An Ensemble Structure and Physicochemical (SPOC) Descriptor for Machine‐Learning Prediction of Chemical Reaction and Molecular Properties. Issue 14 (19th May 2022)

• Record Type: Journal Article
• Authors: Yang, Qi; Liu, Yidi; Cheng, Junjie; Li, Yao; Liu, Siyuan; Duan, Yingdong; Zhang, Long; Luo, Sanzhong
• Is Part Of: Chemphyschem. Volume 23:Issue 14(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-05-19
• Subjects: machine learning -- molecular representation -- property prediction -- reaction prediction -- bioactivity prediction; Chemistry, Physical and theoretical -- Periodicals; 541.05
• DOI: 10.1002/cphc.202200255
• Languages: English
• ISSNs: 1439-4235

Abstract

Abstract: Feature representations, or descriptors, are machines' chemical language that largely shapes the prediction capability, generalizability and interpretability of machine learning models. To develop a generally applicable descriptor is highly warranted for chemists to deal with conventional prediction tasks in the context of sparsely distributed and small datasets. Inspired by the chemist's vision on molecules, we presented herein an ensemble descriptor, SPOC, curated on the principles of physical organic chemistry that integrates S tructure and P hysicoc hemical property (SPOC) of a molecule. SPOC could be readily constructed by combining molecular fingerprints, representing the structure of a given molecule, and molecular physicochemical properties extracted from RDKit or Mordred molecular descriptors. The applicability of SPOC was fully surveyed in a range of well‐structured chemical databases with machine learning tasks varying from regression to classifications. Abstract : Structure and Physicochemical property (SPOC) is an ensemble descriptor that integrates the structure and physicochemical descriptors of a given molecule. It is found to be efficient for machine learning predictions in a variety of chemical or biological prediction tasks.