Topologically inferring active miRNA‐mediated subpathways toward precise cancer classification by directed random walk

Accurate predictions of classification biomarkers and disease status are indispensable for clinical cancer diagnosis and research. However, the robustness of conventional gene biomarkers is limited by issues with reproducibility across different measurement platforms and cohorts of patients. In this...

Full description

Saved in:
Bibliographic Details
Main Authors: Ziyu Ning, Chenchen Feng, Chao Song, Wei Liu, Desi Shang, Meng Li, Qiuyu Wang, Jianmei Zhao, Yuejuan Liu, Jiaxin Chen, Xiaoyang Yu, Jian Zhang, Chunquan Li
Format: Article
Language:English
Published: Wiley 2019-10-01
Series:Molecular Oncology
Subjects:
Online Access:https://doi.org/10.1002/1878-0261.12563
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849314979713384448
author Ziyu Ning
Chenchen Feng
Chao Song
Wei Liu
Desi Shang
Meng Li
Qiuyu Wang
Jianmei Zhao
Yuejuan Liu
Jiaxin Chen
Xiaoyang Yu
Jian Zhang
Chunquan Li
author_facet Ziyu Ning
Chenchen Feng
Chao Song
Wei Liu
Desi Shang
Meng Li
Qiuyu Wang
Jianmei Zhao
Yuejuan Liu
Jiaxin Chen
Xiaoyang Yu
Jian Zhang
Chunquan Li
author_sort Ziyu Ning
collection DOAJ
description Accurate predictions of classification biomarkers and disease status are indispensable for clinical cancer diagnosis and research. However, the robustness of conventional gene biomarkers is limited by issues with reproducibility across different measurement platforms and cohorts of patients. In this study, we collected 4775 samples from 12 different cancer datasets, which contained 4636 TCGA samples and 139 GEO samples. A new method was developed to detect miRNA‐mediated subpathway activities by using directed random walk (miDRW). To calculate the activity of each miRNA‐mediated subpathway, we constructed a global directed pathway network (GDPN) with genes as nodes. We then identified miRNAs with expression levels which were strongly inversely correlated with differentially expressed target genes in the GDPN. Finally, each miRNA‐mediated subpathway activity was integrated with the topological information, differential levels of miRNAs and genes, expression levels of genes, and target relationships between miRNAs and genes. The results showed that the proposed method yielded a more robust and accurate overall performance compared with other existing pathway‐based, miRNA‐based, and gene‐based classification methods. The high‐frequency miRNA‐mediated subpathways are more reliable in classifying samples and for selecting therapeutic strategies.
format Article
id doaj-art-30ae4070851a4b32bb54d4a3a2f83a3f
institution Kabale University
issn 1574-7891
1878-0261
language English
publishDate 2019-10-01
publisher Wiley
record_format Article
series Molecular Oncology
spelling doaj-art-30ae4070851a4b32bb54d4a3a2f83a3f2025-08-20T03:52:16ZengWileyMolecular Oncology1574-78911878-02612019-10-0113102211222610.1002/1878-0261.12563Topologically inferring active miRNA‐mediated subpathways toward precise cancer classification by directed random walkZiyu Ning0Chenchen Feng1Chao Song2Wei Liu3Desi Shang4Meng Li5Qiuyu Wang6Jianmei Zhao7Yuejuan Liu8Jiaxin Chen9Xiaoyang Yu10Jian Zhang11Chunquan Li12School of Medical Informatics Harbin Medical University Daqing ChinaSchool of Medical Informatics Harbin Medical University Daqing ChinaSchool of Pharmacology Harbin Medical University Daqing ChinaDepartment of Mathematics Heilongjiang Institute of Technology Harbin ChinaCollege of Bioinformatics Science and Technology Harbin Medical University ChinaSchool of Medical Informatics Harbin Medical University Daqing ChinaSchool of Medical Informatics Harbin Medical University Daqing ChinaSchool of Medical Informatics Harbin Medical University Daqing ChinaSchool of Medical Informatics Harbin Medical University Daqing ChinaSchool of Medical Informatics Harbin Medical University Daqing ChinaThe Higher Educational Key Laboratory for Measuring & Control Technology and Instrumentations of Heilongjiang Province Harbin University of Science and Technology ChinaSchool of Medical Informatics Harbin Medical University Daqing ChinaSchool of Medical Informatics Harbin Medical University Daqing ChinaAccurate predictions of classification biomarkers and disease status are indispensable for clinical cancer diagnosis and research. However, the robustness of conventional gene biomarkers is limited by issues with reproducibility across different measurement platforms and cohorts of patients. In this study, we collected 4775 samples from 12 different cancer datasets, which contained 4636 TCGA samples and 139 GEO samples. A new method was developed to detect miRNA‐mediated subpathway activities by using directed random walk (miDRW). To calculate the activity of each miRNA‐mediated subpathway, we constructed a global directed pathway network (GDPN) with genes as nodes. We then identified miRNAs with expression levels which were strongly inversely correlated with differentially expressed target genes in the GDPN. Finally, each miRNA‐mediated subpathway activity was integrated with the topological information, differential levels of miRNAs and genes, expression levels of genes, and target relationships between miRNAs and genes. The results showed that the proposed method yielded a more robust and accurate overall performance compared with other existing pathway‐based, miRNA‐based, and gene‐based classification methods. The high‐frequency miRNA‐mediated subpathways are more reliable in classifying samples and for selecting therapeutic strategies.https://doi.org/10.1002/1878-0261.12563cancer biomarkerclassificationmiRNA‐mediated subpathwaytopological information
spellingShingle Ziyu Ning
Chenchen Feng
Chao Song
Wei Liu
Desi Shang
Meng Li
Qiuyu Wang
Jianmei Zhao
Yuejuan Liu
Jiaxin Chen
Xiaoyang Yu
Jian Zhang
Chunquan Li
Topologically inferring active miRNA‐mediated subpathways toward precise cancer classification by directed random walk
Molecular Oncology
cancer biomarker
classification
miRNA‐mediated subpathway
topological information
title Topologically inferring active miRNA‐mediated subpathways toward precise cancer classification by directed random walk
title_full Topologically inferring active miRNA‐mediated subpathways toward precise cancer classification by directed random walk
title_fullStr Topologically inferring active miRNA‐mediated subpathways toward precise cancer classification by directed random walk
title_full_unstemmed Topologically inferring active miRNA‐mediated subpathways toward precise cancer classification by directed random walk
title_short Topologically inferring active miRNA‐mediated subpathways toward precise cancer classification by directed random walk
title_sort topologically inferring active mirna mediated subpathways toward precise cancer classification by directed random walk
topic cancer biomarker
classification
miRNA‐mediated subpathway
topological information
url https://doi.org/10.1002/1878-0261.12563
work_keys_str_mv AT ziyuning topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT chenchenfeng topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT chaosong topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT weiliu topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT desishang topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT mengli topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT qiuyuwang topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT jianmeizhao topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT yuejuanliu topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT jiaxinchen topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT xiaoyangyu topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT jianzhang topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk
AT chunquanli topologicallyinferringactivemirnamediatedsubpathwaystowardprecisecancerclassificationbydirectedrandomwalk