Efficient cross-trait penalized regression increases prediction accuracy in large cohorts using secondary phenotypes

We introduce cross-trait penalized regression (CTPR), a powerful and practical approach for multi-trait polygenic risk prediction in large cohorts. Specifically, we propose a novel cross-trait penalty function with the Lasso and the minimax concave penalty (MCP) to incorporate the shared genetic eff...

Full description

Saved in:
Bibliographic Details
Published in:Nature communications 2019-02, Vol.10 (1), p.569-11, Article 569
Main Authors: Chung, Wonil, Chen, Jun, Turman, Constance, Lindstrom, Sara, Zhu, Zhaozhong, Loh, Po-Ru, Kraft, Peter, Liang, Liming
Format: Article
Language:English
Subjects:
45
45/43
631/114
631/208
639/705
Algorithms
Biobanks
Body mass
Genetic effects
Genome-Wide Association Study - methods
Genotype
Genotypes
Humanities and Social Sciences
Humans
Minimax technique
Models, Genetic
multidisciplinary
Penalty function
Performance prediction
Phenotype
Phenotypes
Polygenic inheritance
Quantitative Trait Loci - genetics
Regression analysis
Science
Science (multidisciplinary)
Single-nucleotide polymorphism
Statistical analysis
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We introduce cross-trait penalized regression (CTPR), a powerful and practical approach for multi-trait polygenic risk prediction in large cohorts. Specifically, we propose a novel cross-trait penalty function with the Lasso and the minimax concave penalty (MCP) to incorporate the shared genetic effects across multiple traits for large-sample GWAS data. Our approach extracts information from the secondary traits that is beneficial for predicting the primary trait based on individual-level genotypes and/or summary statistics. Our novel implementation of a parallel computing algorithm makes it feasible to apply our method to biobank-scale GWAS data. We illustrate our method using large-scale GWAS data (~1M SNPs) from the UK Biobank ( N  = 456,837). We show that our multi-trait method outperforms the recently proposed multi-trait analysis of GWAS (MTAG) for predictive performance. The prediction accuracy for height by the aid of BMI improves from R 2  = 35.8% (MTAG) to 42.5% (MCP + CTPR) or 42.8% (Lasso + CTPR) with UK Biobank data. Information of genetic architectures of complex traits can be leveraged for predicting phenotypes. Here, the authors develop CTPR (Cross-Trait Penalized Regression), a method for multi-trait polygenic risk prediction using individual-level genotypes and/or summary statistics from large cohorts.
ISSN:2041-1723
2041-1723
DOI:10.1038/s41467-019-08535-0