Cancer classification and pathway discovery using non-negative matrix factorization

Zexian Zeng, Andy H. Vo, Chengsheng Mao, Susan E Clare*, Seema Ahsan Khan, Yuan Luo

*Corresponding author for this work

Research output: Contribution to journalArticle

Abstract

Objectives: Extracting genetic information from a full range of sequencing data is important for understanding disease. We propose a novel method to effectively explore the landscape of genetic mutations and aggregate them to predict cancer type. Design: We applied non-smooth non-negative matrix factorization (nsNMF) and support vector machine (SVM) to utilize the full range of sequencing data, aiming to better aggregate genetic mutations and improve their power to predict disease type. More specifically, we introduce a novel classifier to distinguish cancer types using somatic mutations obtained from whole-exome sequencing data. Mutations were identified from multiple cancers and scored using SIFT, PP2, and CADD, and collapsed at the individual gene level. nsNMF was then applied to reduce dimensionality and obtain coefficient and basis matrices. A feature matrix was derived from the obtained matrices to train a classifier for cancer type classification with the SVM model. Results: We have demonstrated that the classifier was able to distinguish four cancer types with reasonable accuracy. In five-fold cross-validations using mutation counts as features, the average prediction accuracy was 80% (SEM = 0.1%), significantly outperforming baselines and outperforming models using mutation scores as features. Conclusion: Using the factor matrices derived from the nsNMF, we identified multiple genes and pathways that are significantly associated with each cancer type. This study presents a generic and complete pipeline to study the associations between somatic mutations and cancers. The proposed method can be adapted to other studies for disease status classification and pathway discovery.

Original languageEnglish (US)
Article number103247
JournalJournal of Biomedical Informatics
Volume96
DOIs
StatePublished - Aug 1 2019

Fingerprint

Factorization
Mutation
Neoplasms
Classifiers
Support vector machines
Genes
Exome
Pipelines
Scanning electron microscopy

Keywords

  • Cancer
  • Classification
  • Non-negative matrix factorization
  • Pathway
  • Somatic mutation
  • Whole-exome sequencing

ASJC Scopus subject areas

  • Computer Science Applications
  • Health Informatics

Cite this

@article{17305219ae944d68b24febfdb043e132,
title = "Cancer classification and pathway discovery using non-negative matrix factorization",
abstract = "Objectives: Extracting genetic information from a full range of sequencing data is important for understanding disease. We propose a novel method to effectively explore the landscape of genetic mutations and aggregate them to predict cancer type. Design: We applied non-smooth non-negative matrix factorization (nsNMF) and support vector machine (SVM) to utilize the full range of sequencing data, aiming to better aggregate genetic mutations and improve their power to predict disease type. More specifically, we introduce a novel classifier to distinguish cancer types using somatic mutations obtained from whole-exome sequencing data. Mutations were identified from multiple cancers and scored using SIFT, PP2, and CADD, and collapsed at the individual gene level. nsNMF was then applied to reduce dimensionality and obtain coefficient and basis matrices. A feature matrix was derived from the obtained matrices to train a classifier for cancer type classification with the SVM model. Results: We have demonstrated that the classifier was able to distinguish four cancer types with reasonable accuracy. In five-fold cross-validations using mutation counts as features, the average prediction accuracy was 80{\%} (SEM = 0.1{\%}), significantly outperforming baselines and outperforming models using mutation scores as features. Conclusion: Using the factor matrices derived from the nsNMF, we identified multiple genes and pathways that are significantly associated with each cancer type. This study presents a generic and complete pipeline to study the associations between somatic mutations and cancers. The proposed method can be adapted to other studies for disease status classification and pathway discovery.",
keywords = "Cancer, Classification, Non-negative matrix factorization, Pathway, Somatic mutation, Whole-exome sequencing",
author = "Zexian Zeng and Vo, {Andy H.} and Chengsheng Mao and Clare, {Susan E} and Khan, {Seema Ahsan} and Yuan Luo",
year = "2019",
month = "8",
day = "1",
doi = "10.1016/j.jbi.2019.103247",
language = "English (US)",
volume = "96",
journal = "Journal of Biomedical Informatics",
issn = "1532-0464",
publisher = "Academic Press Inc.",

}

Cancer classification and pathway discovery using non-negative matrix factorization. / Zeng, Zexian; Vo, Andy H.; Mao, Chengsheng; Clare, Susan E; Khan, Seema Ahsan; Luo, Yuan.

