Relationship between gene expression and GC-content in mammals: statistical significance and biological relevance

Hum Mol Genet. 2005 Feb 1;14(3):421-7. doi: 10.1093/hmg/ddi038. Epub 2004 Dec 8.

Abstract

Mammalian chromosomes are characterized by large-scale variations of DNA base composition (the so-called isochores). In contradiction with previous studies, Lercher et al. (Hum. Mol. Genet., 12, 2411, 2003) recently reported a strong correlation between gene expression breadth and GC-content, suggesting that there might be a selective pressure favoring the concentration of housekeeping genes in GC-rich isochores. We reassessed this issue by examining in human and mouse the correlation between gene expression and GC-content, using different measures of gene expression (EST, SAGE and microarray) and different measures of GC-content. We show that correlations between GC-content and expression are very weak, and may vary according to the method used to measure expression. Such weak correlations have a very low predictive value. The strong correlations reported by Lercher et al. (2003) are because of the fact that they measured variables over neighboring genes windows. We show here that using gene windows artificially enhances the correlation. The assertion that the expression of a given gene depends on the GC-content of the region where it is located is therefore not supported by the data.

Publication types

  • Comparative Study

MeSH terms

  • Animals
  • Base Composition
  • Expressed Sequence Tags
  • GC Rich Sequence*
  • Gene Expression
  • Genome*
  • Humans
  • Isochores / genetics
  • Mice
  • Models, Genetic
  • Oligonucleotide Array Sequence Analysis

Substances

  • Isochores