Chemical Feature-Based Machine Learning Model for Predicting Photophysical Properties of BODIPY Compounds: Density Functional Theory and Quantitative Structure–Property Relationship Modeling DOI Creative Commons
Gerardo M. Casañola‐Martín, Jing Wang, Jian‐Ge Zhou

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 14, 2024

Abstract Boron-dipyrromethene (BODIPY) compounds have unique photophysical properties and been applied in fluorescence imaging, sensing, optoelectronics, beyond. In order to design effective BODIPY compounds, it is crucial acquire a comprehensive understanding of the relationships between structures corresponding photoproperties. present study, DFT/TDDFT was optimize studied models obtain their absorpton spectrum. Based upon theoretical computaional results, machine learning-based Quantitative Structure-Property Relationship (ML/QSPR) model employed for predicting maximum absorption wavelength (λ) by combining hand-crafted molecular descriptors (MD) Explainable Machine Learning (EML) techniques. A dataset 131 with experimental properties, used generate diverse set capturing information about size, shape, connectivity other structural features these compounds. Then genetic algorithm (GA) wrapper Multi-Linear Regression (MLR) performed. Fifteen were identified be strongly correlated wavelength. The developed ML/QSPR exhibited good predictive performance, coefficients determination (R2) 0.945 training 0.734 test set, demonstrating robustness reliability. posterior analysis some selected provided insights into that influence compound meanwhile also emphasizes importance branching, specific functional groups. Our work shows plausible learning approaches screen novel enhanced performance spectra.

Language: Английский

Prediction of Drug-like Compounds Solubility in Supercritical Carbon Dioxide: A Comparative Study between Classical Density Functional Theory and Machine Learning Approaches DOI
Dmitriy M. Makarov, Nikolai N. Kalikin, Yury A. Budkov

et al.

Industrial & Engineering Chemistry Research, Journal Year: 2024, Volume and Issue: 63(3), P. 1589 - 1603

Published: Jan. 15, 2024

Supercritical carbon dioxide (scCO2) plays an essential role in various technological procedures, making the solubility of drugs scCO2 a crucial aspect drug formulation process. This study focuses on utilizing theoretical approaches to predict drug-like compounds order select optimum parameters for subsequent experimental procedures. Several machine learning models were developed and compared with previously established approach based classical density functional theory (cDFT). The CatBoost model, alvaDesc descriptors, demonstrated reasonably accurate predictions 187 (AARD = 1.8%). Meanwhile, incorporating CDK descriptors melting points as input parameters, exhibited satisfactory accuracy 14.3%) extrapolating new compounds. Comparing results between cDFT-based one revealed, average, higher faster prediction speed former. However, cDFT more physical behavior isotherms models. was particularly evident when ML struggled accurately extrapolate values beyond range supercritical state. Model CatBoost/CDK is freely accessible at http://chem-predictor.isc-ras.ru/individual/scco/.

Language: Английский

Citations

9
Bionic inspired multifunctional modular energetic materials: an exploration of new generation of application-oriented energetic materials DOI
Yujia Wen, Linyuan Wen, Bojun Tan

et al.

Journal of Materials Chemistry A, Journal Year: 2024, Volume and Issue: 12(16), P. 9427 - 9437

Published: Jan. 1, 2024

Aiming to balance the pertinence and universality of energetic materials, this study proposes a new concept bionic inspired multifunctional modular materials seeks out potential monomers via high-throughput screening strategy.

Language: Английский

Citations

4
Assisted Energetic Material Property Prediction through Advanced Transfer Learning with Graph Neural Networks DOI

Jianjian Hu,

Jun-Xuan Jin,

Xiao‐Jing Hou

et al.

Industrial & Engineering Chemistry Research, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 16, 2025

In this study, we explore the use of transfer learning to predict properties energetic materials using a force-field-inspired transformer graph neural network (FFiTrNet). We began by pretraining model on large data set CHNOF compounds and then fine-tuning it smaller experimental enthalpy formation for materials. Our results show that significantly enhances accuracy predicting formation, with reduction in mean absolute error root-mean-square compared direct training set. Furthermore, demonstrate effectiveness other materials, highlighting its potential improve predictive capabilities machine models range properties. The result is most accurate among state-of-the-art material used enriches database materials' properties, making valuable publicly available future research.

