/tools

SMILES-GUI is a Streamlit interface designed for single-step retrosynthesis prediction from product SMILES. It employs PyTorch models for inference and RDKit for validation and visualization, allowing users to input SMILES, predict reactants, and export results.

RetroVAETransformer is a PyTorch-based model that utilizes a VAE-Transformer architecture to predict reactants from given product SMILES sequences in chemical reactions. It includes features like SMILES tokenization and autoregressive sampling, making it a useful tool for retrosynthesis in drug discovery.

The 'retrosynthesis-transformer' repository provides a TensorFlow-based Transformer model designed to predict reactant molecules from target molecules represented as SMILES strings. It includes tutorials for beginners to understand neural networks and how to implement sequence-to-sequence models in computational chemistry.

Retromix is a tool designed for exploring reaction space by integrating template-free and template-based approaches. It extends the capabilities of AiZynthFinder to identify and optimize molecular templates for target molecules, facilitating the design and generation of new compounds.

Retrosynthetic Planning is a tool that predicts the feasibility of retrosynthetic pathways using deep learning methods. It employs various scoring methods to evaluate potential synthetic routes for molecules.

This tool predicts reaction conditions for chemical reactions by modifying an existing model. It allows users to extract and export models in ONNX format and provides a straightforward way to check results using Python.

The Retrochemists Toolbox is designed to assist in the retrosynthetic analysis of chemical reactions, providing useful tools for chemists. It aims to facilitate the design and generation of molecules by analyzing potential synthetic pathways.

labmate-mcp is an AI lab companion that offers 81 tools for literature search, compound synthesis, analysis, and publication writing. It includes functionalities for predicting molecular properties, planning syntheses, and generating molecular data, making it a valuable resource for researchers in computational chemistry and molecular biology.

This repository presents a project on machine learning-assisted retrosynthesis planning, aiming to automate the design of synthetic routes for organic small molecules. It discusses various methodologies, including template-based and template-free approaches, to predict how target molecules can be synthesized from simpler starting materials.

BiRetroSys is a multi-step retrosynthesis planning system that includes modules for single-step prediction and multi-step search. It utilizes advanced algorithms to generate synthetic routes for target molecules based on their SMILES representations.

RetroCluster is a code suite designed for clustering large amounts of reaction data to aid in retrosynthesis planning. It processes and organizes chemical reactions based on conditions, mechanisms, and structures, providing interpretable outputs for further analysis.

This repository provides a toolkit for generative virtual screening that integrates seed-based molecule generation, QSAR predictions, and retrosynthesis. It utilizes Hugging Face transformer models to estimate reaction yields and facilitate molecular design and evaluation.