Historical constraints on vertebrate genome evolution.

Recent analyses indicated that genes with larger effect of knockout or mutation and with larger probability to revert to single copy after whole genome duplication are expressed earlier in development. Here, we further investigate whether tissue specificity of gene expression is constrained by the age of origin of the corresponding genes. We use 38 metazoan genomes and a comparative genomic application system to integrate inference of gene duplication with expression data from 17,503 human genes into a strictly phylogenetic framework. We show that the number of anatomical systems in which genes are expressed decreases steadily with decreased age of the genes' first appearance in the phylogeny: the oldest genes are expressed, on average, in twice as many anatomical systems than the genes gained recently in evolution. These results are robust to different sources of expression data, to different levels of the anatomical system hierarchy, and to the use of gene families rather than duplication events. Finally, we show that the rate of increase in gene tissue specificity correlates with the relative rate of increase in the maximum number of cell types in the corresponding taxa. Although subfunctionalization and increase in cell type number throughout evolution could constitute, respectively, the proximal and ultimate causes of this correlation, the two phenomena are intermingled. Our analyses identify a striking historical constraint in gene expression: the number of cell types in existence at the time of a gene appearance (through duplication or de novo origination) tends to determine its level of tissue specificity for tens or hundreds of millions of years.