Bootstrapped Sparse Canonical Correlation Analysis

Mining Stable Imaging and Genetic Associations With Implicit Structure Learning. Mining Stable Imaging and Genetic Associations With Implicit Structure Learning.

Jingwen Yan, Lei Du, Sungeun Kim, Shannon L. Risacher, Heng Huang, Mark Inlow, Jason H. Moore, Andrew Saykin, Li Shen

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Sparse canonical correlation analysis (SCCA) based on lasso and structured lasso has been widely studied to explore the complex associations between brain imaging and genetics features. Although those based on lasso have a better control of overall sparsity, they capture only a small portion of signals because of competition within correlated features. Advanced structure-based models provide a partial solution, but final patterns mostly depend on the prior structures applied. In this work, we propose a new framework, bootstrapped sparse canonical correlation analysis (BoSCCA), to explore the stable associations between correlated imaging and genetic data sets and to implicitly reconstruct the hidden structures. We compare the performances of BoSCCA and traditional SCCA using both synthetic and real data. In synthetic data, BoSCCA outperforms traditional SCCA in both association identification and group structure extraction, especially when the signal proportion goes below 5%. In real data, BoSCCA better captures the group structure within regions of interest and linkage disequilibrium blocks among single-nucleotide polymorphisms and yielded more biologically meaningful results.

Original languageEnglish (US)
Title of host publicationImaging Genetics
PublisherElsevier Inc.
Pages101-117
Number of pages17
ISBN (Electronic)9780128139691
ISBN (Print)9780128139684
DOIs
StatePublished - Sep 26 2017

Fingerprint

Group Structure
Learning
Linkage Disequilibrium
Neuroimaging
Single Nucleotide Polymorphism
alachlor
Datasets

Keywords

  • Bootstrap sampling
  • Feature selection
  • Imaging genetics
  • Selection stability
  • Sparse canonical correlation analysis
  • Structure learning

ASJC Scopus subject areas

  • Medicine (miscellaneous)

Cite this

Bootstrapped Sparse Canonical Correlation Analysis : Mining Stable Imaging and Genetic Associations With Implicit Structure Learning. Mining Stable Imaging and Genetic Associations With Implicit Structure Learning. / Yan, Jingwen; Du, Lei; Kim, Sungeun; Risacher, Shannon L.; Huang, Heng; Inlow, Mark; Moore, Jason H.; Saykin, Andrew; Shen, Li.

Imaging Genetics. Elsevier Inc., 2017. p. 101-117.

Research output: Chapter in Book/Report/Conference proceedingChapter

@inbook{00e354059ef543d399876260bcb66499,
title = "Bootstrapped Sparse Canonical Correlation Analysis: Mining Stable Imaging and Genetic Associations With Implicit Structure Learning. Mining Stable Imaging and Genetic Associations With Implicit Structure Learning.",
abstract = "Sparse canonical correlation analysis (SCCA) based on lasso and structured lasso has been widely studied to explore the complex associations between brain imaging and genetics features. Although those based on lasso have a better control of overall sparsity, they capture only a small portion of signals because of competition within correlated features. Advanced structure-based models provide a partial solution, but final patterns mostly depend on the prior structures applied. In this work, we propose a new framework, bootstrapped sparse canonical correlation analysis (BoSCCA), to explore the stable associations between correlated imaging and genetic data sets and to implicitly reconstruct the hidden structures. We compare the performances of BoSCCA and traditional SCCA using both synthetic and real data. In synthetic data, BoSCCA outperforms traditional SCCA in both association identification and group structure extraction, especially when the signal proportion goes below 5{\%}. In real data, BoSCCA better captures the group structure within regions of interest and linkage disequilibrium blocks among single-nucleotide polymorphisms and yielded more biologically meaningful results.",
keywords = "Bootstrap sampling, Feature selection, Imaging genetics, Selection stability, Sparse canonical correlation analysis, Structure learning",
author = "Jingwen Yan and Lei Du and Sungeun Kim and Risacher, {Shannon L.} and Heng Huang and Mark Inlow and Moore, {Jason H.} and Andrew Saykin and Li Shen",
year = "2017",
month = "9",
day = "26",
doi = "10.1016/B978-0-12-813968-4.00006-7",
language = "English (US)",
isbn = "9780128139684",
pages = "101--117",
booktitle = "Imaging Genetics",
publisher = "Elsevier Inc.",

}

TY - CHAP

T1 - Bootstrapped Sparse Canonical Correlation Analysis

T2 - Mining Stable Imaging and Genetic Associations With Implicit Structure Learning. Mining Stable Imaging and Genetic Associations With Implicit Structure Learning.

AU - Yan, Jingwen

AU - Du, Lei

AU - Kim, Sungeun

AU - Risacher, Shannon L.

AU - Huang, Heng

AU - Inlow, Mark

AU - Moore, Jason H.

AU - Saykin, Andrew

AU - Shen, Li

PY - 2017/9/26

Y1 - 2017/9/26

N2 - Sparse canonical correlation analysis (SCCA) based on lasso and structured lasso has been widely studied to explore the complex associations between brain imaging and genetics features. Although those based on lasso have a better control of overall sparsity, they capture only a small portion of signals because of competition within correlated features. Advanced structure-based models provide a partial solution, but final patterns mostly depend on the prior structures applied. In this work, we propose a new framework, bootstrapped sparse canonical correlation analysis (BoSCCA), to explore the stable associations between correlated imaging and genetic data sets and to implicitly reconstruct the hidden structures. We compare the performances of BoSCCA and traditional SCCA using both synthetic and real data. In synthetic data, BoSCCA outperforms traditional SCCA in both association identification and group structure extraction, especially when the signal proportion goes below 5%. In real data, BoSCCA better captures the group structure within regions of interest and linkage disequilibrium blocks among single-nucleotide polymorphisms and yielded more biologically meaningful results.

AB - Sparse canonical correlation analysis (SCCA) based on lasso and structured lasso has been widely studied to explore the complex associations between brain imaging and genetics features. Although those based on lasso have a better control of overall sparsity, they capture only a small portion of signals because of competition within correlated features. Advanced structure-based models provide a partial solution, but final patterns mostly depend on the prior structures applied. In this work, we propose a new framework, bootstrapped sparse canonical correlation analysis (BoSCCA), to explore the stable associations between correlated imaging and genetic data sets and to implicitly reconstruct the hidden structures. We compare the performances of BoSCCA and traditional SCCA using both synthetic and real data. In synthetic data, BoSCCA outperforms traditional SCCA in both association identification and group structure extraction, especially when the signal proportion goes below 5%. In real data, BoSCCA better captures the group structure within regions of interest and linkage disequilibrium blocks among single-nucleotide polymorphisms and yielded more biologically meaningful results.

KW - Bootstrap sampling

KW - Feature selection

KW - Imaging genetics

KW - Selection stability

KW - Sparse canonical correlation analysis

KW - Structure learning

UR - http://www.scopus.com/inward/record.url?scp=85054248272&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054248272&partnerID=8YFLogxK

U2 - 10.1016/B978-0-12-813968-4.00006-7

DO - 10.1016/B978-0-12-813968-4.00006-7

M3 - Chapter

SN - 9780128139684

SP - 101

EP - 117

BT - Imaging Genetics

PB - Elsevier Inc.

ER -