Analysis and predictive modeling of asthma phenotypes

Allan R. Brasier, Hyunsu Ju

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Molecular classification using robust biochemical measurements provides a level of diagnostic precision that is unattainable using indirect phenotypic measurements. Multidimensional measurements of proteins, genes, or metabolites (analytes) can identify subtle differences in the pathophysiology of patients with asthma in a way that is not otherwise possible using physiological or clinical assessments. We overview a method for relating biochemical analyte measurements to generate predictive models of discrete (categorical) clinical outcomes, a process referred to as "supervised classification." We consider problems inherent in wide (small n and large p ) high-dimensional data, including the curse of dimensionality, collinearity and lack of information content. We suggest methods for reducing the data to the most informative features. We describe different approaches for phenotypic modeling, using logistic regression, classification and regression trees, random forest and nonparametric regression spline modeling. We provide guidance on post hoc model evaluation and methods to evaluate model performance using ROC curves and generalized additive models. The application of validated predictive models for outcome prediction will significantly impact the clinical management of asthma.

Original languageEnglish
Title of host publicationAdvances in Experimental Medicine and Biology
Pages273-288
Number of pages16
Volume795
DOIs
StatePublished - 2014

Publication series

NameAdvances in Experimental Medicine and Biology
Volume795
ISSN (Print)00652598

Fingerprint

Asthma
Phenotype
ROC Curve
Logistic Models
Metabolites
Splines
Logistics
Proteins
Forests

Keywords

  • False discovery rate
  • Feature reduction
  • Generalized additive models (GAMs)
  • Logistic regression
  • Multivariate adaptive regression splines (MARS)
  • Multivariate analysis
  • Random forest
  • Receiver operating characteristic (ROC) curve
  • Significance of microarrays (SAM)
  • Supervised learning

ASJC Scopus subject areas

  • Biochemistry, Genetics and Molecular Biology(all)
  • Medicine(all)

Cite this

Brasier, A. R., & Ju, H. (2014). Analysis and predictive modeling of asthma phenotypes. In Advances in Experimental Medicine and Biology (Vol. 795, pp. 273-288). (Advances in Experimental Medicine and Biology; Vol. 795). https://doi.org/10.1007/978-1-4614-8603-9-17

Analysis and predictive modeling of asthma phenotypes. / Brasier, Allan R.; Ju, Hyunsu.

Advances in Experimental Medicine and Biology. Vol. 795 2014. p. 273-288 (Advances in Experimental Medicine and Biology; Vol. 795).

Research output: Chapter in Book/Report/Conference proceedingChapter

Brasier, AR & Ju, H 2014, Analysis and predictive modeling of asthma phenotypes. in Advances in Experimental Medicine and Biology. vol. 795, Advances in Experimental Medicine and Biology, vol. 795, pp. 273-288. https://doi.org/10.1007/978-1-4614-8603-9-17
Brasier AR, Ju H. Analysis and predictive modeling of asthma phenotypes. In Advances in Experimental Medicine and Biology. Vol. 795. 2014. p. 273-288. (Advances in Experimental Medicine and Biology). https://doi.org/10.1007/978-1-4614-8603-9-17
Brasier, Allan R. ; Ju, Hyunsu. / Analysis and predictive modeling of asthma phenotypes. Advances in Experimental Medicine and Biology. Vol. 795 2014. pp. 273-288 (Advances in Experimental Medicine and Biology).
@inbook{f67c1b307d034502b3056b140da457ed,
title = "Analysis and predictive modeling of asthma phenotypes",
abstract = "Molecular classification using robust biochemical measurements provides a level of diagnostic precision that is unattainable using indirect phenotypic measurements. Multidimensional measurements of proteins, genes, or metabolites (analytes) can identify subtle differences in the pathophysiology of patients with asthma in a way that is not otherwise possible using physiological or clinical assessments. We overview a method for relating biochemical analyte measurements to generate predictive models of discrete (categorical) clinical outcomes, a process referred to as {"}supervised classification.{"} We consider problems inherent in wide (small n and large p ) high-dimensional data, including the curse of dimensionality, collinearity and lack of information content. We suggest methods for reducing the data to the most informative features. We describe different approaches for phenotypic modeling, using logistic regression, classification and regression trees, random forest and nonparametric regression spline modeling. We provide guidance on post hoc model evaluation and methods to evaluate model performance using ROC curves and generalized additive models. The application of validated predictive models for outcome prediction will significantly impact the clinical management of asthma.",
keywords = "False discovery rate, Feature reduction, Generalized additive models (GAMs), Logistic regression, Multivariate adaptive regression splines (MARS), Multivariate analysis, Random forest, Receiver operating characteristic (ROC) curve, Significance of microarrays (SAM), Supervised learning",
author = "Brasier, {Allan R.} and Hyunsu Ju",
year = "2014",
doi = "10.1007/978-1-4614-8603-9-17",
language = "English",
isbn = "9781461486022",
volume = "795",
series = "Advances in Experimental Medicine and Biology",
pages = "273--288",
booktitle = "Advances in Experimental Medicine and Biology",

