Evolution and functional classification of vertebrate gene deserts

Abstract

Large tracts of the human genome, known as gene deserts, are devoid of protein-coding genes. Dichotomy in their level of conservation with chicken separates these regions into two distinct categories, stable and variable. The separation is not caused by differences in rates of neutral evolution but instead appears to be related to different biological functions of stable and variable gene deserts in the human genome. Gene Ontology categories of the adjacent genes are strongly biased toward transcriptional regulation and development for the stable gene deserts, and toward distinctively different functions for the variable gene deserts. Stable gene deserts resist chromosomal rearrangements and appear to harbor multiple distant regulatory elements physically linked to their neighboring genes, with the linearity of conservation invariant throughout vertebrate evolution.

Figure 1.

Chromosome coverage by gene deserts (in blue) and regular intergenic regions (in red).

Figure 2.

Ratio of different categories of repetitive elements populating different human genomic regions. Gene deserts are in blue, average counts for the human genome are in gray, regular intergenic regions are in light blue, and gene-rich regions are in red.

Figure 3.

Correlation of nonrepetitive conservation of the human gene deserts with chicken (A) and mouse (B) versus repeat content. Red color depicts the stable gene deserts that are >2% conserved with chicken throughout their length. Negative correlation of the nonrepetitive conservation level and repeat content is very weak in both chicken and mouse comparisons, with R² reaching 0.06 in the case of mouse comparisons. Two stable gene deserts located upstream of the DACH1 and OTX2 gene and two other ones surrounding the SOX2 gene are in yellow.

Figure 4.

Percentage of genes with UTRs conserved in chicken (vertical axis) versus the gene density (based on RefSeq annotation; in genes per 1 Mb of sequence as plotted at the horizontal axis). Red dots describe different human chromosomes.

Figure 5.

Density of synteny breakpoints per 1 Mb of sequence. Human–mouse comparisons are in orange; human–chicken in lilac.

Figure 6.

Length of orthologous stable gene desert counterparts in the chicken and mouse genomes compared with the human genome. Gene deserts from chicken microchromosomes are in red.

Figure 7.

Distribution of stable gene deserts in the chicken genome (plotted as red lines). Chicken chromosomes are grouped into macro, intermediate, micro, and sex categories with the numerical characterization of average chromosome coverage by the stable gene deserts.