/tools

The DeepLearningExamples repository by NVIDIA provides state-of-the-art deep learning scripts that can be utilized for various applications, including drug discovery. It offers easy-to-train and deploy models that leverage NVIDIA's deep learning software stack.

The DeepMind Research repository contains implementations and code for various research projects, including those focused on molecular simulations and protein structure prediction. It aims to accelerate scientific progress by providing tools and datasets for the research community.

AlphaFold is an open-source implementation of a deep learning model that predicts protein structures from amino acid sequences. It utilizes advanced algorithms to provide accurate predictions, significantly aiding in molecular biology research and applications.

Awesome-Diffusion-Models is a collection of resources and papers on diffusion models, primarily in the context of machine learning and generative modeling. It includes tutorials, papers, and resources related to the theory and applications of diffusion models across various domains, but lacks a direct focus on molecular tools.

Claude Scientific Skills is a collection of 140 ready-to-use scientific skills that enable users to perform complex workflows in drug discovery and cheminformatics. It includes functionalities for molecular property prediction, virtual screening, and molecular docking, making it a valuable resource for researchers in computational chemistry and molecular biology.

AlphaFold 3 is an inference pipeline that allows users to predict the three-dimensional structures of proteins based on their amino acid sequences. It leverages advanced machine learning techniques to provide accurate predictions, which are essential for understanding biomolecular interactions and functions.

DeepChem provides an open-source toolchain that facilitates the application of deep learning in drug discovery, quantum chemistry, and biology. It supports various molecular tasks such as property prediction, molecular generation, and offers extensive tutorials for users to learn and apply these techniques.

NVIDIA Warp is a Python framework designed for accelerated simulation and spatial computing, enabling high-performance physics simulations and data generation. It supports differentiable programming, making it suitable for integration into machine learning pipelines, particularly in the context of molecular simulations.

The 'esm' repository provides pretrained transformer models for proteins, enabling tasks such as protein design and structure prediction. It includes tools for generating protein structures from sequences and predicting the effects of mutations, making it a valuable resource in molecular biology.

Boltz is a family of models designed for predicting biomolecular interactions, including binding affinities. It aims to provide accurate in silico screening for drug discovery, leveraging deep learning techniques to enhance molecular design processes.

RDKit is a comprehensive cheminformatics library that provides tools for molecular property prediction, descriptor generation, and various molecular operations. It supports both small molecules and proteins, making it a versatile tool for researchers in computational chemistry and molecular biology.

The 'Awesome AI for Science' repository is a curated collection of resources related to AI applications in various scientific fields, particularly focusing on drug discovery and molecular design. It includes models, datasets, and frameworks that facilitate the prediction and generation of molecular properties and structures.

PennyLane is an open-source quantum software platform designed for quantum computing, quantum machine learning, and quantum chemistry. It provides tools for building quantum circuits and algorithms, with applications in molecular property prediction and simulations.

LAMMPS is a classical molecular dynamics simulation code that efficiently runs on parallel computers. It is used for simulating a wide range of molecular systems, including small molecules and proteins, and is maintained as an open-source project.

RFdiffusion is an open-source method for generating protein structures, allowing for both unconditional and motif-based design. It supports various protein design challenges, including binder design and scaffold generation, making it a valuable tool in molecular biology and computational chemistry.

Biomni is a general-purpose biomedical AI agent that enhances research productivity by integrating large language model reasoning with planning and execution. It can predict molecular properties, generate hypotheses, and evaluate biological reasoning tasks, making it a versatile tool in the biomedical field.

Chemprop is a machine learning package that utilizes message passing neural networks to predict various molecular properties. It is particularly useful in drug discovery, enabling researchers to assess properties such as ADMET and binding affinity.

RoseTTAFold is a deep learning package that provides models and scripts for predicting protein structures and their interactions. It utilizes a three-track neural network architecture to enhance the accuracy of these predictions, making it a valuable tool in molecular biology and computational chemistry.

The ESM repository provides flagship generative models for protein sequences, allowing users to generate and predict protein structures and functions. It includes tools for both protein design and representation learning, making it a valuable resource in molecular biology.

MMseqs2 is an ultra-fast and sensitive software suite for searching and clustering large sets of protein and nucleotide sequences. It significantly improves the speed and sensitivity of sequence searches compared to traditional methods like BLAST, making it a valuable tool for molecular biology research.

FAIR Chemistry's `fairchem` library offers a centralized repository of machine learning models and methods tailored for chemistry and materials science. It supports various tasks such as predicting molecular properties, running molecular dynamics simulations, and optimizing molecular structures.

This repository is a curated list of papers that explore the use of deep learning techniques in protein design. It includes resources on benchmarks and datasets relevant to the field, making it a valuable tool for researchers in computational biology.

Chai-1 is a multi-modal foundation model for predicting the structures of various biomolecules, including proteins, small molecules, DNA, and RNA. It utilizes advanced techniques to achieve state-of-the-art performance in molecular structure prediction across multiple benchmarks.

DeePMD-kit is a deep learning package designed to create models for interatomic potential energy and perform molecular dynamics simulations. It interfaces with various deep learning frameworks and classical molecular dynamics packages, making it suitable for a wide range of molecular systems.