/tools

ProteinGym is a comprehensive repository of benchmarks for Deep Mutational Scanning (DMS) assays, allowing for the evaluation of various mutation effect predictors. It includes extensive datasets of clinical variants and experimental measurements, facilitating research in protein design and mutation effect prediction.

The Rosetta Bio-macromolecule modeling package is a comprehensive suite for computational modeling and analysis of protein structures. It includes algorithms for de novo protein design, enzyme design, and ligand docking, facilitating significant advancements in computational biology.

ProteinGenerator is a tool that generates sequence-structure pairs using a diffusion model based on RoseTTAFold. It allows users to explore and create new protein sequences conditioned on structural motifs, facilitating advancements in protein design.

The SE3-Transformer-PyTorch repository provides an implementation of SE3-Transformers for equivariant self-attention, specifically aimed at applications in protein structure prediction and drug discovery. It allows for the modeling of molecular interactions and features, making it a valuable tool in computational chemistry and molecular biology.

The DPLM repository provides implementations of diffusion protein language models that excel in generating and predicting protein sequences and structures. It includes features for unconditional and conditional protein generation, as well as representation learning for various protein-related tasks.

ProstT5 is a bilingual language model designed for translating between protein sequences and their corresponding 3D structures. It utilizes advanced machine learning techniques to derive embeddings and facilitate the understanding of protein structures from sequences.

Mol-Instructions is a dataset that contains a large collection of instructions for biomolecular tasks, including molecule-oriented and protein-oriented tasks. It aims to facilitate the development of large language models for generating and understanding molecular and protein-related information.

Equiformer is a PyTorch implementation of an SE3/E3 equivariant attention network that achieves state-of-the-art results in protein folding. It utilizes advanced techniques in deep learning to model molecular interactions and predict protein structures.

RFDesign is a tool for protein hallucination and inpainting using the RoseTTAFold framework. It allows for the generation of protein structures, which is a key aspect of molecular design and computational biology.

ProteinFlow is an open-source Python library that streamlines the pre-processing of protein structure data for deep learning applications. It enables users to filter, cluster, and generate datasets from protein structure databases, facilitating various protein design tasks.

La-Proteina is a tool for the joint generation of protein amino acid sequences and their full atomistic structures. It utilizes a partially latent flow matching model to effectively design proteins, achieving state-of-the-art performance in various generation benchmarks.

OpenComplex is an open-source platform designed for developing protein and RNA complex models, leveraging features from AlphaFold 2 and OpenFold. It allows for high-precision modeling and inference of RNA and protein-RNA complexes, making it a valuable tool in computational biology.

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.

Ankh is an optimized protein language model that enhances general-purpose modeling for protein engineering. It offers pre-trained models and datasets for various protein-related tasks, including secondary structure prediction and solubility assessment.

This repository provides scripts for evolving human antibodies based on general protein language models. It allows users to recommend mutations to antibody sequences and reproduce analyses from a related research paper.

ProLLaMA is a multitask protein language model that facilitates protein generation and understanding. It employs an Evolutionary Protein Generation Framework to ensure generated proteins are biologically viable while also predicting properties such as solubility and superfamily classification.

PocketGen is a tool that generates full-atom ligand-binding protein pockets using generative models. It benchmarks its performance against established datasets like CrossDocked and Binding MOAD, providing processed datasets for training and evaluation of pocket generation methods.

ProteinInvBench is an open-source project that benchmarks structure-based protein design methods. It integrates various models, datasets, and evaluation metrics into a unified framework, facilitating the analysis and development of protein design algorithms.

Genie 2 is a protein structure diffusion model designed for unconditional protein generation and motif scaffolding. It includes training and inference code, allowing users to generate diverse protein structures and evaluate their designability and novelty.

Genie is a tool for de novo protein design that utilizes equivariant diffusion models to generate protein structures. It includes functionalities for training models, sampling generated domains, and evaluating the results against established benchmarks.

PiFold is a tool designed for effective and efficient protein inverse folding, generating protein sequences that fold into specified structures. It employs novel features and a graph neural network approach to enhance the accuracy and speed of protein design.

IgGM is a generative foundation model aimed at antibody design, enabling the creation of novel antibodies and nanobodies through various design tasks such as affinity maturation and structure prediction. It provides tools for designing sequences based on given frameworks and predicting their structures.

ParallelFold accelerates the AlphaFold protein structure prediction process by dividing the computational workload between CPU and GPU, allowing for faster predictions of multiple protein structures. It is particularly useful for researchers in molecular biology looking to enhance their structural predictions.

ImmuneBuilder is a deep learning tool designed to predict the structures of immune receptor proteins, including antibodies, nanobodies, and T-cell receptors. It offers state-of-the-art accuracy and speed in structure prediction, making it a valuable resource for researchers in molecular biology and biotherapeutics.