Consistency of QSAR models: Correct split of training and test sets, ranking of models and performance parameters

Recent implementations of QSAR modelling software provide the user with numerous models and a wealth of information. In this work, we provide some guidance on how one should interpret the results of QSAR modelling, compare and assess the resulting models, and select the best and most consistent ones...

Full description

Saved in:
Bibliographic Details
Published in:SAR and QSAR in environmental research 2015-09, Vol.26 (7-9), p.683-700
Main Authors: Rácz, A., Bajusz, D., Héberger, K.
Format: Article
Language:English
Subjects:
Amidohydrolases - antagonists & inhibitors
Amidohydrolases - chemistry
Animals
Benzene Derivatives - chemistry
Benzene Derivatives - toxicity
cross-validation
Cyprinidae
Decision Support Techniques
Humans
Maleimides - chemistry
Maleimides - toxicity
model selection
Models, Statistical
Molecular Docking Simulation
Monoacylglycerol Lipases - antagonists & inhibitors
Monoacylglycerol Lipases - chemistry
performance parameters
Quantitative Structure-Activity Relationship
ranking
Software
sum of ranking differences
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Recent implementations of QSAR modelling software provide the user with numerous models and a wealth of information. In this work, we provide some guidance on how one should interpret the results of QSAR modelling, compare and assess the resulting models, and select the best and most consistent ones. Two QSAR datasets are applied as case studies for the comparison of model performance parameters and model selection methods. We demonstrate the capabilities of sum of ranking differences (SRD) in model selection and ranking, and identify the best performance indicators and models. While the exchange of the original training and (external) test sets does not affect the ranking of performance parameters, it provides improved models in certain cases (despite the lower number of molecules in the training set). Performance parameters for external validation are substantially separated from the other merits in SRD analyses, highlighting their value in data fusion.
ISSN:1062-936X
1029-046X
DOI:10.1080/1062936X.2015.1084647