G6PD and machine learning algorithms as prognostic and diagnostic indicators of liver hepatocellular carcinoma

Abstract Background Liver Hepatocellular carcinoma (LIHC) exhibits a high incidence of liver cancer with escalating mortality rates over time. Despite this, the underlying pathogenic mechanism of LIHC remains poorly understood. Materials & methods To address this gap, we conducted a comprehensiv...

Full description

Bibliographic Details
Main Authors: Fei Li, Boshen Wang, Hao Li, Lu Kong, Baoli Zhu
Format: Article
Language:English
Published: BMC 2024-01-01
Series:BMC Cancer
Subjects:
Online Access:https://doi.org/10.1186/s12885-024-11887-6
_version_ 1797274294227566592
author Fei Li
Boshen Wang
Hao Li
Lu Kong
Baoli Zhu
author_facet Fei Li
Boshen Wang
Hao Li
Lu Kong
Baoli Zhu
author_sort Fei Li
collection DOAJ
description Abstract Background Liver Hepatocellular carcinoma (LIHC) exhibits a high incidence of liver cancer with escalating mortality rates over time. Despite this, the underlying pathogenic mechanism of LIHC remains poorly understood. Materials & methods To address this gap, we conducted a comprehensive investigation into the role of G6PD in LIHC using a combination of bioinformatics analysis with database data and rigorous cell experiments. LIHC samples were obtained from TCGA, ICGC and GEO databases, and the differences in G6PD expression in different tissues were investigated by differential expression analysis, followed by the establishment of Nomogram to determine the percentage of G6PD in causing LIHC by examining the relationship between G6PD and clinical features, and the subsequent validation of the effect of G6PD on the activity, migration, and invasive ability of hepatocellular carcinoma cells by using the low expression of LI-7 and SNU-449. Additionally, we employed machine learning to validate and compare the predictive capacity of four algorithms for LIHC patient prognosis. Results Our findings revealed significantly elevated G6PD expression levels in liver cancer tissues as compared to normal tissues. Meanwhile, Nomogram and Adaboost, Catboost, and Gbdt Regression analyses showed that G6PD accounted for 46%, 31%, and 49% of the multiple factors leading to LIHC. Furthermore, we observed that G6PD knockdown in hepatocellular carcinoma cells led to reduced proliferation, migration, and invasion abilities. Remarkably, the Decision Tree C5.0 decision tree algorithm demonstrated superior discriminatory performance among the machine learning methods assessed. Conclusion The potential diagnostic utility of G6PD and Decision Tree C5.0 for LIHC opens up a novel avenue for early detection and improved treatment strategies for hepatocellular carcinoma.
first_indexed 2024-03-07T14:56:20Z
format Article
id doaj.art-4c7cbab1930d4e929f5c5810267f9120
institution Directory Open Access Journal
issn 1471-2407
language English
last_indexed 2024-03-07T14:56:20Z
publishDate 2024-01-01
publisher BMC
record_format Article
series BMC Cancer
spelling doaj.art-4c7cbab1930d4e929f5c5810267f91202024-03-05T19:23:42ZengBMCBMC Cancer1471-24072024-01-0124111910.1186/s12885-024-11887-6G6PD and machine learning algorithms as prognostic and diagnostic indicators of liver hepatocellular carcinomaFei Li0Boshen Wang1Hao Li2Lu Kong3Baoli Zhu4Key Laboratory of Environmental Medicine Engineering of Ministry of Education, School of Public Health, Southeast UniversityKey Laboratory of Environmental Medicine Engineering of Ministry of Education, School of Public Health, Southeast UniversityKey Laboratory of Environmental Medicine Engineering of Ministry of Education, School of Public Health, Southeast UniversityKey Laboratory of Environmental Medicine Engineering of Ministry of Education, School of Public Health, Southeast UniversityInstitute of Occupational Disease Prevention, Jiangsu Provincial Center for Disease Prevention and ControlAbstract Background Liver Hepatocellular carcinoma (LIHC) exhibits a high incidence of liver cancer with escalating mortality rates over time. Despite this, the underlying pathogenic mechanism of LIHC remains poorly understood. Materials & methods To address this gap, we conducted a comprehensive investigation into the role of G6PD in LIHC using a combination of bioinformatics analysis with database data and rigorous cell experiments. LIHC samples were obtained from TCGA, ICGC and GEO databases, and the differences in G6PD expression in different tissues were investigated by differential expression analysis, followed by the establishment of Nomogram to determine the percentage of G6PD in causing LIHC by examining the relationship between G6PD and clinical features, and the subsequent validation of the effect of G6PD on the activity, migration, and invasive ability of hepatocellular carcinoma cells by using the low expression of LI-7 and SNU-449. Additionally, we employed machine learning to validate and compare the predictive capacity of four algorithms for LIHC patient prognosis. Results Our findings revealed significantly elevated G6PD expression levels in liver cancer tissues as compared to normal tissues. Meanwhile, Nomogram and Adaboost, Catboost, and Gbdt Regression analyses showed that G6PD accounted for 46%, 31%, and 49% of the multiple factors leading to LIHC. Furthermore, we observed that G6PD knockdown in hepatocellular carcinoma cells led to reduced proliferation, migration, and invasion abilities. Remarkably, the Decision Tree C5.0 decision tree algorithm demonstrated superior discriminatory performance among the machine learning methods assessed. Conclusion The potential diagnostic utility of G6PD and Decision Tree C5.0 for LIHC opens up a novel avenue for early detection and improved treatment strategies for hepatocellular carcinoma.https://doi.org/10.1186/s12885-024-11887-6G6PDLiver hepatocellular carcinomaPrognosticMachine learningImmunologyDrug Sensitivity
spellingShingle Fei Li
Boshen Wang
Hao Li
Lu Kong
Baoli Zhu
G6PD and machine learning algorithms as prognostic and diagnostic indicators of liver hepatocellular carcinoma
BMC Cancer
G6PD
Liver hepatocellular carcinoma
Prognostic
Machine learning
Immunology
Drug Sensitivity
title G6PD and machine learning algorithms as prognostic and diagnostic indicators of liver hepatocellular carcinoma
title_full G6PD and machine learning algorithms as prognostic and diagnostic indicators of liver hepatocellular carcinoma
title_fullStr G6PD and machine learning algorithms as prognostic and diagnostic indicators of liver hepatocellular carcinoma
title_full_unstemmed G6PD and machine learning algorithms as prognostic and diagnostic indicators of liver hepatocellular carcinoma
title_short G6PD and machine learning algorithms as prognostic and diagnostic indicators of liver hepatocellular carcinoma
title_sort g6pd and machine learning algorithms as prognostic and diagnostic indicators of liver hepatocellular carcinoma
topic G6PD
Liver hepatocellular carcinoma
Prognostic
Machine learning
Immunology
Drug Sensitivity
url https://doi.org/10.1186/s12885-024-11887-6
work_keys_str_mv AT feili g6pdandmachinelearningalgorithmsasprognosticanddiagnosticindicatorsofliverhepatocellularcarcinoma
AT boshenwang g6pdandmachinelearningalgorithmsasprognosticanddiagnosticindicatorsofliverhepatocellularcarcinoma
AT haoli g6pdandmachinelearningalgorithmsasprognosticanddiagnosticindicatorsofliverhepatocellularcarcinoma
AT lukong g6pdandmachinelearningalgorithmsasprognosticanddiagnosticindicatorsofliverhepatocellularcarcinoma
AT baolizhu g6pdandmachinelearningalgorithmsasprognosticanddiagnosticindicatorsofliverhepatocellularcarcinoma