Dataset Information

Correcting for population structure and kinship using the linear mixed model: theory and extensions.

ABSTRACT: Population structure and kinship are widespread confounding factors in genome-wide association studies (GWAS). It has been standard practice to include principal components of the genotypes in a regression model in order to account for population structure. More recently, the linear mixed model (LMM) has emerged as a powerful method for simultaneously accounting for population structure and kinship. The statistical theory underlying the differences in empirical performance between modeling principal components as fixed versus random effects has not been thoroughly examined. We undertake an analysis to formalize the relationship between these widely used methods and elucidate the statistical properties of each. Moreover, we introduce a new statistic, effective degrees of freedom, that serves as a metric of model complexity and a novel low rank linear mixed model (LRLMM) to learn the dimensionality of the correction for population structure and kinship, and we assess its performance through simulations. A comparison of the results of LRLMM and a standard LMM analysis applied to GWAS data from the Multi-Ethnic Study of Atherosclerosis (MESA) illustrates how our theoretical results translate into empirical properties of the mixed model. Finally, the analysis demonstrates the ability of the LRLMM to substantially boost the strength of an association for HDL cholesterol in Europeans.

SUBMITTER: Hoffman GE

PROVIDER: S-EPMC3810480 | biostudies-literature | 2013

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Correcting for population structure and kinship using the linear mixed model: theory and extensions.

Hoffman Gabriel E GE

PloS one 20131028 10

Population structure and kinship are widespread confounding factors in genome-wide association studies (GWAS). It has been standard practice to include principal components of the genotypes in a regression model in order to account for population structure. More recently, the linear mixed model (LMM) has emerged as a powerful method for simultaneously accounting for population structure and kinship. The statistical theory underlying the differences in empirical performance between modeling princ ...[more]

PMID: 24204578

Similar Datasets

Project description:Background: Genome-wide detection of single feature polymorphisms (SFP) in swine using transcriptome profiling of day 25 placental RNA by contrasting probe intensities from either Meishan or an occidental composite breed with Affymetrix porcine microarrays is presented. A linear mixed model analysis was used to identify significant breed-by-probe interactions. Results: Gene specific linear mixed models were fit to each of the log2 transformed probe intensities on these arrays, using fixed effects for breed, probe, breed-by-probe interaction, and a random effect for array. After surveying the day 25 placental transcriptome, 789 probes with a q-value ≤ 0.05 and |fold change| ≥ 2 for the breed-by-probe interaction were identified as candidates containing SFP. To address the quality of the bioinformatics approach, universal pyrosequencing assays were designed from Affymetrix exemplar sequences to independently assess polymorphisms within a subset of probes. Of those probes sampled from high-, medium-, and low-ranking categories, 20 of 27 were confirmed by pyrosequencing to contain SFPs. In most cases, the 25-mer probe sequence printed on the microarray diverged from Meishan, not occidental crosses. This analysis was used to define a set of highly reliable predicted SFPs according to their probability scores. Conclusions: By this method we detected transition and transversion single nucleotide polymorphisms, as well as insertions/deletions. These results demonstrate that this approach can identify polymorphisms between two breeds and/or lines of any species for which a short oligonucleotide array is available, and can be used to rapidly develop markers for genetic mapping and association analysis in species where high density genotyping platforms are otherwise unavailable. SNPs and INDELS discovered by this approach have been publicly deposited in NCBI’s SNP repository dbSNP. This method is an attractive bioinformatics tool for uncovering breed-by-probe interactions, for rapidly identifying expressed SNPs, for investigating potential functional correlations between gene expression and breed polymorphisms, and is robust enough to be used on any Affymetrix gene expression platform. Keywords: Transcriptional profiling of Day 25 porcine placentas

Project description:Background: Genome-wide detection of single feature polymorphisms (SFP) in swine using transcriptome profiling of day 25 placental RNA by contrasting probe intensities from either Meishan or an occidental composite breed with Affymetrix porcine microarrays is presented. A linear mixed model analysis was used to identify significant breed-by-probe interactions. Results: Gene specific linear mixed models were fit to each of the log2 transformed probe intensities on these arrays, using fixed effects for breed, probe, breed-by-probe interaction, and a random effect for array. After surveying the day 25 placental transcriptome, 789 probes with a q-value â¤ 0.05 and |fold change| â¥ 2 for the breed-by-probe interaction were identified as candidates containing SFP. To address the quality of the bioinformatics approach, universal pyrosequencing assays were designed from Affymetrix exemplar sequences to independently assess polymorphisms within a subset of probes. Of those probes sampled from high-, medium-, and low-ranking categories, 20 of 27 were confirmed by pyrosequencing to contain SFPs. In most cases, the 25-mer probe sequence printed on the microarray diverged from Meishan, not occidental crosses. This analysis was used to define a set of highly reliable predicted SFPs according to their probability scores. Conclusions: By this method we detected transition and transversion single nucleotide polymorphisms, as well as insertions/deletions. These results demonstrate that this approach can identify polymorphisms between two breeds and/or lines of any species for which a short oligonucleotide array is available, and can be used to rapidly develop markers for genetic mapping and association analysis in species where high density genotyping platforms are otherwise unavailable. SNPs and INDELS discovered by this approach have been publicly deposited in NCBIâs SNP repository dbSNP. This method is an attractive bioinformatics tool for uncovering breed-by-probe interactions, for rapidly identifying expressed SNPs, for investigating potential functional correlations between gene expression and breed polymorphisms, and is robust enough to be used on any Affymetrix gene expression platform. Keywords: Transcriptional profiling of Day 25 porcine placentas 6 samples: Breed (Meishan, Occidental) Placental Tissue, Day 25

Dataset Information

Correcting for population structure and kinship using the linear mixed model: theory and extensions.

Publications

Correcting for population structure and kinship using the linear mixed model: theory and extensions.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets