Modeling gene-wise dependencies improves the identification of drug response biomarkers in cancer studies

Olga Nikolova, Russell Moser, Christopher Kemp, Mehmet Gonen, Adam Margolin

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Motivation: In recent years, vast advances in biomedical technologies and comprehensive sequencing have revealed the genomic landscape of common forms of human cancer in unprecedented detail. The broad heterogeneity of the disease calls for rapid development of personalized therapies. Translating the readily available genomic data into useful knowledge that can be applied in the clinic remains a challenge. Computational methods are needed to aid these efforts by robustly analyzing genome-scale data from distinct experimental platforms for prioritization of targets and treatments. Results: We propose a novel, biologically motivated, Bayesian multitask approach, which explicitly models gene-centric dependencies across multiple and distinct genomic platforms. We introduce a gene-wise prior and present a fully Bayesian formulation of a group factor analysis model. In supervised prediction applications, our multitask approach leverages similarities in response profiles of groups of drugs that are more likely to be related to true biological signal, which leads to more robust performance and improved generalization ability. We evaluate the performance of our method on molecularly characterized collections of cell lines profiled against two compound panels, namely the Cancer Cell Line Encyclopedia and the Cancer Therapeutics Response Portal. We demonstrate that accounting for the gene-centric dependencies enables leveraging information from multi-omic input data and improves prediction and feature selection performance. We further demonstrate the applicability of our method in an unsupervised dimensionality reduction application by inferring genes essential to tumorigenesis in the pancreatic ductal adenocarcinoma and lung adenocarcinoma patient cohorts from The Cancer Genome Atlas.

Original languageEnglish (US)
Pages (from-to)1362-1369
Number of pages8
JournalBioinformatics
Volume33
Issue number9
DOIs
StatePublished - May 1 2017

Fingerprint

Pharmacological Biomarkers
Biomarkers
Cancer
Drugs
Genes
Gene
Genomics
Modeling
Neoplasms
Genome
Encyclopedias
Distinct
Cell Line
Biomedical Technology
Bayes Theorem
Prioritization
Robust Performance
Atlases
Line
Prediction

ASJC Scopus subject areas

  • Statistics and Probability
  • Medicine(all)
  • Biochemistry
  • Molecular Biology
  • Computer Science Applications
  • Computational Theory and Mathematics
  • Computational Mathematics

Cite this

Modeling gene-wise dependencies improves the identification of drug response biomarkers in cancer studies. / Nikolova, Olga; Moser, Russell; Kemp, Christopher; Gonen, Mehmet; Margolin, Adam.

In: Bioinformatics, Vol. 33, No. 9, 01.05.2017, p. 1362-1369.

Research output: Contribution to journalArticle

@article{acf9eb62bfd44f5ea1674b0eb1403d37,
title = "Modeling gene-wise dependencies improves the identification of drug response biomarkers in cancer studies",
abstract = "Motivation: In recent years, vast advances in biomedical technologies and comprehensive sequencing have revealed the genomic landscape of common forms of human cancer in unprecedented detail. The broad heterogeneity of the disease calls for rapid development of personalized therapies. Translating the readily available genomic data into useful knowledge that can be applied in the clinic remains a challenge. Computational methods are needed to aid these efforts by robustly analyzing genome-scale data from distinct experimental platforms for prioritization of targets and treatments. Results: We propose a novel, biologically motivated, Bayesian multitask approach, which explicitly models gene-centric dependencies across multiple and distinct genomic platforms. We introduce a gene-wise prior and present a fully Bayesian formulation of a group factor analysis model. In supervised prediction applications, our multitask approach leverages similarities in response profiles of groups of drugs that are more likely to be related to true biological signal, which leads to more robust performance and improved generalization ability. We evaluate the performance of our method on molecularly characterized collections of cell lines profiled against two compound panels, namely the Cancer Cell Line Encyclopedia and the Cancer Therapeutics Response Portal. We demonstrate that accounting for the gene-centric dependencies enables leveraging information from multi-omic input data and improves prediction and feature selection performance. We further demonstrate the applicability of our method in an unsupervised dimensionality reduction application by inferring genes essential to tumorigenesis in the pancreatic ductal adenocarcinoma and lung adenocarcinoma patient cohorts from The Cancer Genome Atlas.",
author = "Olga Nikolova and Russell Moser and Christopher Kemp and Mehmet Gonen and Adam Margolin",
year = "2017",
month = "5",
day = "1",
doi = "10.1093/bioinformatics/btw836",
language = "English (US)",
volume = "33",
pages = "1362--1369",
journal = "Bioinformatics",
issn = "1367-4803",
publisher = "Oxford University Press",
number = "9",

