lme4qtl: linear mixed models with flexible covariance structure for genetic studies of related individuals

Abstract

Background: Quantitative trait locus (QTL) mapping in genetic data often involves analysis of correlated observations, which need to be accounted for to avoid false association signals. This is commonly performed by modeling such correlations as random effects in linear mixed models (LMMs). The R package lme4 is a well-established tool that implements major LMM features using sparse matrix methods; however, it is not fully adapted for QTL mapping association and linkage studies. In particular, two LMM features are lacking in the base version of lme4: the definition of random effects by custom covariance matrices; and parameter constraints, which are essential in advanced QTL models. Apart from applications in linkage studies of related individuals, such functionalities are of high interest for association studies in situations where multiple covariance matrices need to be modeled, a scenario not covered by many genome-wide association study (GWAS) software.

A small‐sample kernel association test for correlated data with application to microbiome association studies

“…However, along with the benefits of these complex designs come statistical challenges, particularly regarding how to address the correlation among related samples. The null LMM in CSKAT can be fitted by many softwares such as the lme4 R package (Bates, Mächler, Bolker, & Walker, 2015) and lme4qtl R package (Ziyatdinov et al, 2018), allowing for many different types of correlated data including longitudinal data and pedigree data by specifying the kinship matrix. In this paper, we propose CSKAT to test the association between microbiome compositions and an outcome of interest, where microbiome samples and outcome measurements within the same cluster are related to each other.…”

“…Besides the improved power using small-sample correction, CSKAT is more flexible than famSKAT and LSKAT in that it can test many types of correlated data. The null LMM in CSKAT can be fitted by many softwares such as the lme4 R package (Bates, Mächler, Bolker, & Walker, 2015) and lme4qtl R package (Ziyatdinov et al, 2018), allowing for many different types of correlated data including longitudinal data and pedigree data by specifying the kinship matrix. Although we illustrate CSKAT using microbiome data as an example, the methodology we developed is general and would have substantial applicability to many types of genetic and genomic data (Chen et al, 2013;Wang et al, 2017;Zhan, Girirajan, Zhao, Wu & Ghosh, 2016).…”

Delineating the Genetic Component of Gene Expression in Major Depression

“…A linear mixed model was used to test for an association between z-scores indicating non-zero association for each feature and gene set membership, while adjusting for gene length and numbers of SNPs within the gene region, and accounting for correlation between features. Linear mixed models were fitted using the R package lme4qtl (22). TWAS-GSEA used 7,246 hypothesis-free gene sets from MSigDB (v6.1) and 76 candidate gene sets from (5), which included FMRP binding partners, de novo mutations, GWAS top SNPs, and ion channels.…”

High-resolution QTL mapping with Diversity Outbred mice identifies genetic variants that impact gut microbiome composition

“…The unit of inference for phenotypes was the relative abundance of each taxon in each individual, while the units of genetic inference were the SNP genotypes at each of 57,973 sites for each mouse using the mouse array. We estimated narrow-sense "SNP" heritability (h 2 ) using a linear mixed model in R-package lme4qtl [Ziyatdinov et al 2018]. A linear mixed model was used to predict whether the effects of the autosomal genotype on the phenotype is proportional to the genetic similarity between the mice, after adjustment for known factors.…”