/tools

AbExpress is a tool for assessing and optimizing the expressability of antibodies using Long-Short Term Memory (LSTM) models. It includes functionalities for training models, predicting expressability, and designing antibody structures based on engineered sequences.

The PROPHET-Ab repository provides supporting material for a platform that assesses the developability of antibodies using high-throughput methods. It facilitates AI/ML model training, which is essential for predicting molecular properties and optimizing antibody design.

RAAD is a tool for designing and optimizing antibodies using relation-aware equivariant graph networks. It includes functionalities for antigen-binding CDR-H3 design and affinity optimization, making it a valuable resource in the field of drug discovery.

LlamaAffinity is a model designed for predicting the binding affinity between antibodies and antigens using a LLaMA3 transformer backbone. It employs 5-fold cross-validation for robust evaluation and provides various performance metrics to assess its predictive capabilities.

This repository provides a framework for predicting the binding affinities of antibody-antigen complexes by interpreting protein structures as graphs. It utilizes graph neural networks and includes datasets for training and validation, making it a valuable tool for molecular property prediction.

The diffab-pytorch repository provides an unofficial re-implementation of a diffusion-based generative model aimed at designing and optimizing antigen-specific antibodies. It includes data preprocessing scripts for handling antibody-antigen structures, making it a valuable tool for researchers in protein design.

Locuaz is an antibody optimization platform that utilizes DLPacker-based mutators for designing and generating antibody sequences. It integrates molecular dynamics tools and provides a framework for optimizing antibody structures.

LICHEN is a tool for generating light-chain immunoglobulin sequences conditioned on heavy-chain sequences and experimental needs. It allows for customization of sequence generation based on various parameters and includes filtering options to ensure the generated sequences meet specific criteria.

The PIA-KRASv2-Nb repository provides a computational framework for designing a high-affinity nanobody against the KRAS oncoprotein. It includes detailed validation of binding affinity and specificity through molecular dynamics simulations and structural modeling.

SE3Bind is a tool designed for predicting the binding affinity of antibody-antigen complexes using an SE(3)-equivariant model. It includes functionalities for training models, generating datasets, and performing docking simulations, making it a valuable resource in molecular biology and computational chemistry.

The 'protstruc' repository offers a Python package for easy handling and manipulation of protein structures, specifically designed for deep learning applications. It includes features for computing geometric properties, selecting specific chains, and diffusing atom coordinates, making it a valuable tool for protein design and analysis.

The deepab-pytorch repository provides an unofficial re-implementation of DeepAb, a deep learning model designed for predicting the structure of antibodies. It allows users to generate coarse-grained structural predictions based on antibody sequences, making it a valuable tool in the field of computational biology.

DeepSCAb is a tool for predicting the structure of antibodies by utilizing deep learning techniques to improve the accuracy of side chain conformations. It provides code, data, and weights for non-commercial use, facilitating advancements in antibody design.

TADA is an AI-powered pipeline designed for the discovery and optimization of therapeutic antibodies, integrating target identification and in silico antibody design. It utilizes multi-source data and advanced modeling techniques to streamline the drug development process.

This repository develops a toolkit for characterizing the maturation pathways of broadly neutralizing antibodies through reinforcement learning techniques. It aims to enhance understanding of antibody-mediated immunity and optimize vaccination protocols against rapidly mutating pathogens.

TeBaAb is a framework for redesigning antibodies conditioned on antigen sequences and textual descriptions, utilizing generative modeling and directed evolution. It includes a dataset of antibody-antigen pairs and aims to improve binding affinity while maintaining structural integrity.

This repository contains code that analyzes and compares the binding sites of antibodies and single-domain antibodies, providing insights into their structural differences and implications for biotherapeutics. It utilizes sequence and structural datasets to explore interactions between these molecules.

The igfold-pytorch repository provides an unofficial re-implementation of IgFold, a method for predicting the structure of antibodies using deep learning techniques in PyTorch. It leverages embedding vectors and attention matrices to generate accurate structural predictions for antibodies.

This repository provides a computational pipeline for enhancing antibody-antigen binding affinity using AI-based mutation generation and structural modeling. It employs tools like ProteinMPNN and FoldX to predict binding affinities and evaluate antibody variants, making it a valuable resource for antibody engineering.

This repository contains supporting material and code for conducting large-scale molecular dynamics simulations of CDR loops in antibodies and TCRs. It includes Jupyter notebooks for analysis, datasets, and scripts for setting up and running simulations.

DeepBioLab is an interactive platform that explores AI algorithms with applications in biology, including projects related to antibody design. It serves as a collaborative space for researchers and developers interested in the intersection of AI and biological sciences.

This repository provides code for regenerating figures related to the computational design of nanobodies targeting the Ras-binding domain of ELMO1. It is focused on protein-protein interactions and the design of antibodies, which are key aspects of molecular biology and computational chemistry.

The Antibody Evolution Wizard is a PyMOL plugin that facilitates machine learning-based antibody evolution by integrating tools for evaluating binding affinity and generating mutation suggestions. It allows users to track mutations and their effects on binding free energy, making it a valuable resource for antibody design.

This repository provides a Nextflow pipeline for reconstructing antibody repertoires from RNA-seq data, filtering viable chains, and optionally modeling antibody structures. It includes features for pairing heavy and light chains and outputs high-confidence sequences for further analysis.