Analysis of Sequence Data Under Multivariate Trait-Dependent Sampling

Abstract

High-throughput DNA sequencing allows for the genotyping of common and rare variants for genetic association studies. At the present time and for the foreseeable future, it is not economically feasible to sequence all individuals in a large cohort. A cost-effective strategy is to sequence those individuals with extreme values of a quantitative trait. We consider the design under which the sampling depends on multiple quantitative traits. Under such trait-dependent sampling, standard linear regression analysis can result in bias of parameter estimation, inflation of Type I error, and loss of power. We construct a likelihood function that properly reflects the sampling mechanism and uses all available data. We implement a computationally efficient EM algorithm and establish the theoretical properties of the resulting maximum likelihood estimators. Our methods can be used to perform separate inference on each trait or simultaneous inference on multiple traits. We pay special attention to gene-level association tests for rare variants. We demonstrate the superiority of the proposed methods over standard linear regression through extensive simulation studies. We provide applications to the Cohorts for Heart and Aging Research in Genomic Epidemiology Targeted Sequencing Study and the National Heart, Lung, and Blood Institute Exome Sequencing Project. Supplementary materials for this article are available online.

• Association studies
• Gene-level tests
• Linear regression
• Quantitative traits
• Rare variants
• Sequencing studies

Additional information

Notes on contributors

Ran Tao

Ran Tao is Doctoral Student (E-mail: [email protected])

Donglin Zeng

Donglin Zeng is Professor (E-mail: [email protected])

Nora Franceschini

Dan-Yu Lin is Dennis Gillings Distinguished Professor (E-mail: [email protected]), Department of Biostatistics, University of North Carolina, Chapel Hill, NC 27599.

Kari E. North

Nora Franceschini is Research Assistant Professor (E-mail: [email protected])

Eric Boerwinkle

Kari E. North is Associate Professor (E-mail: [email protected]), Department of Epidemiology, University of North Carolina, Chapel Hill, NC 27599.

Dan-Yu Lin

Eric Boerwinkle is Professor and Director, Human Genetics Center, University of Texas Health Science Center, Houston, TX 77030 (E-mail: [email protected]).