Identifying Candidate Genetic Associations with MRI-derived AD-related ROI via Tree-guided Sparse Learning

Xiaoke Hao, Xiaohui Yao, Shannon Risacher, Andrew Saykin, Jintai Yu, Huifu Wang, Lan Tan, Li Shen, Daoqiang Zhang

Research output: Contribution to journalArticle

Abstract

Imaging genetics has attracted significant interests in recent studies. Traditional work has focused on univariate statistical approaches that identify important single nucleotide polymorphisms (SNPs) associated with quantitative traits (QTs). Recently, to address the problem of multiple comparison and weak detection, multivariate analysis methods such as Lasso are often used for selecting the most relevant SNPs associated with QTs. However, one problem of Lasso based feature selection methods for imaging genetics is that priori information, i.e., the hierarchical structure among SNPs are rarely used for designing powerful model. In this paper, we propose to identify the associations between candidate genetic features (i.e., SNPs) and MRI-derived measures using a tree-guided sparse learning (TGSL) method. The advantage of our method is that it explicitly models the complex hierarchical structure among the SNPs. Specifically, motivated by the biological phenomenon, the hierarchical structures involving gene groups, LD blocks and individual SNPs are imposed as a tree-guided regularization term in TGSL model. Experimental results on simulation and the ADNI database studies show that our method not only achieves better predictions than competing methods on the MRI-derived measures of AD-related ROIs, but also identifies sparse SNP patterns at the block level to better guide the biological interpretation.

Original languageEnglish (US)
JournalIEEE/ACM Transactions on Computational Biology and Bioinformatics
DOIs
StateAccepted/In press - May 5 2018

Fingerprint

Genetic Association
Single nucleotide Polymorphism
Nucleotides
Polymorphism
Magnetic resonance imaging
Single Nucleotide Polymorphism
Learning
Hierarchical Structure
Lasso
Imaging
Biological Phenomena
Imaging techniques
Multiple Comparisons
Multivariate Analysis
Complex Structure
Feature Selection
Univariate
Feature extraction
Regularization
Genes

Keywords

  • hierarchical structure
  • Imaging genetics
  • ROIs
  • SNPs
  • tree-guided sparse learning

ASJC Scopus subject areas

  • Biotechnology
  • Genetics
  • Applied Mathematics

Cite this

Identifying Candidate Genetic Associations with MRI-derived AD-related ROI via Tree-guided Sparse Learning. / Hao, Xiaoke; Yao, Xiaohui; Risacher, Shannon; Saykin, Andrew; Yu, Jintai; Wang, Huifu; Tan, Lan; Shen, Li; Zhang, Daoqiang.

In: IEEE/ACM Transactions on Computational Biology and Bioinformatics, 05.05.2018.

Research output: Contribution to journalArticle

@article{3b72029c53014873a811704a187226d5,
title = "Identifying Candidate Genetic Associations with MRI-derived AD-related ROI via Tree-guided Sparse Learning",
abstract = "Imaging genetics has attracted significant interests in recent studies. Traditional work has focused on univariate statistical approaches that identify important single nucleotide polymorphisms (SNPs) associated with quantitative traits (QTs). Recently, to address the problem of multiple comparison and weak detection, multivariate analysis methods such as Lasso are often used for selecting the most relevant SNPs associated with QTs. However, one problem of Lasso based feature selection methods for imaging genetics is that priori information, i.e., the hierarchical structure among SNPs are rarely used for designing powerful model. In this paper, we propose to identify the associations between candidate genetic features (i.e., SNPs) and MRI-derived measures using a tree-guided sparse learning (TGSL) method. The advantage of our method is that it explicitly models the complex hierarchical structure among the SNPs. Specifically, motivated by the biological phenomenon, the hierarchical structures involving gene groups, LD blocks and individual SNPs are imposed as a tree-guided regularization term in TGSL model. Experimental results on simulation and the ADNI database studies show that our method not only achieves better predictions than competing methods on the MRI-derived measures of AD-related ROIs, but also identifies sparse SNP patterns at the block level to better guide the biological interpretation.",
keywords = "hierarchical structure, Imaging genetics, ROIs, SNPs, tree-guided sparse learning",
author = "Xiaoke Hao and Xiaohui Yao and Shannon Risacher and Andrew Saykin and Jintai Yu and Huifu Wang and Lan Tan and Li Shen and Daoqiang Zhang",
year = "2018",
month = "5",
day = "5",
doi = "10.1109/TCBB.2018.2833487",
language = "English (US)",
journal = "IEEE/ACM Transactions on Computational Biology and Bioinformatics",
issn = "1545-5963",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Identifying Candidate Genetic Associations with MRI-derived AD-related ROI via Tree-guided Sparse Learning

