MolFeSCue: enhancing molecular property prediction in data-limited and imbalanced contexts using few-shot and contrastive learning
Open Access
- 29 February 2024
- journal article
- research article
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 40 (4)
Abstract
Motivation: Predicting molecular properties is a pivotal task in various scientific domains, including drug discovery, material science, and computational chemistry. This problem is often hindered by the lack of annotated data and imbalanced class distributions, which pose significant challenges in developing accurate and robust predictive models. Results: This study tackles these issues by employing pretrained molecular models within a few-shot learning framework. A novel dynamic contrastive loss function is utilized to further improve model performance in the situation of class imbalance. The proposed MolFeSCue framework not only facilitates rapid generalization from minimal samples, but also employs a contrastive loss function to extract meaningful molecular representations from imbalanced datasets. Extensive evaluations and comparisons of MolFeSCue and state-of-the-art algorithms have been conducted on multiple benchmark datasets, and the experimental data demonstrate our algorithm’s effectiveness in molecular representations and its broad applicability across various pretrained models. Our findings underscore MolFeSCues potential to accelerate advancements in drug discovery. Availability and implementation: We have made all the source code utilized in this study publicly accessible via GitHub at http://www.healthinformaticslab.org/supp/ or https://github.com/zhangruochi/MolFeSCue. The code (MolFeSCue-v1-00) is also available as the supplementary file of this paper.Keywords
Funding Information
- Senior and Junior Technological Innovation Team (20210509055RQ)
- Guizhou Provincial Science and Technology Projects (ZK2023-297)
- Science and Technology Foundation of Health Commission of Guizhou Province (gzwkj2023-565)
- National Natural Science Foundation of China (62072212, U19A2061)
- Jilin Provincial Key Laboratory of Big Data Intelligent Computing (20180622002JC)
This publication has 35 references indexed in Scilit:
- Artificial intelligence in drug discovery and developmentDrug Discovery Today, 2021
- Applications of Deep Learning in Molecule Generation and Molecular Property PredictionAccounts of Chemical Research, 2020
- PySpark and RDKit: Moving towards Big Data in CheminformaticsMolecular Informatics, 2019
- Machine learning for molecular and materials scienceNature, 2018
- MoleculeNet: a benchmark for molecular machine learningChemical Science, 2017
- Low Data Drug Discovery with One-Shot LearningACS Central Science, 2017
- Deep learning for computational chemistryJournal of Computational Chemistry, 2017
- In Silico ADMET Prediction: Recent Advances, Current Challenges and Future TrendsCurrent Topics in Medicinal Chemistry, 2013
- Activity Cliffs: Facts or Artifacts?Chemical Biology & Drug Design, 2013
- In silicoprediction of ADMET properties: how far have we come?Expert Opinion on Drug Metabolism & Toxicology, 2007