Iam hiQ—a novel pair of accuracy indices for imputed genotypes

Abstract

Background Imputation of untyped markers is a standard tool in genome-wide association studies to close the gap between directly genotyped and other known DNA variants. However, high accuracy with which genotypes are imputed is fundamental. Several accuracy measures have been proposed and some are implemented in imputation software, unfortunately diversely across platforms. In the present paper, we introduce Iam hiQ, an independent pair of accuracy measures that can be applied to dosage files, the output of all imputation software. Iam (imputation accuracy measure) quantifies the average amount of individual-specific versus population-specific genotype information in a linear manner. hiQ (heterogeneity in quantities of dosages) addresses the inter-individual heterogeneity between dosages of a marker across the sample at hand. Results Applying both measures to a large case–control sample of the International Lung Cancer Consortium (ILCCO), comprising 27,065 individuals, we found meaningful thresholds for Iam and hiQ suitable to classify markers of poor accuracy. We demonstrate how Manhattan-like plots and moving averages of Iam and hiQ can be useful to identify regions enriched with less accurate imputed markers, whereas these regions would by missed when applying the accuracy measure info (implemented in IMPUTE2). Conclusion We recommend using Iam hiQ additional to other accuracy scores for variant filtering before stepping into the analysis of imputed GWAS data.

ImputAccur: fast and user-friendly calculation of genotype-imputation accuracy-measures

“…According to the accuracy indices Iam chance (0.546) and Iam HWE (0.448) it is reasonable to assume that about half of the information contained in dosages comes from the true (but unknown) genotypes of the individuals in the sample, the other half comes from the population used as a reference for genotype imputation. The values are borderline near the recommended threshold of 0.47 [ 6 ]. The difference between Iam chance and Iam HWE is the “anchor point” used, which is either purely populations-related dosages (Hardy–Weinberg Equilibrium HWE, taking MAF into account) or pure chance (1/3 probability for each of the three possible genotypes of a SNP).…”

“…Several quality indices have been developed and are routinely applied in studies [ 2 , 3 , 5 ]. These comprise inter alia MACH’s r 2 , BEAGLES’s r 2 , IMPUTE2’s info or the recently proposed Iam hiQ, including a regional classification across markers [ 6 ]. Unfortunately, these accuracy measurements are implemented on different platforms.…”

“…Furthermore, ImputAccur classifies markers to be located in a “cold”, “tepid”, “hot”, or “very hot” region, the last indicating massively inaccurate imputation, as outlined by Rosenberger et al [ 6 ] Details and equations of these accuracy indices and the classification are summarized in the Additional file 1 . The validity of the calculations was tested by comparison with output files from IMPUTE2 (for info) and with known results of carefully selected sample data (all indices).…”

ImputAccur: fast and user-friendly calculation of genotype-imputation accuracy-measures

“…1 rs00001 913 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 2 rs00002 402 0.01 0.99 0 0.01 0.99 0 0.01 0.99 0 0.01 0.99 0 0.01 0.99 0 3 rs00003 644 0.333 0.334 0.333 0.333 0.334 0.333 0.333 0.334 0.333 0.333 0.334 0.333 0.333 0.334 0.333 4 rs00004 222 0.25 0.5 0.25 0.25 0.5 0.25 0.25 0.5 0.25 0.25 0.5 0.25 0.25 0.5 0.25 5 rs00005 221 0.47 0.18 0.35 0.89 0.02 0.09 0.03 0.96 0.01 0.94 0 0.06 0.62 0.34 0.04 6 rs00006 955 0.975 0.002 0.023 0.52 0.154 0.326 0.309 0.21 0.481 0.48 0.509 0.011 0.969 0.004 0.027 7 rs00007 518 0.63 0. A regional classification across markers (as outlined by Rosenberger et al (2022)) is given in the column accuracy.…”

“…Several quality indices have been developed and are routinely applied in studies (Marchini and Howie, 2010;Das et al, 2018;Browning and Browning, 2009). These comprise inter alia MACH's r², BEAGLES's r², IMPUTE2's info or the recently proposed Iam hiQ, including a regional classification across markers (Rosenberger et al, 2022). Unfortunately, these accuracy measurements are implemented on different platforms.…”