Related Experiment Video
Updated: May 13, 2026
06:29
Workflow and Tools for Crystallographic Fragment Screening at the Helmholtz-Zentrum Berlin
Published on: March 3, 2021
5.4K
Modeling High Energy Molecules and Screening to Find Novel High Energy Candidates
Mazal Rachamim1, Abraham J Domb2, Amiram Goldblum1
1Molecular Modelling and Drug Design Lab, Institute for Drug Research and Fraunhofer Project Center for Drug Discovery and Delivery, Faculty of Medicine, The Hebrew University of Jerusalem, Jerusalem 91905, Israel.
ACS Omega
|October 28, 2024
Summary
The Iterative Stochastic Elimination (ISE) algorithm effectively identifies novel high-energy materials (HEMs). This machine learning approach accelerates the discovery of potent compounds for various applications.
Area of Science:
- Materials Science
- Computational Chemistry
- Machine Learning
Background:
- High energy materials (HEMs) are crucial for military and commercial applications due to their significant energy release.
- Traditional experimental methods for discovering HEMs are time-consuming and resource-intensive.
- Machine learning (ML) offers a promising avenue to accelerate the identification of novel high-energy compounds.
Purpose of the Study:
- To apply the in-house Iterative Stochastic Elimination (ISE) algorithm for the discovery of high-energy materials (HEMs).
- To evaluate the efficacy of ISE in identifying potent HEM candidates by analyzing physicochemical properties.
- To screen a large chemical database for novel HEMs using a refined ISE model.
Main Methods:
- Development and application of two distinct Iterative Stochastic Elimination (ISE) models (Model A and Model B) incorporating known HEMs and non-HEMs.
- Creation of a merged model (Model C) using all active molecules from Models A and B for enhanced predictive power.
- Screening of approximately 2 million molecules from the Enamine database using Model C to identify top-scoring HEM candidates.
Main Results:
- Model A successfully identified 69% of active molecules from Model B, with 62% achieving the highest scores.
- Model B identified 80% of active molecules from Model A, with 61% receiving the highest scores.
- Model C, trained on 261 active molecules, identified 66 top-scoring novel HEM candidates from the Enamine database, with 32% containing a nitro group.
Conclusions:
- The Iterative Stochastic Elimination (ISE) algorithm demonstrates significant potential as a computational tool for discovering novel high-energy materials.
- This ML-driven approach offers an efficient and sustainable pathway for advancing HEM research.
- The study validates ISE's capability in molecular discovery beyond biomolecules, specifically for high-energy materials.
