Design of New Dispersants Using Machine Learning and Visual Analytics

Artificial intelligence (AI) is an emerging technology that is revolutionizing the discovery of new materials. One key application of AI is virtual screening of chemical libraries, which enables the accelerated discovery of materials with desired properties. In this study, we developed computational...

Full description

Bibliographic Details
Main Authors: María Jimena Martínez, Roi Naveiro, Axel J. Soto, Pablo Talavante, Shin-Ho Kim Lee, Ramón Gómez Arrayas, Mario Franco, Pablo Mauleón, Héctor Lozano Ordóñez, Guillermo Revilla López, Marco Bernabei, Nuria E. Campillo, Ignacio Ponzoni
Format: Article
Language:English
Published: MDPI AG 2023-03-01
Series:Polymers
Subjects:
Online Access:https://www.mdpi.com/2073-4360/15/5/1324
_version_ 1797614553988595712
author María Jimena Martínez
Roi Naveiro
Axel J. Soto
Pablo Talavante
Shin-Ho Kim Lee
Ramón Gómez Arrayas
Mario Franco
Pablo Mauleón
Héctor Lozano Ordóñez
Guillermo Revilla López
Marco Bernabei
Nuria E. Campillo
Ignacio Ponzoni
author_facet María Jimena Martínez
Roi Naveiro
Axel J. Soto
Pablo Talavante
Shin-Ho Kim Lee
Ramón Gómez Arrayas
Mario Franco
Pablo Mauleón
Héctor Lozano Ordóñez
Guillermo Revilla López
Marco Bernabei
Nuria E. Campillo
Ignacio Ponzoni
author_sort María Jimena Martínez
collection DOAJ
description Artificial intelligence (AI) is an emerging technology that is revolutionizing the discovery of new materials. One key application of AI is virtual screening of chemical libraries, which enables the accelerated discovery of materials with desired properties. In this study, we developed computational models to predict the dispersancy efficiency of oil and lubricant additives, a critical property in their design that can be estimated through a quantity named blotter spot. We propose a comprehensive approach that combines machine learning techniques with visual analytics strategies in an interactive tool that supports domain experts’ decision-making. We evaluated the proposed models quantitatively and illustrated their benefits through a case study. Specifically, we analyzed a series of virtual polyisobutylene succinimide (PIBSI) molecules derived from a known reference substrate. Our best-performing probabilistic model was Bayesian Additive Regression Trees (BART), which achieved a mean absolute error of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>5.50</mn><mo>±</mo><mn>0.34</mn></mrow></semantics></math></inline-formula> and a root mean square error of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>7.56</mn><mo>±</mo><mn>0.47</mn></mrow></semantics></math></inline-formula>, as estimated through 5-fold cross-validation. To facilitate future research, we have made the dataset, including the potential dispersants used for modeling, publicly available. Our approach can help accelerate the discovery of new oil and lubricant additives, and our interactive tool can aid domain experts in making informed decisions based on blotter spot and other key properties.
first_indexed 2024-03-11T07:13:04Z
format Article
id doaj.art-f437d7b0a41c4524bff817acb649ed59
institution Directory Open Access Journal
issn 2073-4360
language English
last_indexed 2024-03-11T07:13:04Z
publishDate 2023-03-01
publisher MDPI AG
record_format Article
series Polymers
spelling doaj.art-f437d7b0a41c4524bff817acb649ed592023-11-17T08:29:17ZengMDPI AGPolymers2073-43602023-03-01155132410.3390/polym15051324Design of New Dispersants Using Machine Learning and Visual AnalyticsMaría Jimena Martínez0Roi Naveiro1Axel J. Soto2Pablo Talavante3Shin-Ho Kim Lee4Ramón Gómez Arrayas5Mario Franco6Pablo Mauleón7Héctor Lozano Ordóñez8Guillermo Revilla López9Marco Bernabei10Nuria E. Campillo11Ignacio Ponzoni12ISISTAN (CONICET-UNCPBA) Campus Universitario—Paraje Arroyo Seco, Tandil 7000, ArgentinaInstitute of Mathematical Sciences (ICMAT-CSIC), Nicolás Cabrera, nº 13-15, Campus de Cantoblanco, UAM, 28049 Madrid, SpainInstitute for Computer Science and Engineering (UNS–CONICET), San Andrés 800, Campus Palihue, Bahía Blanca 8000, ArgentinaAItenea Biotech, Parque Científico de Madrid, Ciudad Universitaria de Cantoblanco, Calle Faraday, 7, 28049 Madrid, SpainAItenea Biotech, Parque Científico de Madrid, Ciudad Universitaria de Cantoblanco, Calle Faraday, 7, 28049 Madrid, SpainAItenea Biotech, Parque Científico de Madrid, Ciudad Universitaria de Cantoblanco, Calle Faraday, 7, 28049 Madrid, SpainDepartment of Organic Chemistry, Institute for Advanced Research in Chemical Sciences (IAdChem) UAM, 28049 