Integrating gene set analysis and nonlinear predictive modeling of disease phenotypes using a Bayesian multitask formulation

Research output: Contribution to journalReview article

1 Citation (Scopus)

Abstract

Background: Identifying molecular signatures of disease phenotypes is studied using two mainstream approaches: (i) Predictive modeling methods such as linear classification and regression algorithms are used to find signatures predictive of phenotypes from genomic data, which may not be robust due to limited sample size or highly correlated nature of genomic data. (ii) Gene set analysis methods are used to find gene sets on which phenotypes are linearly dependent by bringing prior biological knowledge into the analysis, which may not capture more complex nonlinear dependencies. Thus, formulating an integrated model of gene set analysis and nonlinear predictive modeling is of great practical importance. Results: In this study, we propose a Bayesian binary classification framework to integrate gene set analysis and nonlinear predictive modeling. We then generalize this formulation to multitask learning setting to model multiple related datasets conjointly. Our main novelty is the probabilistic nonlinear formulation that enables us to robustly capture nonlinear dependencies between genomic data and phenotype even with small sample sizes. We demonstrate the performance of our algorithms using repeated random subsampling validation experiments on two cancer and two tuberculosis datasets by predicting important disease phenotypes from genome-wide gene expression data. Conclusions: We are able to obtain comparable or even better predictive performance than a baseline Bayesian nonlinear algorithm and to identify sparse sets of relevant genes and gene sets on all datasets. We also show that our multitask learning formulation enables us to further improve the generalization performance and to better understand biological processes behind disease phenotypes.

Original languageEnglish (US)
Pages (from-to)123-135
Number of pages13
JournalBMC bioinformatics
Volume17
DOIs
StatePublished - Dec 13 2016
Externally publishedYes

Fingerprint

Predictive Modeling
Nonlinear Modeling
Phenotype
Genes
Gene
Formulation
Multi-task Learning
Genomics
Sample Size
Signature
Learning
Bayesian Classification
Biological Phenomena
Subsampling
Tuberculosis
Binary Classification
Small Sample Size
Integrated Model
Multiple Models
Gene Expression Data

Keywords

  • Cancer
  • Disease phenotypes
  • Gene set analysis
  • Multiple kernel learning
  • Nonlinear predictive modeling
  • Tuberculosis

ASJC Scopus subject areas

  • Structural Biology
  • Biochemistry
  • Molecular Biology
  • Computer Science Applications
  • Applied Mathematics

Cite this

@article{e77a27e690ea4af58073dc10e8793203,
title = "Integrating gene set analysis and nonlinear predictive modeling of disease phenotypes using a Bayesian multitask formulation",
abstract = "Background: Identifying molecular signatures of disease phenotypes is studied using two mainstream approaches: (i) Predictive modeling methods such as linear classification and regression algorithms are used to find signatures predictive of phenotypes from genomic data, which may not be robust due to limited sample size or highly correlated nature of genomic data. (ii) Gene set analysis methods are used to find gene sets on which phenotypes are linearly dependent by bringing prior biological knowledge into the analysis, which may not capture more complex nonlinear dependencies. Thus, formulating an integrated model of gene set analysis and nonlinear predictive modeling is of great practical importance. Results: In this study, we propose a Bayesian binary classification framework to integrate gene set analysis and nonlinear predictive modeling. We then generalize this formulation to multitask learning setting to model multiple related datasets conjointly. Our main novelty is the probabilistic nonlinear formulation that enables us to robustly capture nonlinear dependencies between genomic data and phenotype even with small sample sizes. We demonstrate the performance of our algorithms using repeated random subsampling validation experiments on two cancer and two tuberculosis datasets by predicting important disease phenotypes from genome-wide gene expression data. Conclusions: We are able to obtain comparable or even better predictive performance than a baseline Bayesian nonlinear algorithm and to identify sparse sets of relevant genes and gene sets on all datasets. We also show that our multitask learning formulation enables us to further improve the generalization performance and to better understand biological processes behind disease phenotypes.",
keywords = "Cancer, Disease phenotypes, Gene set analysis, Multiple kernel learning, Nonlinear predictive modeling, Tuberculosis",
author = "Mehmet Gonen",
year = "2016",
month = "12",
day = "13",
doi = "10.1186/s12859-016-1311-3",
language = "English (US)",
volume = "17",
pages = "123--135",
journal = "BMC Bioinformatics",
issn = "1471-2105",
publisher = "BioMed Central",

