MRlogP: Transfer Learning Enables Accurate logP Prediction Using Small Experimental Training Datasets

Small molecule lipophilicity is often included in generalized rules for medicinal chemistry. These rules aim to reduce time, effort, costs, and attrition rates in drug discovery, allowing the rejection or prioritization of compounds without the need for synthesis and testing. The availability of hig...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Yan-Kai Chen, Steven Shave, Manfred Auer
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	lipophilicity prediction logP prediction transfer learning physicochemical property prediction Chemical technology TP1-1185 Chemistry QD1-999
Acceso en línea:	https://doaj.org/article/9e399a6a966d4276ac9fafce1d8b1586
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:9e399a6a966d4276ac9fafce1d8b1586
record_format	dspace
spelling	oai:doaj.org-article:9e399a6a966d4276ac9fafce1d8b15862021-11-25T18:51:35ZMRlogP: Transfer Learning Enables Accurate logP Prediction Using Small Experimental Training Datasets10.3390/pr91120292227-9717https://doaj.org/article/9e399a6a966d4276ac9fafce1d8b15862021-11-01T00:00:00Zhttps://www.mdpi.com/2227-9717/9/11/2029https://doaj.org/toc/2227-9717Small molecule lipophilicity is often included in generalized rules for medicinal chemistry. These rules aim to reduce time, effort, costs, and attrition rates in drug discovery, allowing the rejection or prioritization of compounds without the need for synthesis and testing. The availability of high quality, abundant training data for machine learning methods can be a major limiting factor in building effective property predictors. We utilize transfer learning techniques to get around this problem, first learning on a large amount of low accuracy predicted logP values before finally tuning our model using a small, accurate dataset of 244 druglike compounds to create MRlogP, a neural network-based predictor of logP capable of outperforming state of the art freely available logP prediction methods for druglike small molecules. MRlogP achieves an average root mean squared error of 0.988 and 0.715 against druglike molecules from Reaxys and PHYSPROP. We have made the trained neural network predictor and all associated code for descriptor generation freely available. In addition, MRlogP may be used online via a web interface.Yan-Kai ChenSteven ShaveManfred AuerMDPI AGarticlelipophilicity predictionlogP predictiontransfer learningphysicochemical property predictionChemical technologyTP1-1185ChemistryQD1-999ENProcesses, Vol 9, Iss 2029, p 2029 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	lipophilicity prediction logP prediction transfer learning physicochemical property prediction Chemical technology TP1-1185 Chemistry QD1-999
spellingShingle	lipophilicity prediction logP prediction transfer learning physicochemical property prediction Chemical technology TP1-1185 Chemistry QD1-999 Yan-Kai Chen Steven Shave Manfred Auer MRlogP: Transfer Learning Enables Accurate logP Prediction Using Small Experimental Training Datasets
description	Small molecule lipophilicity is often included in generalized rules for medicinal chemistry. These rules aim to reduce time, effort, costs, and attrition rates in drug discovery, allowing the rejection or prioritization of compounds without the need for synthesis and testing. The availability of high quality, abundant training data for machine learning methods can be a major limiting factor in building effective property predictors. We utilize transfer learning techniques to get around this problem, first learning on a large amount of low accuracy predicted logP values before finally tuning our model using a small, accurate dataset of 244 druglike compounds to create MRlogP, a neural network-based predictor of logP capable of outperforming state of the art freely available logP prediction methods for druglike small molecules. MRlogP achieves an average root mean squared error of 0.988 and 0.715 against druglike molecules from Reaxys and PHYSPROP. We have made the trained neural network predictor and all associated code for descriptor generation freely available. In addition, MRlogP may be used online via a web interface.
format	article
author	Yan-Kai Chen Steven Shave Manfred Auer
author_facet	Yan-Kai Chen Steven Shave Manfred Auer
author_sort	Yan-Kai Chen
title	MRlogP: Transfer Learning Enables Accurate logP Prediction Using Small Experimental Training Datasets
title_short	MRlogP: Transfer Learning Enables Accurate logP Prediction Using Small Experimental Training Datasets
title_full	MRlogP: Transfer Learning Enables Accurate logP Prediction Using Small Experimental Training Datasets
title_fullStr	MRlogP: Transfer Learning Enables Accurate logP Prediction Using Small Experimental Training Datasets
title_full_unstemmed	MRlogP: Transfer Learning Enables Accurate logP Prediction Using Small Experimental Training Datasets
title_sort	mrlogp: transfer learning enables accurate logp prediction using small experimental training datasets
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/9e399a6a966d4276ac9fafce1d8b1586
work_keys_str_mv	AT yankaichen mrlogptransferlearningenablesaccuratelogppredictionusingsmallexperimentaltrainingdatasets AT stevenshave mrlogptransferlearningenablesaccuratelogppredictionusingsmallexperimentaltrainingdatasets AT manfredauer mrlogptransferlearningenablesaccuratelogppredictionusingsmallexperimentaltrainingdatasets
_version_	1718410664091844608

MRlogP: Transfer Learning Enables Accurate logP Prediction Using Small Experimental Training Datasets

Ejemplares similares