Giter Club home page Giter Club logo

molecular-generation's Introduction

Molecular-generation

Artificial Intelligence-Assisted Electrolyte Design to Improve Li+ Diffusion Coefficient

Introduction

The problem of electrolyte freezing and power loss on lithium batteries under low-temperature conditions requires immediate attention, especially applications to high altitude, high latitude regions and aerospace have been limited. Enhancing the Li+ diffusion coefficient represents a crucial direction for improvement. However, improving Li+ diffusion coefficient in existed electrolytes has proven to be challenging, exploring new solvents or additives in electrolyte may have great potential. In this study, we present a novel strategy that utilizes advanced techniques to design and find five novel molecules as additives with high Li+ diffusion coefficient. Our methodology involves two rounds of Molecular generation (MG) and Molecular dynamics (MD), and three rounds of Machine learning (ML), resulting in the highest Li+ diffusion coefficient of the generated additive being 3.96 times that of the benchmark system (1.72×10-11 m/s2). These findings are of great significance in improving Li+ diffusion coefficient in practical applications and addressing this persistent issue. Furthermore, these strategies and results provide valuable insights for further investigation into other challenges of lithium batteries.

Dependencies

rdkit==2022.03.3
scipy==1.8.0
numpy==1.21.5
pandas==1.4.1
scikit-learn==1.0.2
xgboost==1.5.0
shap==0.41.0
moses

Installation

conda create --name version python=3.8

Generative model GraphInvent

  1. git clone https://github.com/MolecularAI/GraphINVENT.git,
  2. pretraining and training the generative model using collected dataset,
  3. generated plenty of novel molecules,
  4. filtering the invalid and repeated molecules using scripts.

MD simulations

MD engine GROMACS 2020 and further optimized OPLS-AA force field are used for calculating Li+ diffusion coefficient.

Machine learning methods

Classification models were trained by Scikit-learn except XGBoost,
the features for machine learning models are obatined by python get_functional_group_features.py,
you can also define some important physical descriptors on the bulk and interface.

Synthesizable analysis

SA

A heuristic estimate of how hard (10) or how easy (1). it is to synthesize a given molecule. SA score is based on a combination of the molecule’s fragments contributions. And SA score threshold change from 6.0 to -4.5, SA score would yield similar results as SYBA, here, we set the threshold of SA score 4.5. SA was assembled in MOSES https://github.com/molecularsets/moses .

SYBA

SYBA is a fragment-based method for dividing the organic compounds into easy-synthesize (ES) or hard-to-synthesize (HS). It is based on their frequencies in the dataset of ES and HS. It is based on a Bernoulli naïve Bayes classifier that is used to assign SYBA score contributions to individual fragment based on the frequencies in the database of ES and HS. SYBA was trained previously on ES molecules in ZINC15 database and on HS molecules generated by the Nonpher methodology. And SYBA is publicly available at https://github.com/lich-uct/syba . Here, we set the threshold of SYBA score -18.6.

molecular-generation's People

Contributors

bujianbusan123-alt avatar

Watchers

 avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.