Dataset Information

Reducing bias of allele frequency estimates by modeling SNP genotype data with informative missingness.

ABSTRACT: The presence of missing single-nucleotide polymorphism (SNP) genotypes is common in genetic studies. For studies with low-density SNPs, the most commonly used approach to dealing with genotype missingness is to simply remove the observations with missing genotypes from the analyses. This naïve method is straightforward but is valid only when the missingness is random. However, a given assay often has a different capability in genotyping heterozygotes and homozygotes, causing the phenomenon of "differential dropout" in the sense that the missing rates of heterozygotes and homozygotes are different. In practice, differential dropout among genotypes exists in even carefully designed studies, such as the data from the HapMap project and the Wellcome Trust Case Control Consortium. Under the assumption of Hardy-Weinberg equilibrium and no genotyping error, we here propose a statistical method to model the differential dropout among different genotypes. Compared with the naïve method, our method provides more accurate allele frequency estimates when the differential dropout is present. To demonstrate its practical use, we further apply our method to the HapMap data and a scleroderma data set.

SUBMITTER: Lin WY

PROVIDER: S-EPMC3376470 | biostudies-literature | 2012

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Reducing bias of allele frequency estimates by modeling SNP genotype data with informative missingness.

Lin Wan-Yu WY Liu Nianjun N

Frontiers in genetics 20120618

The presence of missing single-nucleotide polymorphism (SNP) genotypes is common in genetic studies. For studies with low-density SNPs, the most commonly used approach to dealing with genotype missingness is to simply remove the observations with missing genotypes from the analyses. This naïve method is straightforward but is valid only when the missingness is random. However, a given assay often has a different capability in genotyping heterozygotes and homozygotes, causing the phenomenon of "d ...[more]

PMID: 22719749

Similar Datasets

Project description:Microsatellite loci play an important role as markers for identification, disease gene mapping, and evolutionary studies. Mutation rate, which is of fundamental importance, can be obtained from interspecies comparisons, which, however, are subject to ascertainment bias. This bias arises, for example, when a locus is selected on the basis of its large allele size in one species (cognate species 1), in which it is first discovered. This bias is reflected in average allele length in any noncognate species 2 being smaller than that in species 1. This phenomenon was observed in various pairs of species, including comparisons of allele sizes in human and chimpanzee. Various mechanisms were proposed to explain observed differences in mean allele lengths between two species. Here, we examine the framework of a single-step asymmetric and unrestricted stepwise mutation model with genetic drift. Analysis is based on coalescent theory. Analytical results are confirmed by simulations using the simuPOP software. The mechanism of ascertainment bias in this model is a tighter correlation of allele sizes within a cognate species 1 than of allele sizes in two different species 1 and 2. We present computations of the expected average allele size difference, given the mutation rate, population sizes of species 1 and 2, time of separation of species 1 and 2, and the age of the allele. We show that when the past demographic histories of the cognate and noncognate taxa are different, the rate and directionality of mutations affect the allele sizes in the two taxa differently from the simple effect of ascertainment bias. This effect may exaggerate or reverse the effect of difference in mutation rates. We reanalyze literature data, which indicate that despite the bias, the microsatellite mutation rate estimate in the ancestral population is consistently greater than that in either human or chimpanzee and the mutation rate estimate in human exceeds or equals that in chimpanzee with the rate of allele length expansion in human being greater than that in chimpanzee. We also demonstrate that population bottlenecks and expansions in the recent human history have little impact on our conclusions.