}

TY - JOUR

T1 - Integrating gene set analysis and nonlinear predictive modeling of disease phenotypes using a Bayesian multitask formulation

AU - Gonen, Mehmet

PY - 2016/12/13

Y1 - 2016/12/13

N2 - Background: Identifying molecular signatures of disease phenotypes is studied using two mainstream approaches: (i) Predictive modeling methods such as linear classification and regression algorithms are used to find signatures predictive of phenotypes from genomic data, which may not be robust due to limited sample size or highly correlated nature of genomic data. (ii) Gene set analysis methods are used to find gene sets on which phenotypes are linearly dependent by bringing prior biological knowledge into the analysis, which may not capture more complex nonlinear dependencies. Thus, formulating an integrated model of gene set analysis and nonlinear predictive modeling is of great practical importance. Results: In this study, we propose a Bayesian binary classification framework to integrate gene set analysis and nonlinear predictive modeling. We then generalize this formulation to multitask learning setting to model multiple related datasets conjointly. Our main novelty is the probabilistic nonlinear formulation that enables us to robustly capture nonlinear dependencies between genomic data and phenotype even with small sample sizes. We demonstrate the performance of our algorithms using repeated random subsampling validation experiments on two cancer and two tuberculosis datasets by predicting important disease phenotypes from genome-wide gene expression data. Conclusions: We are able to obtain comparable or even better predictive performance than a baseline Bayesian nonlinear algorithm and to identify sparse sets of relevant genes and gene sets on all datasets. We also show that our multitask learning formulation enables us to further improve the generalization performance and to better understand biological processes behind disease phenotypes.

AB - Background: Identifying molecular signatures of disease phenotypes is studied using two mainstream approaches: (i) Predictive modeling methods such as linear classification and regression algorithms are used to find signatures predictive of phenotypes from genomic data, which may not be robust due to limited sample size or highly correlated nature of genomic data. (ii) Gene set analysis methods are used to find gene sets on which phenotypes are linearly dependent by bringing prior biological knowledge into the analysis, which may not capture more complex nonlinear dependencies. Thus, formulating an integrated model of gene set analysis and nonlinear predictive modeling is of great practical importance. Results: In this study, we propose a Bayesian binary classification framework to integrate gene set analysis and nonlinear predictive modeling. We then generalize this formulation to multitask learning setting to model multiple related datasets conjointly. Our main novelty is the probabilistic nonlinear formulation that enables us to robustly capture nonlinear dependencies between genomic data and phenotype even with small sample sizes. We demonstrate the performance of our algorithms using repeated random subsampling validation experiments on two cancer and two tuberculosis datasets by predicting important disease phenotypes from genome-wide gene expression data. Conclusions: We are able to obtain comparable or even better predictive performance than a baseline Bayesian nonlinear algorithm and to identify sparse sets of relevant genes and gene sets on all datasets. We also show that our multitask learning formulation enables us to further improve the generalization performance and to better understand biological processes behind disease phenotypes.

KW - Cancer

KW - Disease phenotypes

KW - Gene set analysis

KW - Multiple kernel learning

KW - Nonlinear predictive modeling

KW - Tuberculosis

UR - http://www.scopus.com/inward/record.url?scp=85003430034&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85003430034&partnerID=8YFLogxK

U2 - 10.1186/s12859-016-1311-3

DO - 10.1186/s12859-016-1311-3

M3 - Review article

C2 - 28105911

AN - SCOPUS:85003430034

VL - 17

SP - 123

EP - 135

JO - BMC Bioinformatics

JF - BMC Bioinformatics

SN - 1471-2105

ER -