}

TY - JOUR

T1 - Modeling gene-wise dependencies improves the identification of drug response biomarkers in cancer studies

AU - Nikolova, Olga

AU - Moser, Russell

AU - Kemp, Christopher

AU - Gonen, Mehmet

AU - Margolin, Adam

PY - 2017/5/1

Y1 - 2017/5/1

N2 - Motivation: In recent years, vast advances in biomedical technologies and comprehensive sequencing have revealed the genomic landscape of common forms of human cancer in unprecedented detail. The broad heterogeneity of the disease calls for rapid development of personalized therapies. Translating the readily available genomic data into useful knowledge that can be applied in the clinic remains a challenge. Computational methods are needed to aid these efforts by robustly analyzing genome-scale data from distinct experimental platforms for prioritization of targets and treatments. Results: We propose a novel, biologically motivated, Bayesian multitask approach, which explicitly models gene-centric dependencies across multiple and distinct genomic platforms. We introduce a gene-wise prior and present a fully Bayesian formulation of a group factor analysis model. In supervised prediction applications, our multitask approach leverages similarities in response profiles of groups of drugs that are more likely to be related to true biological signal, which leads to more robust performance and improved generalization ability. We evaluate the performance of our method on molecularly characterized collections of cell lines profiled against two compound panels, namely the Cancer Cell Line Encyclopedia and the Cancer Therapeutics Response Portal. We demonstrate that accounting for the gene-centric dependencies enables leveraging information from multi-omic input data and improves prediction and feature selection performance. We further demonstrate the applicability of our method in an unsupervised dimensionality reduction application by inferring genes essential to tumorigenesis in the pancreatic ductal adenocarcinoma and lung adenocarcinoma patient cohorts from The Cancer Genome Atlas.

AB - Motivation: In recent years, vast advances in biomedical technologies and comprehensive sequencing have revealed the genomic landscape of common forms of human cancer in unprecedented detail. The broad heterogeneity of the disease calls for rapid development of personalized therapies. Translating the readily available genomic data into useful knowledge that can be applied in the clinic remains a challenge. Computational methods are needed to aid these efforts by robustly analyzing genome-scale data from distinct experimental platforms for prioritization of targets and treatments. Results: We propose a novel, biologically motivated, Bayesian multitask approach, which explicitly models gene-centric dependencies across multiple and distinct genomic platforms. We introduce a gene-wise prior and present a fully Bayesian formulation of a group factor analysis model. In supervised prediction applications, our multitask approach leverages similarities in response profiles of groups of drugs that are more likely to be related to true biological signal, which leads to more robust performance and improved generalization ability. We evaluate the performance of our method on molecularly characterized collections of cell lines profiled against two compound panels, namely the Cancer Cell Line Encyclopedia and the Cancer Therapeutics Response Portal. We demonstrate that accounting for the gene-centric dependencies enables leveraging information from multi-omic input data and improves prediction and feature selection performance. We further demonstrate the applicability of our method in an unsupervised dimensionality reduction application by inferring genes essential to tumorigenesis in the pancreatic ductal adenocarcinoma and lung adenocarcinoma patient cohorts from The Cancer Genome Atlas.

UR - http://www.scopus.com/inward/record.url?scp=85019728022&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85019728022&partnerID=8YFLogxK

U2 - 10.1093/bioinformatics/btw836

DO - 10.1093/bioinformatics/btw836

M3 - Article

VL - 33

SP - 1362

EP - 1369

JO - Bioinformatics

JF - Bioinformatics

SN - 1367-4803

IS - 9

ER -