/tools

ToolUniverse is a platform that democratizes the creation of AI scientists by integrating a wide range of machine learning models, datasets, and scientific tools. It enables users to perform tasks related to molecular property prediction, molecular design, and scientific workflows, making it a versatile tool in computational chemistry and molecular biology.

This repository provides a comprehensive list of papers and resources related to molecular and material design using generative AI and deep learning techniques. It covers various methodologies for drug design, molecular optimization, and includes datasets and benchmarks relevant to the field.

BoltzGen is a tool for designing and generating protein structures based on specified design criteria. It utilizes advanced algorithms to produce ranked sets of protein designs, facilitating the exploration of novel molecular architectures.

AIRS is an open-source collection of software tools and datasets focused on artificial intelligence applications in quantum, atomistic, and continuum systems. It includes resources for predicting molecular properties, designing molecules, and conducting simulations, making it highly relevant to the fields of computational chemistry and molecular biology.

REINVENT4 is an AI molecular design tool that focuses on generating and optimizing small molecules through techniques like reinforcement learning. It supports various design tasks such as scaffold hopping and R-group replacement, making it a valuable resource for drug discovery.

DiffSBDD is an implementation of an equivariant diffusion model aimed at structure-based drug design. It allows users to generate new ligands, optimize existing molecules, and benchmark performance using various datasets, making it a comprehensive tool for molecular design and analysis.

RFantibody is a pipeline for the structure-based design of de novo antibodies and nanobodies. It utilizes methods for protein backbone design, sequence design, and in silico filtering of designs, making it a comprehensive tool for antibody engineering.

ReLeaSE is a tool that utilizes deep reinforcement learning to facilitate the de-novo design of drug molecules. It includes functionalities for optimizing molecular properties and generating new molecular structures based on learned patterns.

BoFire is a Bayesian optimization framework designed for real experiments, particularly in the chemical and pharmaceutical industries. It supports optimization of molecular properties and experimental designs, utilizing advanced techniques like multi-objective optimization and active learning.

PyXtal is an open-source Python library that facilitates the generation of atomic and molecular structures while adhering to symmetry constraints. It includes features for geometry optimization and supports various structural formats for further analysis.

GAUCHE is a library that facilitates probabilistic modeling and optimization techniques tailored for molecular representations. It includes various kernels for molecules, chemical reactions, and proteins, enabling users to perform tasks such as property prediction and Bayesian optimization.

The 'mosaic' repository provides a framework for functional, multi-objective protein design using continuous relaxation and machine learning models. It allows users to combine various predictors to optimize protein properties such as binding affinity and solubility, making it a valuable tool for computational biology and molecular design.

Bio-Diffusion is a geometry-complete diffusion generative model designed for generating and optimizing 3D molecular structures. It allows for both unconditional and property-conditional generation of small molecules, making it a valuable tool in molecular design and optimization.

DrugEx is a deep learning toolkit designed for scaffold-constrained drug design using graph transformer-based reinforcement learning. It allows users to generate novel drug molecules by optimizing multiple objectives, making it a valuable tool in the field of drug discovery.

geomeTRIC is a geometry optimization code that facilitates the optimization of molecular structures using various quantum chemistry and molecular mechanics software. It supports multiple external quantum chemistry packages and provides a command line interface for executing optimizations.

Fragmenstein is a tool that merges and links compounds by stitching them together based on atomic overlap, allowing for the generation of new molecular conformers. It also places follow-up molecules in relation to parent compounds, facilitating molecular design and optimization in drug discovery.

QUICK is an open-source software package designed for ab initio quantum chemistry calculations, utilizing GPU acceleration for efficient electronic structure calculations. It supports various methods including Hartree-Fock and density functional theory, and is capable of performing geometry optimizations and QM/MM simulations.

CodonTransformer is a deep learning tool for optimizing DNA sequences to enhance protein expression in various organisms. It leverages a large dataset of DNA-protein pairs to predict and generate host-specific DNA sequences with optimized codon usage.

Coscientist is a tool designed for autonomous chemical research, leveraging large language models to assist in synthesis planning and optimization problems. It includes various datasets and implementations that support molecular design and generation tasks.

Auto3D is a tool that automatically generates low-energy 3D molecular conformers from SMILES or SDF input using neural network potentials. It includes features for tautomer enumeration, stereoisomer generation, and geometry optimization, making it useful for molecular design and optimization tasks.

MolGen is a tool for domain-agnostic molecular generation that incorporates chemical feedback to optimize molecular properties. It supports de novo molecule generation and fine-tuning for specific properties like QED and logP, making it valuable for molecular design and optimization tasks.

The ReinventCommunity repository provides a collection of Jupyter notebook tutorials for using REINVENT 3.2, focusing on molecular design through reinforcement learning and QSAR modeling. It includes examples of data preparation, model building, and various reinforcement learning scenarios to generate and optimize novel compounds.

ChemTS is a software tool that utilizes Monte Carlo Tree Search combined with neural networks to design novel molecules with desired properties such as HOMO-LUMO gap and logP. It also allows for the design of molecules that are active against target proteins, facilitating advancements in drug discovery.

DrugEx is an open-source software library designed for the de novo design of small molecules using deep learning generative models within a multi-objective reinforcement learning framework. It provides various generator architectures and scoring tools to facilitate the generation and optimization of drug-like compounds.