TY - JOUR

T1 - Use of Genotypes of Common Variants for Genome-Wide Regional Association Analysis

AU - Kirichenko, A. V.

AU - Zorkoltseva, I. V.

AU - Belonogova, N. M.

AU - Axenovich, T. I.

PY - 2018/2/1

Y1 - 2018/2/1

N2 - Regional association analysis is a new statistical method which simultaneously considers all variants in a selected genome region. This method was created for the analysis of rare genetic variants, whose genotypes are determined by exome or genome sequencing. The gene is usually considered as a region. It was also proposed to use a regional analysis for testing of the association between a complex trait and a set of common variants genotyped by the panels developed for genome-wide association analysis. In this case, overlapping genome regions (sliding windows) are usually considered as a region. Since the size of such regions can be rather large, there is a risk of overestimation (inflation) of the test statistic and an increase in the type I error. In this work, the effect of the size of the region on the type I error was studied for traits with different heritability. The results of simulating experiments demonstrated that the physical size of the region but not the number of genetic variants in it is a limiting factor. The higher the trait heritability, the greater the type I error differs from the declared value. The analysis of a large number of real traits confirmed these conclusions. It is necessary to take into account these results during the interpretation of the results of regional association analysis conducted on large regions using common genetic variants.

AB - Regional association analysis is a new statistical method which simultaneously considers all variants in a selected genome region. This method was created for the analysis of rare genetic variants, whose genotypes are determined by exome or genome sequencing. The gene is usually considered as a region. It was also proposed to use a regional analysis for testing of the association between a complex trait and a set of common variants genotyped by the panels developed for genome-wide association analysis. In this case, overlapping genome regions (sliding windows) are usually considered as a region. Since the size of such regions can be rather large, there is a risk of overestimation (inflation) of the test statistic and an increase in the type I error. In this work, the effect of the size of the region on the type I error was studied for traits with different heritability. The results of simulating experiments demonstrated that the physical size of the region but not the number of genetic variants in it is a limiting factor. The higher the trait heritability, the greater the type I error differs from the declared value. The analysis of a large number of real traits confirmed these conclusions. It is necessary to take into account these results during the interpretation of the results of regional association analysis conducted on large regions using common genetic variants.

KW - common genetic variants

KW - inflation factor

KW - quantitative traits

KW - regional association analysis

KW - simulation

KW - single nucleotide polymorphic markers

KW - type I error

UR - http://www.scopus.com/inward/record.url?scp=85043482439&partnerID=8YFLogxK

U2 - 10.1134/S1022795418010076

DO - 10.1134/S1022795418010076

M3 - Article

AN - SCOPUS:85043482439

VL - 54

SP - 250

EP - 258

JO - Russian Journal of Genetics

JF - Russian Journal of Genetics

SN - 1022-7954

IS - 2

ER -