Genetic analysis of transcriptional regulation is a rapidly emerging field of investigation that promises to shed light on the regulatory networks that control gene expression. Although a number of such studies have been carried out, the nature and extent of the heritability of gene expression traits have not been well established. We describe the inheritance of transcript levels in liver tissue in the first filial (F1) generation of mice obtained from reciprocal crosses between the common inbred strains A/J and C57BL/6J. We obtain estimates of genetic and technical variance components from these data and demonstrate that shrinkage estimators can increase detectable heritability. Estimates of heritability vary widely from transcript to transcript, with one-third of transcripts showing essentially no heritability (<0.01) and one-quarter showing very high heritability (>0.50). Roughly half of all transcripts are differentially expressed between the two parental strains. Most transcripts show an additive pattern of inheritance. Dominance effects were observed for 20% of transcripts and a small number of transcripts were identified as showing an overdominance mode of inheritance. In addition, we identified 314 transcripts with expression levels that differ between the reciprocal F1 animals. These genes may be related to maternal effect.
Design file for statistical modeling of the data: AffyDesignFinal.txt
RMA-preprocessed original data file - rma.expr.affy.txt
RMA processed data file with SNP probes removed - rma.expr.customcdf.txt
The script files may contain some early-stage explorations in addition to what was presented in the paper. You are encouraged to contact Xiangqin Cui (firstname.lastname@example.org) before making serious effort to reproduce the results.
Script for carrying out all the statistical tests on original data: script.paper.revise.Tests.R
Script for the statistical tests on data with SNP probes removed: script.paper.revise.p.snp.7.26.R
Script for generating the figures in the paper: script.AffyEredo.paperFig.revise.v9.R
Two custom functions used in the analysis: