Abstract
One central goal of genome biology is to understand how the usage of the genome differs between organisms. Our knowledge of genome composition, needed for downstream inferences, is critically dependent on gene annotations, yet problems associated with gene annotation and assembly errors are usually ignored in comparative genomics. Here we analyze the genomes of 68 species across all animal groups and some single-cell eukaryotes for general trends in genome usage and composition, taking into account problems of gene annotation. We show that, regardless of genome size, the ratio of introns to intergenic sequence is comparable across essentially all animals, with nearly all deviations dominated by increased intergenic sequence. Genomes of model organisms have ratios much closer to 1:1, suggesting that the majority of published genomes of non-model organisms are underannotated and consequently omit substantial numbers of genes, with likely negative impact on evolutionary interpretations. Finally, our results also indicate that most animals transcribe half or more of their genomes arguing against differences in genome usage between animal groups, and also suggesting that the transcribed portion is more dependent on genome size than previously thought.
Author's Summary Within our anthropocentric genomic framework, many analyses try to define humans, mammals, or vertebrates relative to the so-called “lower” animals. This implicitly posits that vertebrates are complex organisms with large genomes and invertebrates are simple organisms with small genomes. This has the problem that genome size is therefore presumed to correlate with complexity and ignores any unknown complexity of vast numbers of invertebrate groups, many with large genomes. Animals vary widely in genome size, by almost three orders of magnitude, but when sequencing new animal genomes preference is given to those with smaller genomes for reasons of cost. In trying to understand how genomes are used in general, there is an added layer of complication from quality of the assembly and annotation. We have examined genome usage across a wide range of animals and have described ways to account for errors of low-quality annotations. We also show that the genomes of invertebrates and vertebrates are not so different, and that when large-genome invertebrates are considered, the fraction of the genome that is genes appears to be strongly influenced by genome size.