Madrid, SpainDepartment of Organic Chemistry, Institute for Advanced Research in Chemical Sciences (IAdChem) UAM, 28049 Madrid, SpainRepsol Technology Lab DC Technology & Corporate Venturing, Agustín de Betancourt s/n, Móstoles, 28935 Madrid, SpainRepsol Technology Lab DC Technology & Corporate Venturing, Agustín de Betancourt s/n, Móstoles, 28935 Madrid, SpainRepsol Technology Lab DC Technology & Corporate Venturing, Agustín de Betancourt s/n, Móstoles, 28935 Madrid, SpainInstitute of Mathematical Sciences (ICMAT-CSIC), Nicolás Cabrera, nº 13-15, Campus de Cantoblanco, UAM, 28049 Madrid, SpainInstitute for Computer Science and Engineering (UNS–CONICET), San Andrés 800, Campus Palihue, Bahía Blanca 8000, ArgentinaArtificial intelligence (AI) is an emerging technology that is revolutionizing the discovery of new materials. One key application of AI is virtual screening of chemical libraries, which enables the accelerated discovery of materials with desired properties. In this study, we developed computational models to predict the dispersancy efficiency of oil and lubricant additives, a critical property in their design that can be estimated through a quantity named blotter spot. We propose a comprehensive approach that combines machine learning techniques with visual analytics strategies in an interactive tool that supports domain experts’ decision-making. We evaluated the proposed models quantitatively and illustrated their benefits through a case study. Specifically, we analyzed a series of virtual polyisobutylene succinimide (PIBSI) molecules derived from a known reference substrate. Our best-performing probabilistic model was Bayesian Additive Regression Trees (BART), which achieved a mean absolute error of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>5.50</mn><mo>±</mo><mn>0.34</mn></mrow></semantics></math></inline-formula> and a root mean square error of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>7.56</mn><mo>±</mo><mn>0.47</mn></mrow></semantics></math></inline-formula>, as estimated through 5-fold cross-validation. To facilitate future research, we have made the dataset, including the potential dispersants used for modeling, publicly available. Our approach can help accelerate the discovery of new oil and lubricant additives, and our interactive tool can aid domain experts in making informed decisions based on blotter spot and other key properties.https://www.mdpi.com/2073-4360/15/5/1324polyisobutyleneblotter spotartificial intelligenceBayesian regression
spellingShingle María Jimena Martínez
Roi Naveiro
Axel J. Soto
Pablo Talavante
Shin-Ho Kim Lee
Ramón Gómez Arrayas
Mario Franco
Pablo Mauleón
Héctor Lozano Ordóñez
Guillermo Revilla López
Marco Bernabei
Nuria E. Campillo
Ignacio Ponzoni
Design of New Dispersants Using Machine Learning and Visual Analytics
Polymers
polyisobutylene
blotter spot
artificial intelligence
Bayesian regression
title Design of New Dispersants Using Machine Learning and Visual Analytics
title_full Design of New Dispersants Using Machine Learning and Visual Analytics
title_fullStr Design of New Dispersants Using Machine Learning and Visual Analytics
title_full_unstemmed Design of New Dispersants Using Machine Learning and Visual Analytics
title_short Design of New Dispersants Using Machine Learning and Visual Analytics
title_sort design of new dispersants using machine learning and visual analytics
topic polyisobutylene
blotter spot
artificial intelligence
Bayesian regression
url https://www.mdpi.com/2073-4360/15/5/1324
work_keys_str_mv AT mariajimenamartinez designofnewdispersantsusingmachinelearningandvisualanalytics
AT roinaveiro designofnewdispersantsusingmachinelearningandvisualanalytics
AT axeljsoto designofnewdispersantsusingmachinelearningandvisualanalytics
AT pablotalavante designofnewdispersantsusingmachinelearningandvisualanalytics
AT shinhokimlee designofnewdispersantsusingmachinelearningandvisualanalytics
AT ramongomezarrayas designofnewdispersantsusingmachinelearningandvisualanalytics
AT mariofranco designofnewdispersantsusingmachinelearningandvisualanalytics
AT pablomauleon designofnewdispersantsusingmachinelearningandvisualanalytics
AT hectorlozanoordonez designofnewdispersantsusingmachinelearningandvisualanalytics
AT guillermorevillalopez designofnewdispersantsusingmachinelearningandvisualanalytics
AT marcobernabei designofnewdispersantsusingmachinelearningandvisualanalytics
AT nuriaecampillo designofnewdispersantsusingmachinelearningandvisualanalytics
AT ignacioponzoni designofnewdispersantsusingmachinelearningandvisualanalytics