PMID- 34722995
OWN - NLM
STAT- PubMed-not-MEDLINE
LR  - 20240403
IS  - 2470-1343 (Electronic)
IS  - 2470-1343 (Linking)
VI  - 6
IP  - 42
DP  - 2021 Oct 26
TI  - Knowledge-Embedded Message-Passing Neural Networks: Improving Molecular Property 
      Prediction with Human Knowledge.
PG  - 27955-27967
LID - 10.1021/acsomega.1c03839 [doi]
AB  - The graph neural network (GNN) has become a promising method to predict molecular 
      properties with end-to-end supervision, as it can learn molecular features 
      directly from chemical graphs in a black-box manner. However, to achieve high 
      prediction accuracy, it is essential to supervise a huge amount of property data, 
      which is often accompanied by a high property experiment cost. Prior to the deep 
      learning method, descriptor-based quantitative structure-property relationships 
      (QSPR) studies have investigated physical and chemical knowledge to manually 
      design descriptors for effectively predicting properties. In this study, we 
      extend a message-passing neural network (MPNN) to include a novel MPNN 
      architecture called the knowledge-embedded MPNN (KEMPNN) that can be supervised 
      together with nonquantitative knowledge annotations by human experts on a 
      chemical graph that contains information on the important substructure of a 
      molecule and its effect on the target property (e.g., positive or negative 
      effect). We evaluated the performance of the KEMPNN in a small training data 
      setting using a physical chemistry dataset in MoleculeNet (ESOL, FreeSolv, 
      Lipophilicity) and a polymer property (glass-transition temperature) dataset with 
      virtual knowledge annotations. The results demonstrate that the KEMPNN with 
      knowledge supervision can improve the prediction accuracy obtained from the MPNN. 
      The results also demonstrate that the accuracy of the KEMPNN is better than or 
      comparable to those of descriptor-based methods even in the case of small 
      training data.
CI  - (c) 2021 The Author. Published by American Chemical Society.
FAU - Hasebe, Tatsuya
AU  - Hasebe T
AUID- ORCID: 0000-0003-3812-7800
AD  - Research & Development Group, Hitachi, Ltd., 832-2, Horiguchi, Hitachinaka, 
      Ibaraki 312-0034, Japan.
LA  - eng
PT  - Journal Article
DEP - 20211014
PL  - United States
TA  - ACS Omega
JT  - ACS omega
JID - 101691658
PMC - PMC8552328
COIS- The author declares no competing financial interest.
EDAT- 2021/11/02 06:00
MHDA- 2021/11/02 06:01
PMCR- 2021/10/14
CRDT- 2021/11/01 09:31
PHST- 2021/07/20 00:00 [received]
PHST- 2021/10/01 00:00 [accepted]
PHST- 2021/11/01 09:31 [entrez]
PHST- 2021/11/02 06:00 [pubmed]
PHST- 2021/11/02 06:01 [medline]
PHST- 2021/10/14 00:00 [pmc-release]
AID - 10.1021/acsomega.1c03839 [doi]
PST - epublish
SO  - ACS Omega. 2021 Oct 14;6(42):27955-27967. doi: 10.1021/acsomega.1c03839. 
      eCollection 2021 Oct 26.