Language: Английский

Citations

0
Accelerated predictions of the sublimation enthalpy of organic materials with machine learning DOI Creative Commons
Yifan Liu, Tran Doan Huan,

Chaofan Huang

et al.

Materials Genome Engineering Advances, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 19, 2025

Abstract The sublimation enthalpy, , is a key thermodynamic parameter governing the phase transformation of substance between its solid and gas phases. This at core many important materials' purification, deposition, etching processes. While can be measured experimentally estimated computationally, these approaches have their own different challenges. Here, we develop machine learning (ML) approach to rapidly predict from data generated using density functional theory (DFT). We further demonstrate how combining ML DFT methods with active efficient in exploring materials space, expanding coverage computed dataset, systematically improving predictive model . With an error kJ/mol instantaneous predictions developed this work will useful for community.

Language: Английский

Citations

0
Predictive Modeling for Energetic Materials DOI
Didier Mathieu

Challenges and advances in computational chemistry and physics, Journal Year: 2025, Volume and Issue: unknown, P. 265 - 310

Published: Jan. 1, 2025

Language: Английский

Citations

0
First-principles calculations of solid-phase enthalpy of formation of energetic materials DOI Creative Commons
Lixiang Zhong, Danyang Liu, Mingwei Hu

et al.

Communications Chemistry, Journal Year: 2025, Volume and Issue: 8(1)

Published: May 10, 2025

Language: Английский

Citations

0
Screening heat-resistant energetic molecules via deep learning and high-throughput computation DOI
Jian Liu, Jie Tian, Rui Liu

et al.

Chemical Engineering Journal, Journal Year: 2025, Volume and Issue: unknown, P. 160218 - 160218

Published: Feb. 1, 2025

Language: Английский

Citations

0
Vaporization enthalpy prediction of ionic liquids based on back-propagation artificial neural network DOI Creative Commons

Changzheng Ji,

Zhaochong Shi,

Yufeng Zheng

et al.

Green Chemical Engineering, Journal Year: 2025, Volume and Issue: unknown

Published: March 1, 2025

Language: Английский

Citations

0
MDs-NP: a property prediction model construction procedure for naphtha based on molecular dynamics simulation DOI
Yixin Wei, Tong Qiu

Journal of Physics Condensed Matter, Journal Year: 2024, Volume and Issue: 36(31), P. 315402 - 315402

Published: April 24, 2024

Abstract In the context of carbon neutrality and peaking, molecular management has become a focus petrochemical industry. The key to achieving is reconstruction, which relies on rapid accurate calculation oil properties. Focusing naphtha, we proposed novel property prediction model construction procedure (MDs-NP) employing dynamics simulations for collections gamma distribution from real analytical data calculating mole fractions simulation mixtures. We calculated 348 sets mixture properties in range 273 K–300 K by simulations. Molecular feature extraction was based descriptors. addition descriptors open-source toolkits (RDKit Mordred), designed 12 naphtha knowledge (NK) with naphtha. Three machine learning algorithms (support vector regression, extreme gradient boosting artificial neural network) were applied compared establish models density viscosity Mordred NK + support regression algorithm achieved best performance density. selected RDKFp network viscosity. Using ablation studies, T, P_w CC(C)C are three effective that can improve models. MDs-NP potential be extended more as well more-complex petroleum systems. used reconstruction facilitate data-driven intelligent transformation processes.

Language: Английский

Citations

2
Micro-scale crystallization thermodynamics study of typical energetic compounds integrating optofluidics and machine learning DOI
Xingyi Zhou,

Li Liu,

Yipeng Fei

et al.

Chemical Engineering Science, Journal Year: 2024, Volume and Issue: 298, P. 120443 - 120443

Published: Oct. 1, 2024

Language: Английский

Citations

1