Assessing natural variations in gene expression in humans by comparing with monozygotic twins using microarrays

Quantitative variationin gene expressionin humans is the outcome of various factors, including differences in genetic background, gender, age, and environment. However, the extent of the influence of these factors on gene expressionis not clear. We attempted to address this issue by carrying out gene expression profilingin blood leukocytes with 13 individuals (including 5 pairs of monozygotic twins) on 10,000 genes using HG-U95Av2 oligonucleotide microarrays. The proportion of differentially expressed genes between monozygotic twins was low (up to 1.76%). Most of the variationsbelonged to the least variable category. These genes, exhibiting “ random variations,” did not show clear preference to any functional class, although “signaling and communication” and “immune and related functions” generallytopped the list. The extent of variationin gene expressionincreased in comparisons between unrelated individuals (up to 14.13%). Most of the genes (89%) exhibiting random variationsin twins also varied in expression in unrelated individuals. As with twins, signaling and communication topped the list, and substantial variationswere observed in all three categories: least variable, moderately variable, and most variable. An important outcome of this study was that the housekeeping geneswere nearly insensitive to random variationsbut appeared to be more susceptible to genetic differences. However, the highly expressed housekeeping genesexhibited low variationand appeared to be insensitive to all known factors. Gene expression profilingin monozygotic twins can provide useful data for the assessment of natural variationin gene expressionin humans.