In: Journal of Biomedical Informatics, Vol. 96, 103247, 01.08.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Cancer classification and pathway discovery using non-negative matrix factorization

AU - Zeng, Zexian

AU - Vo, Andy H.

AU - Mao, Chengsheng

AU - Clare, Susan E

AU - Khan, Seema Ahsan

AU - Luo, Yuan

PY - 2019/8/1

Y1 - 2019/8/1

N2 - Objectives: Extracting genetic information from a full range of sequencing data is important for understanding disease. We propose a novel method to effectively explore the landscape of genetic mutations and aggregate them to predict cancer type. Design: We applied non-smooth non-negative matrix factorization (nsNMF) and support vector machine (SVM) to utilize the full range of sequencing data, aiming to better aggregate genetic mutations and improve their power to predict disease type. More specifically, we introduce a novel classifier to distinguish cancer types using somatic mutations obtained from whole-exome sequencing data. Mutations were identified from multiple cancers and scored using SIFT, PP2, and CADD, and collapsed at the individual gene level. nsNMF was then applied to reduce dimensionality and obtain coefficient and basis matrices. A feature matrix was derived from the obtained matrices to train a classifier for cancer type classification with the SVM model. Results: We have demonstrated that the classifier was able to distinguish four cancer types with reasonable accuracy. In five-fold cross-validations using mutation counts as features, the average prediction accuracy was 80% (SEM = 0.1%), significantly outperforming baselines and outperforming models using mutation scores as features. Conclusion: Using the factor matrices derived from the nsNMF, we identified multiple genes and pathways that are significantly associated with each cancer type. This study presents a generic and complete pipeline to study the associations between somatic mutations and cancers. The proposed method can be adapted to other studies for disease status classification and pathway discovery.

AB - Objectives: Extracting genetic information from a full range of sequencing data is important for understanding disease. We propose a novel method to effectively explore the landscape of genetic mutations and aggregate them to predict cancer type. Design: We applied non-smooth non-negative matrix factorization (nsNMF) and support vector machine (SVM) to utilize the full range of sequencing data, aiming to better aggregate genetic mutations and improve their power to predict disease type. More specifically, we introduce a novel classifier to distinguish cancer types using somatic mutations obtained from whole-exome sequencing data. Mutations were identified from multiple cancers and scored using SIFT, PP2, and CADD, and collapsed at the individual gene level. nsNMF was then applied to reduce dimensionality and obtain coefficient and basis matrices. A feature matrix was derived from the obtained matrices to train a classifier for cancer type classification with the SVM model. Results: We have demonstrated that the classifier was able to distinguish four cancer types with reasonable accuracy. In five-fold cross-validations using mutation counts as features, the average prediction accuracy was 80% (SEM = 0.1%), significantly outperforming baselines and outperforming models using mutation scores as features. Conclusion: Using the factor matrices derived from the nsNMF, we identified multiple genes and pathways that are significantly associated with each cancer type. This study presents a generic and complete pipeline to study the associations between somatic mutations and cancers. The proposed method can be adapted to other studies for disease status classification and pathway discovery.

KW - Cancer

KW - Classification

KW - Non-negative matrix factorization

KW - Pathway

KW - Somatic mutation

KW - Whole-exome sequencing

UR - http://www.scopus.com/inward/record.url?scp=85068422313&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068422313&partnerID=8YFLogxK

U2 - 10.1016/j.jbi.2019.103247

DO - 10.1016/j.jbi.2019.103247

M3 - Article

C2 - 31271844

AN - SCOPUS:85068422313

VL - 96

JO - Journal of Biomedical Informatics

JF - Journal of Biomedical Informatics

SN - 1532-0464

M1 - 103247

ER -