Project description:Highly specific amplification of complex DNA pools without bias or template-independent products (TIPs) remains a challenge. We have developed a procedure using phi29 DNA polymerase and trehalose and optimized control of amplification to create micrograms of specific amplicons without TIPs from down to sub-femtograms of DNA. The amplicons from 5 ng and 0.5 ng DNA, which were from originally good quality of gDNA (05-050), or partially degraded gDNA (04-018), were validate with Illumina HumanHap550-Duo Genotyping Beadchip. As seen in (Suppl. Table 5a), the call rates (97.30% to 99.07%) and accuracy or concordance ( > 99.85% for the SNPs called in both amplicon and natural reference) for 5 ng derived amplicons with both Wpa and Gv2 were close to each other and close to native gDNA (call rate: 98.3% to 99.75%). These call rates were better than a recent report (amplicon 95.9% vs. un-amplified 98.5%), in which the early kit Repli-g 625S was applied, and re-genotyping was performed when the performance was low and duplicate samples were filtered for the highest call rate. The genotyping accuracy of Wpa was actually in the same range as the variation in technical replicates with similar SNP typing arrays (99.87% and 99.88%, replicated Affymetrix array, or between Affymetrix and Illumina arrays). Importantly, the genotyping concordance for amplicons generated from 0.5 ng with Wpa (99.88% and 99.69%) were also close to the technical replicates. In this case, the call rates of Wpa were slightlyreduced compared to that with 5 ng input, but the call rate for the partially degraded sample 04-018, was modestly improved over Gv2 (92.06 % vs. 90.53%). Wpa data also showed some amplification non-uniformity among different locations, resulting in some “artificial CNVs” similar to Gv2 (exampled as in Suppl. Fig. 5 and Suppl. Table 6), with the outputs obtained by taking unamplified gDNAs as their reference. This imbalance however was consistent and reproducible for each method but different between Wpa and Gv2. These artificial CNVs can be efficiently cancelled if pair-wise amplified test and reference are compared, as observed in CGH result (Fig. 4 and Suppl. Fig. 4), also supported by others {Pugh 2008}. It is interesting to note that the representation of chromosomal terminal sequences was greatly improved with Wpa compared with Gv2 (Fig. 5), and that some of these regions were significantly under-amplified or even lost with Gv2 (Suppl. Fig. 5 and Suppl. Table 6, 7), as also independently reported recently {Pugh 2008}. This occurred especially in the terminal 3 to 5 Mb and sometimes extended to 10 Mb in many chromosome termini, and was particularly serious when low levels or degraded DNA was as input. An analysis for 5 Mb termini is shown (Suppl. Table 5b calculated all involved SNPs as a cohort. Fig. 5 and Suppl. Tables 6 and 7 were the result for each chromosome terminus). Importantly, the SNP typing was also greatly improved, outstandingly exemplified by the amplicons of 0.5 ng input for the partially degraded 04-018, with Wpa versus Gv2 call rate of 91.9% vs. 84.45% and accuracy of 99.57% vs. 98.62%. The result also showed that these terminal regions underrepresentation in Gv2 was not absolutely associated with the distance-to-end, but possibly was a sequence related issue. Keywords: Whole-pool amplification, whole genome SNP typing

Project description:The absence or 'missingness' of single nucleotide polymorphism (SNP) assay values because of genotype or related factors of interest may bias association and other studies. Missingness was determined for the Type 1 Diabetes Genetics Consortium (T1DGC) Major Histocompatibility Complex (MHC) data and was found to vary across the region, ranging up to 11.1% of the non-null proband SNPs, with a median of 0.3%. We consider factors related to missingness in the T1DGC data and briefly assess its possible influence on association studies.We assessed associations of missingness in the SNP assay data with human leucocyte antigen (HLA) genotype of the individual and with SNP genotypes of the parents. Within-cohort analyses were combined (over all cohorts) using (i) Mantel-Haenszel tests for two-by-two tables or (ii) by combining test statistics for larger tables and regression models. Mixed effect regression models were used to assess association of the SNP genotypes with affected status of the offspring after adjustment for parental SNP genotypes, cohort membership and HLA genotypes. Log-linear models were used to assess association of missingness in the unaffected sib assays with SNP genotypes of the probands.Missingness of SNP values near the HLA class I (A, B and C) and class II (DR, DQ and DP) loci is strongly associated with carriage of corresponding HLA genotypes within these groups. Similar associations pertain to missing values among the microsatellite data. In at least some of these cases, regions of missingness coincided with known deletion regions corresponding to the associated HLA haplotype. We conjecture that other regions of associated missingness may point to similar haplotypic deletions. Analysis of association patterns of SNP genotypes with affected status of offspring does not indicate strong informative missingness. However, association of missingness in proband data with parental SNP genotypes may impact transmission disequilibrium test (TDT)-type analyses. Comparisons of affected and unaffected siblings point to possible susceptibility regions additional to the classical HLA-DR3/4 alleles near BAT4-LY6G5B-BAT5 and NOTCH4.Potentially informative missingness in SNP assay values in the MHC region may impact on association and related analyses based on the T1DGC data. These results suggest that it would be prudent to assess the degree to which missingness may abrogate assessed SNP disease markers in such studies. Initial analyses based on comparison of affected and unaffected status in offspring suggest that at least these may be little affected.

Dataset Information

Reducing bias of allele frequency estimates by modeling SNP genotype data with informative missingness.

Publications

Reducing bias of allele frequency estimates by modeling SNP genotype data with informative missingness.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets