Bilateria are the predominant clade of animals on Earth. Despite having evolved a wide variety of body plans and developmental modes, they are characterized by common morphological traits. By default, researchers have tried to link clade-specific genes to these traits, thus distinguishing bilaterians from non-bilaterians, by their gene content. Here we argue that it is rather biological processes that unite Bilateria and set them apart from their non-bilaterian sisters, with a less complex body morphology. To test this hypothesis, we compared proteomes of bilaterian and non-bilaterian species in an elaborate computational pipeline, aiming to search for a set of bilaterian-specific genes. Despite the limited confidence in their bilaterian specificity, we nevertheless detected Bilateria-specific functional and developmental patterns in the sub-set of genes conserved in distantly related Bilateria. Using a novel multi-species GO-enrichment method, we determined the functional repertoire of genes that are widely conserved among Bilateria. Analyzing expression profiles in three very distantly related model species—D. melanogaster, D. rerio and C. elegans—we find characteristic peaks at comparable stages of development and a delayed onset of expression in embryos. In particular, the expression of the conserved genes appears to peak at the phylotypic stage of different bilaterian phyla. In summary, our study illustrate how development connects distantly related Bilateria after millions of years of divergence, pointing to processes potentially separating them from non-bilaterians. We argue that evolutionary biologists should return from a purely gene-centric view of evolution and place more focus on analyzing and defining conserved developmental processes and periods.

Bilateria constitute a monophyletic group of organisms comprising about 99% of all living animals. Since their initial radiation about 540Mya they have evolved a plethora of traits and body forms allowing them to conquer almost any habitat on earth. There are only few truly uniting and shared morphological features retained across the phylum. Unsurprisingly, also the genetic toolkit of bilateria is highly diverged.In the light of this divergence we investigated if a set of Bilaterian-specific genes exists and, beyond this, if such genes are related with respect to function and expression patterns among organisms as distant as fly, worm and fish. Using a conservative pyramidal approach of orthology inference we first employed BLAST to extract a set of protein-coding genes which has orthologs in all major branches of Bilateria, but no homologs in non-Bilaterian species. To characterize the proteins with respect to function, we developed a new method for multi-species GO analysis and augmented it by a human-curated annotation based on an extensive literature search. Finally, we extracted characteristic developmental expression profiles for Bilateria from the detailed data available for three model organisms and we explored the relation between expression and function.We find 85 clusters of pan-bilaterian proteins. Although some of these proteins belong to common developmental processes, they cover a wide range of biological components, from transcription factors to metabolic enzymes. The 85 groups include myoD, an important regulator of mesodermal cell fate and muscle development, and prospero and several other factors involved in nervous system development.They reveal a so far unknown connection between bilaterian key innovations (mesoderm, complex nervous system) and the emergence of new genes in the ancestor of bilaterians. In addition, the majority of these † Current address: The Pirbright Institute,

Bilateria are the predominant clade of animals on earth. Despite having evolved a large variety of body-plans and developmental modes, they are characterized by common morphological traits. However, it is not clear if clade-specific genes can be linked to these traits, distinguishing bilaterians from non-bilaterians, with their less complex body morphology. Comparing proteomes of bilaterian and non-bilaterian species in an elaborate computational pipeline we aimed to find and define a set of of bilaterian-specific genes. Finding no high-confidence set of such genes, we nevertheless detected an evolutionary signal possibly uniting the highly diverse bilaterian taxa. Using a novel multi-species GO-enrichment method, we determined the functional repertoire of genes that are widely conserved among Bilateria. We found that these genes contribute to morphogenesis, neuronal-system and muscle development, processes that have been described as di↵erent between bilaterians and non-bilaterians. Analyzing gene expression profiles in three very distantly related bilaterina species, we find characteristic peaks at comparable stages of development and a delayed onset of expression in embryos. In particular, the expression of the conserved genes appears to peak at the phylotypic stage of di↵erent bilaterian phyla. In summary, our data underpin ⇤ corresponding authors the orthologue conjecture and illustrate how development connects distantly related Bilateria after millions of years of divergence, pointing to processes potentially separating them from non-bilaterians. 93 filtering criteria. 94 2.2. Orthologues conserved across divergent Bilateria 95 In our analysis, we included the highly divergent bilaterian model or-96 ganisms Caenorhabditis elegans, Drosophila melanogaster, and Danio rerio.97 4These have very well curated and annotated genomes and are therefore help-98 ful to find genes which are also functionally conserved since more than 99 500Myrs of independent evolution. To take advantage of this, we decided 100 to use the intersection of sets M and L (termed L 0 = L \ M ) for further 101 analysis. Compared to set L (125 clusters), this intersection lacks 31 clusters 102 missing a C. elegans ortholog and 9 clusters missing an ortholog from D. 103 melanogaster, resulting in a set of 85 orthogroups. Except for A. californica, 104 the least represented species in these clusters is S. purpuratus (genes of this 105 species occur in 63 of 85 clusters). By construction the model organisms (C. 106 elegans, D. melanogaster, D. rerio) are represented in all clusters in L 0 . On 107 average, a (bilaterian) species is represented in 93.1% of the clusters.108 2.3. Clade-specificity declines with data availability 109 A general drawback of our experimental procedure is its reliance on cor-110 rectly identified and annotated genes. Since the majority of the organisms 111 in our study are non-model organisms, they may su↵er from incomplete or 112 erroneous gene annotation. For example, we omitted the bilaterian species 113 A. californic...