Gene expression has been shown to vary within and between populations of a variety of different species, ranging from yeast to humans. Between yeast strains of Saccharomyces cerevisiae, for example, 24% of genes analysed showed significant variation (Brem et al. 2002). Similar results were found in the fly Drosophila mela-nogaster where 25% of genes showed variation in expression (Jin et al. 2001) and telefost fish of the genus Fundulus where 18% of genes analysed showed significant variation in expression between individuals in a population (Oleksiak et al. 2002). In human populations, significant variation in gene expression was observed between individuals with evidence of familial aggregation and tissue-specific variation (Enard et al. 2002; Cheung et al. 2003c; Schadt et al. 2003; Monks et al. 2004; Goring et al. 2007; Stranger et al. 2007b; Emilsson et al. 2008).
In an early study utilizing microarray technology to simultaneously analyse gene expression for many different genes in a high throughput manner, Cheung and colleagues analysed gene expression in 35 lymphob-lastoid cell lines established from unrelated individuals of European descent as part of the CEPH collection (Box 11.1) (Cheung et al. 2003c). The majority of genes showed greater variation between than within individuals. The top 5% most variable genes in terms of expression were scattered across the genome and included previously characterized genes such as HLA-DRB1 in the major histocompatibility complex (MHC), known to show significant variation in expression (Section 12.3.4) (Cheung et al. 2003c). When the expression of five genes that were highly variable based on the expression array dataset were quantified by real time quantitative reverse transcription polymerase chain reaction (RT-PCR), much greater variation was observed between unrelated individuals than between siblings from the same family, who in turn showed greater variation than monozygotic twins (Fig. 11.1) (Cheung et al. 2003c).
Schadt and colleagues also analysed gene expression among lymphoblastoid cell lines from CEPH families and found significant evidence of heritability (Schadt et al. 2003; Monks et al. 2004). An initial study of 56 individuals was based on four CEPH families who comprised pedigree founders (grandparents), parents, and children. This defined 2726 genes among the 24 479 genes analysed which were differentially expressed in at least half of the founders. Evidence of a significant heritable component was found among 25% of these genes (Schadt et al. 2003). In a larger study of 167 individuals among 15 CEPH families, 2430 differentially expressed genes were defined with 31% showing significant heritability at a false discovery rate of 0.05; the median heritability was 0.34 (Monks et al. 2004). Interestingly, this group of genes were enriched for genes with immune-related function. Further evidence of significant heritability was seen among lymphoblastoid cell lines established from family trios (parents and child) in the International HapMap Project (Stranger et al. 2007b).
Strong evidence for heritability in gene expression came from a more recent large scale analysis of primary blood cells among a cohort of 1240 Mexican Americans, most of whom came from 30 extended families of up to four generations (Goring et al. 2007). Gene expression
Epstein-Barr virus immortalized lymphoblastoid cell lines have proved a powerful resource for genetic analysis of gene expression, established from individuals and families such as the Centre d'Etude du Polymorphisme Humain (CEPH) reference family panel (Dausset et al. 1990) and as part of the International HapMap Project (Box 9.1). Lymphoblastoid cell lines provide an ongoing source of a single cell type in which growth and culture conditions can be standardized to minimize environmental variation. Moreover large numbers of cells can be harvested for analysis and experimental replication achieved. The availability of dense SNP genotyping data for collections of lymphoblastoid cell lines established from diverse human populations such as the HapMap Project made these cells particularly suitable for application of the gen-etical genomics approach (Section 11.3.1). The HapMap Project has greatly advanced our understanding of fine scale genetic variation and genomic architecture with several million common SNPs defined and genotyped across diverse populations with the primary goal of facilitating and enabling genome-wide association studies of common disease (Section 9.2.4). The panels of lymphoblastoid cell lines generated as part of the HapMap Project to provide an ongoing source of genomic DNA have subsequently been integral to many analyses of the genetics of gene expression. As lymphoblastoid cell lines are established from circulating peripheral blood lymphocytes they are a relevant cell type for many human traits and show detectable expression of a large number of genes. However, there are important caveats to their use, notably the fact that they are immortalized with Epstein-Barr virus which will itself modulate expression of some genes and may show heterogeneity between individual cell lines with viral load and other factors. Further potential issues on prolonged cell culture relate to the establishment of clonality and differential DNA methylation (Pastinen et al. 2006). It is also important to remember that functionally important genetic variation is often context-specific, and may be only manifested in specific cell or tissue types under specific conditions of stimulation or other environmental variables.
for 18 519 genes was analysed for circulating peripheral blood lymphocytes with a variance components-based heritability analysis performed that included age and sex. This demonstrated that 85% of transcripts show significant heritability with a false discovery rate of 5%. The median heritability of all expressed transcripts was 22.5%. These striking results in a large sample size analysing gene expression in primary blood cells reinforce the substantial genetic component to variation in observed gene expression.
Among primates, when gene expression was analysed at the RNA and protein level for 12 000 genes in blood and liver samples, patterns of gene expression in humans and chimpanzees (Pan troglodytes) were found to be more closely related to each other than to that seen in macaques (Macaca mulatta) (Enard et al. 2002). By contrast, when gene expression in brain samples was analysed, chimpanzees were more similar to orangutans (Pongo pygmaeus) and macaques than to humans, highlighting tissue-specific differences that may be present, and in this study demonstrating accelerated differences in gene expression between human and other primates when brain tissue was analysed (Enard et al. 2002).
Was this article helpful?