/tools

RGN2 is a reference implementation of a recurrent geometric network designed for predicting the 3D structures of orphan or de novo protein sequences. It utilizes deep learning techniques to enhance protein structure prediction, making it a valuable tool in the field of molecular biology.

DrugHIVE is a software tool that implements a deep hierarchical variational autoencoder for structure-based drug design. It allows for the generation and optimization of ligands, making it a valuable resource in the field of molecular design and drug discovery.

LocalRetro is a tool for predicting retrosynthetic pathways for organic molecules using machine learning. It implements a model that derives local reaction templates and predicts reactants based on given products, facilitating the design and generation of new molecules.

AF3Score is a pipeline for evaluating the quality of protein structures using metrics derived from AlphaFold3. It generates various scoring metrics to assess the accuracy of protein models, focusing on both global and local structural features.

AIMNet is a neural network potential that accurately predicts chemical properties using an atoms-in-molecules approach. It provides a framework for molecular dynamics simulations and benchmarks for evaluating performance in property prediction.

PepGLAD is a tool for full-atom peptide design that utilizes geometric latent diffusion models to co-design peptide sequences and structures. It supports binding conformation generation and provides datasets for benchmarking the models.

The 'arnie' repository provides a Python API for estimating and comparing RNA energetics across various secondary structure algorithms. It facilitates structure prediction and analysis of RNA sequences, making it a useful tool in molecular biology.

DelFTa is an open-source toolbox designed for predicting quantum-mechanical properties of drug-like molecules using machine learning techniques. It employs advanced models to approximate DFT reference values and supports various molecular file formats for easy integration into workflows.

SELFormer is a molecular representation learning tool that utilizes SELFIES language models to generate high-quality molecular embeddings. It is pre-trained on drug-like compounds and fine-tuned for various molecular property prediction tasks, making it a valuable resource for drug discovery and cheminformatics.

Psikit is a wrapper library for Psi4 and RDKit that facilitates quantum chemistry calculations, including the prediction of molecular properties such as HOMO and LUMO energies, structure optimization, and charge calculations. It is designed to assist in molecular design and cheminformatics tasks.

Awesome-Molecular-Docking is a curated list of resources aimed at solving molecular docking and related tasks. It includes software for docking, datasets, and references to molecular dynamics simulations, making it a valuable tool for researchers in drug discovery and molecular biology.

QUBEKit is a toolkit for deriving bespoke force fields from quantum mechanical calculations, enabling users to automate the parameterization of molecular mechanics. It supports various molecular types and provides functionalities for both single and bulk analyses.

ResGen is a tool for generating 3D molecular structures that are aware of their binding pockets, utilizing parallel multi-scale modeling. It is designed for molecular generation tasks, particularly in the context of drug discovery and protein-ligand interactions.

MoleculeNet is a collection of datasets and benchmarks designed for evaluating machine learning models in the context of molecular property prediction. It includes various datasets relevant to physical chemistry, biophysics, and materials science, facilitating research in drug discovery and molecular modeling.

APPT is a state-of-the-art model designed to predict protein-protein binding affinity using transformer architectures and the ESM protein language model. It supports drug discovery and protein engineering by providing precise predictions based on protein sequence pairs.

BitBIRCH-Lean is a memory-efficient implementation of the BitBIRCH clustering algorithm designed for high-throughput clustering of large molecular libraries. It allows users to generate molecular fingerprints from SMILES files and cluster them efficiently, making it a valuable tool for drug discovery and cheminformatics applications.

The ProSE repository provides multi-task and masked language model-based protein sequence embedding models. It allows users to train models and embed protein sequences, facilitating research in protein structure and function analysis.

QCSchema is a schema designed to standardize the representation of quantum chemistry data, allowing for easier integration and manipulation of data from various quantum chemistry packages. It aims to streamline workflows by providing a consistent format for output variables, making it easier to work with complex quantum chemistry computations.

OVO is an open-source ecosystem designed for de novo protein design, integrating models, workflows, and data management. It provides a scalable platform for scaffold and binder design, utilizing advanced methods for protein structure generation and validation.

MMseqs2-App is a graphical interface for the MMseqs2 and Foldseek software suites, enabling users to search and annotate large sequence and structure datasets. It supports interactive data exploration and can be deployed on servers or workstations, facilitating molecular biology research.

PPIRef is a Python package designed for working with 3D structures of protein-protein interactions. It includes functionalities for extracting, visualizing, and analyzing protein-protein interfaces, as well as providing a dataset for machine learning applications in this domain.

FlowSite and HarmonicFlow are tools for generating binding structures for single and multi-ligands, as well as designing binding site residues. They utilize advanced generative models to optimize molecular interactions, making them valuable for applications in drug discovery and protein design.

This repository provides a graph neural network implementation for predicting molecular properties based on 3D structures. It allows users to preprocess datasets and train models to predict various molecular properties using a subset of the QM9 dataset.

MEAN is a tool for conditional antibody design utilizing a multi-channel equivariant attention network. It provides functionalities for redesigning antibody CDRs and optimizing binding affinities, making it a valuable resource in the field of molecular design and drug discovery.