AU - Hao, Xiaoke

AU - Yao, Xiaohui

AU - Risacher, Shannon

AU - Saykin, Andrew

AU - Yu, Jintai

AU - Wang, Huifu

AU - Tan, Lan

AU - Shen, Li

AU - Zhang, Daoqiang

PY - 2018/5/5

Y1 - 2018/5/5

N2 - Imaging genetics has attracted significant interests in recent studies. Traditional work has focused on univariate statistical approaches that identify important single nucleotide polymorphisms (SNPs) associated with quantitative traits (QTs). Recently, to address the problem of multiple comparison and weak detection, multivariate analysis methods such as Lasso are often used for selecting the most relevant SNPs associated with QTs. However, one problem of Lasso based feature selection methods for imaging genetics is that priori information, i.e., the hierarchical structure among SNPs are rarely used for designing powerful model. In this paper, we propose to identify the associations between candidate genetic features (i.e., SNPs) and MRI-derived measures using a tree-guided sparse learning (TGSL) method. The advantage of our method is that it explicitly models the complex hierarchical structure among the SNPs. Specifically, motivated by the biological phenomenon, the hierarchical structures involving gene groups, LD blocks and individual SNPs are imposed as a tree-guided regularization term in TGSL model. Experimental results on simulation and the ADNI database studies show that our method not only achieves better predictions than competing methods on the MRI-derived measures of AD-related ROIs, but also identifies sparse SNP patterns at the block level to better guide the biological interpretation.

AB - Imaging genetics has attracted significant interests in recent studies. Traditional work has focused on univariate statistical approaches that identify important single nucleotide polymorphisms (SNPs) associated with quantitative traits (QTs). Recently, to address the problem of multiple comparison and weak detection, multivariate analysis methods such as Lasso are often used for selecting the most relevant SNPs associated with QTs. However, one problem of Lasso based feature selection methods for imaging genetics is that priori information, i.e., the hierarchical structure among SNPs are rarely used for designing powerful model. In this paper, we propose to identify the associations between candidate genetic features (i.e., SNPs) and MRI-derived measures using a tree-guided sparse learning (TGSL) method. The advantage of our method is that it explicitly models the complex hierarchical structure among the SNPs. Specifically, motivated by the biological phenomenon, the hierarchical structures involving gene groups, LD blocks and individual SNPs are imposed as a tree-guided regularization term in TGSL model. Experimental results on simulation and the ADNI database studies show that our method not only achieves better predictions than competing methods on the MRI-derived measures of AD-related ROIs, but also identifies sparse SNP patterns at the block level to better guide the biological interpretation.

KW - hierarchical structure

KW - Imaging genetics

KW - ROIs

KW - SNPs

KW - tree-guided sparse learning

UR - http://www.scopus.com/inward/record.url?scp=85046429283&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85046429283&partnerID=8YFLogxK

U2 - 10.1109/TCBB.2018.2833487

DO - 10.1109/TCBB.2018.2833487

M3 - Article

AN - SCOPUS:85046429283

JO - IEEE/ACM Transactions on Computational Biology and Bioinformatics

JF - IEEE/ACM Transactions on Computational Biology and Bioinformatics

SN - 1545-5963

ER -