Machine learning algorithms identify target genes and the molecular mechanism of matrine against diffuse large B-cell lymphoma

Yidong Zhu; Zhongping Ning; Ximing Li; Zhikang Lin

doi:10.2174/1573409920666230821102806

Machine learning algorithms identify target genes and the molecular mechanism of matrine against diffuse large B-cell lymphoma

Curr Comput Aided Drug Des. 2023 Aug 21. doi: 10.2174/1573409920666230821102806. Online ahead of print.

Authors

Yidong Zhu¹, Zhongping Ning², Ximing Li², Zhikang Lin³

Affiliations

¹ Shanghai Tenth People's Hospital, Tongji University School of Medicine, Shanghai, 200072, China.
² Shanghai Pudong New District Zhoupu Hospital, Shanghai University of Medicine & Health Sciences, Shanghai, 201318, China.
³ Longhua Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai, 200032, China.

PMID: 37605410
DOI: 10.2174/1573409920666230821102806

Abstract

Background: Diffuse large B-cell lymphoma (DLBCL) is the most common type of non-Hodgkin's lymphoma worldwide. Novel treatment strategies are still needed for refractory or relapsed DLBCL.

Objective: The present study aimed to systematically explore the potential targets and molecular mechanisms of matrine in the treatment of DLBCL.

Methods: Potential matrine targets were collected from multiple platforms. Microarray data and clinical characteristics of DLBCL were downloaded from publicly available databases. Differential expression analysis and weighted gene co-expression network analysis (WGCNA) were applied to identify the hub genes of DLBCL using R software. Then, the shared target genes between matrine and DLBCL were identified as the potential targets of matrine against DLBCL. The least absolute shrinkage and selection operator (LASSO) algorithm was used to determine the final core target genes, which were further verified by molecular docking simulation and receiver operating characteristic (ROC) curve analysis. Functional analysis was also performed to elucidate the potential mechanisms.

Results: A total of 222 matrine target genes and 1269 DLBCL hub genes were obtained through multiple databases and machine learning algorithms, respectively. From the nine shared target genes of matrine and DLBCL, five final core target genes, including CTSL, NR1H2, PDPK1, MDM2, and JAK3, were identified. Molecular docking showed that the binding of matrine to the core genes was stable. ROC curves also suggested close associations between the core genes and DLBCL. Additionally, functional analysis showed that the therapeutic effect of matrine against DLBCL may be related to the PI3K-Akt signaling pathway.

Conclusion: Matrine may target five genes and the PI3K-Akt signaling pathway in DLBCL treatment.

Keywords: Machine learning algorithms; PI3K-Akt signaling pathwayMachine learning algorithms; diffuse large B-cell lymphoma; matrine; network pharmacology; network pharmacologyMachine learning algorithms; the least absolute shrinkage and selection operator; weighted gene co-expression network analysis.