Utilizing Molecular Descriptor Importance to Enhance Endpoint Predictions

Quantitative structure–activity relationship (QSAR) models are essential for predicting endpoints that are otherwise challenging to estimate using other in silico approaches. Developing interpretable models for endpoint prediction is valuable as interpretable models may provide valuable insights int...

Full description

Saved in:
Bibliographic Details
Main Authors: Benjamin Bajželj, Marjana Novič, Viktor Drgan
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Toxics
Subjects:
Online Access:https://www.mdpi.com/2305-6304/13/5/383
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849719847655571456
author Benjamin Bajželj
Marjana Novič
Viktor Drgan
author_facet Benjamin Bajželj
Marjana Novič
Viktor Drgan
author_sort Benjamin Bajželj
collection DOAJ
description Quantitative structure–activity relationship (QSAR) models are essential for predicting endpoints that are otherwise challenging to estimate using other in silico approaches. Developing interpretable models for endpoint prediction is valuable as interpretable models may provide valuable insights into the relationship between molecular structure and observed biological or toxicological properties of compounds. In this study, we introduce a novel modification of counter-propagation artificial neural networks that aims to identify key molecular features responsible for classifying molecules into specific endpoint classes. The novel approach presented in this work dynamically adjusts molecular descriptor importance during model training, allowing different molecular descriptor importance values for structurally different molecules, which increases its adaptability to diverse sets of compounds. We applied the method to enzyme inhibition and hepatotoxicity classification datasets. Our findings show that the proposed approach improves the classification of molecules, reduces the number of neurons excited by molecules from different endpoint classes, and increases the number of acceptable models. The proposed approach may be useful in compound toxicity prediction and drug design studies.
format Article
id doaj-art-a04cd8fad6be498cabd0f92dda0eecb3
institution DOAJ
issn 2305-6304
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Toxics
spelling doaj-art-a04cd8fad6be498cabd0f92dda0eecb32025-08-20T03:12:04ZengMDPI AGToxics2305-63042025-05-0113538310.3390/toxics13050383Utilizing Molecular Descriptor Importance to Enhance Endpoint PredictionsBenjamin Bajželj0Marjana Novič1Viktor Drgan2Laboratory for Cheminformatics, Theory Department, National Institute of Chemistry, Hajdrihova 19, 1001 Ljubljana, SloveniaLaboratory for Cheminformatics, Theory Department, National Institute of Chemistry, Hajdrihova 19, 1001 Ljubljana, SloveniaLaboratory for Cheminformatics, Theory Department, National Institute of Chemistry, Hajdrihova 19, 1001 Ljubljana, SloveniaQuantitative structure–activity relationship (QSAR) models are essential for predicting endpoints that are otherwise challenging to estimate using other in silico approaches. Developing interpretable models for endpoint prediction is valuable as interpretable models may provide valuable insights into the relationship between molecular structure and observed biological or toxicological properties of compounds. In this study, we introduce a novel modification of counter-propagation artificial neural networks that aims to identify key molecular features responsible for classifying molecules into specific endpoint classes. The novel approach presented in this work dynamically adjusts molecular descriptor importance during model training, allowing different molecular descriptor importance values for structurally different molecules, which increases its adaptability to diverse sets of compounds. We applied the method to enzyme inhibition and hepatotoxicity classification datasets. Our findings show that the proposed approach improves the classification of molecules, reduces the number of neurons excited by molecules from different endpoint classes, and increases the number of acceptable models. The proposed approach may be useful in compound toxicity prediction and drug design studies.https://www.mdpi.com/2305-6304/13/5/383quantitative structure–activity relationshipmolecular descriptorsmolecular descriptor importanceenzyme inhibitionhepatotoxicity
spellingShingle Benjamin Bajželj
Marjana Novič
Viktor Drgan
Utilizing Molecular Descriptor Importance to Enhance Endpoint Predictions
Toxics
quantitative structure–activity relationship
molecular descriptors
molecular descriptor importance
enzyme inhibition
hepatotoxicity
title Utilizing Molecular Descriptor Importance to Enhance Endpoint Predictions
title_full Utilizing Molecular Descriptor Importance to Enhance Endpoint Predictions
title_fullStr Utilizing Molecular Descriptor Importance to Enhance Endpoint Predictions
title_full_unstemmed Utilizing Molecular Descriptor Importance to Enhance Endpoint Predictions
title_short Utilizing Molecular Descriptor Importance to Enhance Endpoint Predictions
title_sort utilizing molecular descriptor importance to enhance endpoint predictions
topic quantitative structure–activity relationship
molecular descriptors
molecular descriptor importance
enzyme inhibition
hepatotoxicity
url https://www.mdpi.com/2305-6304/13/5/383
work_keys_str_mv AT benjaminbajzelj utilizingmoleculardescriptorimportancetoenhanceendpointpredictions
AT marjananovic utilizingmoleculardescriptorimportancetoenhanceendpointpredictions
AT viktordrgan utilizingmoleculardescriptorimportancetoenhanceendpointpredictions