About this application

Learn more about the 3D molecular clustering application

Application overview

This application provides an interactive 3D visualization of filtered weakly solvating electrolytes (WSEs) based on their chemical structure and properties. The system uses Density Functional Theory(DFT), Molecular Dynamics(MD) calculation and deep-learning model to obtain properties of molecules, exploits K-means clustering to group molecules into 4 clusters with similar characteristics, enabling researchers to identify patterns and relationships in molecular data.

The application was developed to support research in materials science and computational chemistry, particularly providing an intuitive insight for designing proper WSEs for lithium metal batteries.

3D visualization

Interactive 3D scatter plots of molecular clusters with full rotation and zoom capabilities.

Similarity search

Find molecules similar to a query using SMILES strings and adjustable similarity thresholds.

Property analysis

Detailed property visualization and statistical analysis for each molecule and cluster.

Technology stack
Backend technologies
  • Python Flask Web framework
  • RDKit Cheminformatics
  • Scikit-learn Machine learning
  • Pandas & NumPy Data processing
Frontend technologies
  • HTML5 & CSS3 Markup & styling
  • JavaScript (ES6+) Interactivity
  • Bootstrap 5 UI framework
  • Plotly.js Visualization
How to use
1
Explore the 3D visualization

Use your mouse to rotate, zoom, and pan the 3D scatter plot. Click on any point to view detailed information about that molecule.

2
Search for similar molecules

Enter a SMILES string in the search box and adjust the similarity threshold to find molecules with similar structures.

3
Analyze cluster properties

Use the data explorer page to view statistical information about each cluster and their properties.

4
Download data

Download the complete dataset or individual clusters for further analysis in external tools.

Research background

This application is part of a research project focused on developing weakly solvating electrolytes for lithium metal betteries. The clustering approach helps identify groups of molecules with similar electronic properties, which is valuable for materials design and discovery.

The methodology involves:

  • Calculating molecular descriptors from SMILES representations
  • Applying dimensionality reduction techniques (PCA)
  • Clustering molecules using K-means algorithm
  • Visualizing results in interactive 3D space

If you use this application in your research, please cite:

None NoneNone

Contact Information
Project Lead

Dr. Xiang Chen
xiangchen@mail.tsinghua.edu.cn
Department of Chemical Engineering
Tsinghua University