Data

Tools

Indexes

Applications

Experiments

Open Source Science

modelsmatter

https://github.com/alanhassen/modelsmatter

Science Score: 57.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ✓
    CITATION.cff file
    Found CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ✓
    DOI references
    Found 2 DOI reference(s) in README
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (12.3%) to scientific vocabulary
Last synced: 11 months ago · JSON representation ·

Repository

Basic Info
  • Host: GitHub
  • Owner: AlanHassen
  • License: mit
  • Language: Python
  • Default Branch: main
  • Size: 734 KB
Statistics
  • Stars: 8
  • Watchers: 3
  • Forks: 1
  • Open Issues: 0
  • Releases: 0
Created almost 3 years ago · Last pushed almost 2 years ago
Metadata Files
Readme Changelog License Citation

README.md

Models Matter

This package extends AiZynthFinder to seamlessly integrate various single-step retrosynthesis models, as illustrated in our papers Models Matter: The Impact of Single-Step Models on Synthesis Prediction and Mind the Retrosynthesis Gap: Bridging the divide between Single-step and Multi-step Retrosynthesis Prediction

Overview

Models Matter uses the default AiZynthFinder implementation and introduces the following enhancements:

  • Smiles-based Expansion Strategy: This feature allows the integration of any retrosynthesis model operating at smiles-level with AiZynthFinder, not depending on the underlying search algorithm.
  • ModelZoo Integration: The package includes ModelZoo, a dedicated framework that defines the single-step retrosynthesis approach within the smiles-based expansion strategy. Currently supported implementations are Chemformer, MHNreact, and LocalRetro.

Installation Procedure

To install the package, follow the sequential steps below:

```bash

Clone this repository and all submodules

git clone --recursive https://github.com/AlanHassen/modelsmatter

Create and initialize a conda environment

conda env create -f environments/environment.yml -n ssbenchmark conda activate ssbenchmark

Transition to the SSBenchmark directory

cd external/modelsmatter_modelzoo/

First the installation of the ModelZoo

poetry install

Navigate back to the models matter directory

cd ../..

Finalize the installation process

poetry install

Subsequently, install the appropriate single-step models or the necessary libraries.

```

Usage Guidelines

Adaptations to AiZynthFinder config are necessary to accommodate the different single-step models (examples provided in config/). The configurable settings include:

  • gpu_mode: Enable GPU mode for accelerated inference.

  • module_path: The path of the single-step retrosynthesis model repository.

  • model_path: The location of the trained model.

  • Additional parameters can be set, such as specifying the vocabulary for Chemformer via vocab_path.

Utilize the extended functionalities of AiZynthFinder through Models Matter for a comprehensive synthesis prediction experience.

Datasets & Models

Datasets and models that are not proprietary are available in the Models Matter figshare repository.

Funding

This study was partially funded by the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie Innovative Training Network European Industrial Doctorate grant agreement No. 956832 “Advanced machine learning for Innovative Drug Discovery”

Owner

  • Name: Alan Kai
  • Login: AlanHassen
  • Kind: user

Citation (CITATION.cff) 

# YAML 1.2
---
abstract: "We present the open-source AiZynthFinder software that can be readily used in retrosynthetic planning. The algorithm is based on a Monte Carlo tree search that recursively breaks down a molecule to purchasable precursors. The tree search is guided by an artificial neural network policy that suggests possible precursors by utilizing a library of known reaction templates. The software is fast and can typically find a solution in less than 10 s and perform a complete search in less than 1 min. Moreover, the development of the code was guided by a range of software engineering principles such as automatic testing, system design and continuous integration leading to robust software with high maintainability. Finally, the software is well documented to make it suitable for beginners. The software is available at http://www.github.com/MolecularAI/aizynthfinder."
authors:
  -
    family-names: Genheden
    given-names: Samuel
  -
    family-names: Thakkar
    given-names: Amol
  -
    family-names: "Chadimová"
    given-names: Veronika
  -
    family-names: Reymond
    given-names: "Jean-Louis"
  -
    family-names: Engkvist
    given-names: Ola
  -
    family-names: Bjerrum
    given-names: Esben
    orcid: "https://orcid.org/0000-0003-1614-7376"
cff-version: "1.1.0"
date-released: 2020-12-08
doi: "https://doi.org/10.1186/s13321-020-00472-1"
identifiers:
  -
    type: doi
    value: "10.1186/s13321-020-00472-1"
keywords:
  - retrosynthesis
  - casp
  - retrosynthesis
  - cheminformatics
  - "neural-networks"
  - "monte-carlo-tree-search"
  - "chemical-reactions"
  - astrazeneca
  - "reaction-informatics"
license: MIT
message: "If you use this software, please cite it using these metadata."
repository-code: "https://github.com/MolecularAI/aizynthfinder"
title: AiZynthFinder
version: "2.2.1"
...

GitHub Events

Total
  • Watch event: 2
Last Year
  • Watch event: 2