}

TY - CHAP

T1 - Analysis and predictive modeling of asthma phenotypes

AU - Brasier, Allan R.

AU - Ju, Hyunsu

PY - 2014

Y1 - 2014

N2 - Molecular classification using robust biochemical measurements provides a level of diagnostic precision that is unattainable using indirect phenotypic measurements. Multidimensional measurements of proteins, genes, or metabolites (analytes) can identify subtle differences in the pathophysiology of patients with asthma in a way that is not otherwise possible using physiological or clinical assessments. We overview a method for relating biochemical analyte measurements to generate predictive models of discrete (categorical) clinical outcomes, a process referred to as "supervised classification." We consider problems inherent in wide (small n and large p ) high-dimensional data, including the curse of dimensionality, collinearity and lack of information content. We suggest methods for reducing the data to the most informative features. We describe different approaches for phenotypic modeling, using logistic regression, classification and regression trees, random forest and nonparametric regression spline modeling. We provide guidance on post hoc model evaluation and methods to evaluate model performance using ROC curves and generalized additive models. The application of validated predictive models for outcome prediction will significantly impact the clinical management of asthma.

AB - Molecular classification using robust biochemical measurements provides a level of diagnostic precision that is unattainable using indirect phenotypic measurements. Multidimensional measurements of proteins, genes, or metabolites (analytes) can identify subtle differences in the pathophysiology of patients with asthma in a way that is not otherwise possible using physiological or clinical assessments. We overview a method for relating biochemical analyte measurements to generate predictive models of discrete (categorical) clinical outcomes, a process referred to as "supervised classification." We consider problems inherent in wide (small n and large p ) high-dimensional data, including the curse of dimensionality, collinearity and lack of information content. We suggest methods for reducing the data to the most informative features. We describe different approaches for phenotypic modeling, using logistic regression, classification and regression trees, random forest and nonparametric regression spline modeling. We provide guidance on post hoc model evaluation and methods to evaluate model performance using ROC curves and generalized additive models. The application of validated predictive models for outcome prediction will significantly impact the clinical management of asthma.

KW - False discovery rate

KW - Feature reduction

KW - Generalized additive models (GAMs)

KW - Logistic regression

KW - Multivariate adaptive regression splines (MARS)

KW - Multivariate analysis

KW - Random forest

KW - Receiver operating characteristic (ROC) curve

KW - Significance of microarrays (SAM)

KW - Supervised learning

UR - http://www.scopus.com/inward/record.url?scp=84892978633&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84892978633&partnerID=8YFLogxK

U2 - 10.1007/978-1-4614-8603-9-17

DO - 10.1007/978-1-4614-8603-9-17

M3 - Chapter

SN - 9781461486022

VL - 795

T3 - Advances in Experimental Medicine and Biology

SP - 273

EP - 288

BT - Advances in Experimental Medicine and Biology

ER -