affinity of small molecules to protein pockets.
SALSA
Structurally Aware Latent Space Auto-encoder
A contrastive autoencoder for generating a latent space where chemicals of similar graph edit distance are near each other.
Databases
SMACC
Small Molecule Antiviral Compound Collection
A dataset of small molecules that have antiviral activity collected from ChEMBL.
MeDIC
Medicines, Diseases, Indications, and Contraindications
A database featuring highly curated lists of drugs, diseases, indications, contraindications, and additional metadata.
BigBind
A dataset built from CrossDocked and ChEMBL data.
Tools
ExEmPLAR
Extracting, Exploring, and Embedding Pathways Leading to Actionable Research
A user interface for exploring and analyzing biomedical knowledge graphs.
HEALER
Hit Expansion to Advanced Leads Using Enumerated Reactions
A computational chemistry platform designed to accelerate lead optimization by generating synthetically accessible chemical space.