/tools

Molecule-Generator is a Variational Autoencoder-based tool that generates synthetic SMILES strings for molecules composed of specific repeat units. It allows for the creation of a large dataset of molecular representations and facilitates the generation of new molecular structures through perturbation in the latent space.

QADD is a software tool designed for de novo drug design that utilizes iterative multi-objective deep reinforcement learning. It incorporates a graph-based molecular quality assessment model to generate high-quality drug-like molecules while considering their potential drug properties.

De-Novo-Drug-Design is a tool that utilizes Deep Reinforcement Learning to optimize the permeation of drugs across the blood-brain barrier. It aims to facilitate the design of new drug candidates by improving their molecular properties.

AutoGraph is a scalable autoregressive model designed for generating molecular graphs by flattening them into sequences. It achieves state-of-the-art performance on various molecular benchmarks and supports both unconditional and substructure-conditioned generation.

ML Conformer Generator is a tool that utilizes Equivariant Diffusion Models and Graph Convolutional Networks to generate novel 3D molecular conformations that adhere to specific shape constraints. It supports workflows in molecular design by allowing the generation of molecules that are chemically valid and spatially similar to reference structures.

This repository contains code for generating molecular conformers using SO(3) Averaged Flow-Matching and Reflow techniques. It provides a framework for efficient conformer generation, which is essential in molecular design and optimization.

Conformer-Search is a workflow designed for minimum energy conformer search of molecules using force field optimization methods like UFF and MMFF94. It allows users to import molecules from SMILES, generate conformers, optimize them, and select the minimum energy conformer for further analysis.

This repository contains a deep learning project aimed at reconstructing and generating molecules from low-dimensional representations. It utilizes Variational Autoencoders (VAEs) and employs the ZINC250K dataset for training and evaluation, making it a relevant tool for molecular design.

CATNIP is a tool designed to facilitate the prediction of enzyme compatibility with small molecules in biocatalysis. It utilizes machine learning models and a curated dataset to navigate between chemical and protein sequence spaces, aiming to streamline biocatalytic synthetic strategies.

moo-denovo is a tool for automated de novo design of drug-like molecular libraries, leveraging deep learning and multi-objective optimization techniques. It allows users to optimize molecular descriptors and generate new molecular structures based on specified criteria.

TAGMol is a framework for target-aware gradient-guided molecule generation, aimed at optimizing molecular properties for drug design. It includes functionalities for training models and evaluating generated molecules based on various criteria such as binding affinity and drug-likeness.

Graph2Token is a tool that utilizes graph neural networks to process molecular data, potentially for tasks such as classification and regression related to molecular properties. It involves pretraining a GNN encoder and fine-tuning for specific molecular tasks.

The nf-binder-design repository provides a Nextflow pipeline for designing protein binders using advanced molecular modeling techniques such as RFdiffusion and BindCraft. It automates the generation and optimization of protein structures, facilitating research in protein engineering and drug discovery.

ProteinTuneRL is a framework designed for optimizing protein sequences using infilling language models and reinforcement learning. It specifically supports antibody design by modifying regions of protein sequences to enhance properties like stability and binding affinity.

This repository contains code for predictive modeling of late-stage functionalization in chemistry using transfer learning techniques. It includes modules for training models, predicting regioselectivity of new molecules, and handling datasets relevant to molecular properties.

MCTS-RNA is a computational tool that solves the RNA inverse folding problem using Monte Carlo Tree Search. It allows for the design of nested and pseudoknot RNA structures while controlling the GC-content and its deviation precisely.

MaskedProteinEnT provides code to sample protein sequences using a contextual Masked EnTransformer. It allows for the design and generation of protein sequences, making it a useful tool in the field of molecular biology and protein engineering.

iupacGPT is a large-scale molecular pre-trained model that utilizes IUPAC nomenclature for property prediction and molecular generation. It employs transformer-based architectures to enhance the interpretability and performance of molecular tasks compared to traditional representations like SMILES.

Molecule-RNN is a recurrent neural network designed to generate drug-like molecules for drug discovery. It learns from a dataset and samples new molecules that resemble the training data, utilizing various tokenization methods for SMILES representation.

SecretoGen is a conditional autoregressive model that generates signal peptides based on the mature protein sequence and the host organism. It also includes functionality for evaluating the efficiency of these peptides, making it a specialized tool for protein design.

Progen is a Python implementation of a language model for generating protein sequences, based on the ProGen paper. It utilizes various protein sequence datasets for training and evaluation, making it a valuable tool for protein design and generation.

This repository contains materials for a workshop on AI in chemistry, covering topics such as molecular representations, data quality, and hands-on sessions for molecule generation. It aims to educate participants on the application of AI techniques in molecular design and analysis.

AMPTrans-lstm is a deep generative model application designed to discover novel and diverse functional peptides that can combat microbial resistance. It utilizes advanced machine learning techniques, specifically LSTM and transformer models, to generate and optimize peptide sequences.

MolCraftDiffusion is a unified generative-AI framework that facilitates the training and deployment of 3D molecular diffusion models for various molecular generation tasks. It supports property-targeted generation and provides tools for analyzing and optimizing generated molecules.