One of the oldest problems in evolutionary biology remains largely unsolved. Which mutations generate evolutionarily relevant phenotypic variation? What kinds of molecular changes do they entail? What are the phenotypic magnitudes, frequencies of origin, and pleiotropic effects of such mutations? How is the genome constructed to allow the observed abundance of phenotypic diversity? Historically, the neo-Darwinian synthesizers stressed the predominance of micromutations in evolution, whereas others noted the similarities between some dramatic mutations and evolutionary transitions to argue for macromutationism. Arguments on both sides have been biased by misconceptions of the developmental effects of mutations. For example, the traditional view that mutations of important developmental genes always have large pleiotropic effects can now be seen to be a conclusion drawn from observations of a small class of mutations with dramatic effects. It is possible that some mutations, for example, those in cis-regulatory DNA, have few or no pleiotropic effects and may be the predominant source of morphological evolution. In contrast, mutations causing dramatic phenotypic effects, although superficially similar to hypothesized evolutionary transitions, are unlikely to fairly represent the true path of evolution. Recent developmental studies of gene function provide a new way of conceptualizing and studying variation that contrasts with the traditional genetic view that was incorporated into neo-Darwinian theory and population genetics. This new approach in developmental biology is as important for microevolutionary studies as the actual results from recent evolutionary developmental studies. In particular, this approach will assist in the task of identifying the specific mutations generating phenotypic variation and elucidating how they alter gene function. These data will provide the current missing link between molecular and phenotypic variation in natural populations.

Juvenile hormone (JH) signaling underpins both regulatory and developmental pathways in insects. However, the JH receptor is poorly understood. Methoprene tolerant (Met) and germ cell expressed (gce) have been implicated in JH signaling in Drosophila. We investigated the evolution of Met and gce across 12 Drosophila species and found that these paralogs are conserved across at least 63 million years of dipteran evolution. Distinct patterns of selection found using estimates of dN/dS ratios across Drosophila Met and gce coding sequences, along with their incongruent temporal expression profiles in embryonic Drosophila melanogaster, illustrate avenues through which these genes have diverged within the Diptera. Additionally, we demonstrate that the annotated gene CG15032 is the 5’ terminus of gce. In mosquitoes and beetles, a single Met-like homolog displays structural similarity to both Met and gce, and the intron locations are conserved with those of gce. We found that Tribolium and mosquito Met orthologs are assembled from Met- and gce-specific domains in a modular fashion. Our results suggest that Drosophila Met and gce experienced divergent evolutionary pressures following the duplication of an ancestral gce-like gene found in less derived holometabolous insects.

Pheromones, chemical signals that convey social information, mediate many insect social behaviors, including navigation and aggregation. Several studies have suggested that behavior during the immature larval stages of Drosophila development is influenced by pheromones, but none of these compounds or the pheromone-receptor neurons that sense them have been identified. Here we report a larval pheromone-signaling pathway. We found that larvae produce two novel long-chain fatty acids that are attractive to other larvae. We identified a single larval chemosensory neuron that detects these molecules. Two members of the pickpocket family of DEG/ENaC channel subunits (ppk23 and ppk29) are required to respond to these pheromones. This pheromone system is evolving quickly, since the larval exudates of D. simulans, the sister species of D. melanogaster, are not attractive to other larvae. Our results define a new pheromone signaling system in Drosophila that shares characteristics with pheromone systems in a wide diversity of insects.

Biological systems display extraordinary robustness. Robustness of transcriptional enhancers results mainly from clusters of binding sites for the same transcription factor, and it is not clear how robust enhancers can evolve loss of expression through point mutations. Here, we report the high-resolution functional dissection of a robust enhancer of the shavenbaby gene that has contributed to morphological evolution. We found that robustness is encoded by many binding sites for the transcriptional activator Arrowhead and that, during evolution, some of these activator sites were lost, weakening enhancer activity. Complete silencing of enhancer function, however, required evolution of a binding site for the spatially restricted potent repressor Abrupt. These findings illustrate that recruitment of repressor binding sites can overcome enhancer robustness and may minimize pleiotropic consequences of enhancer evolution. Recruitment of repression may be a general mode of evolution to break robust regulatory linkages.

Novel body structures are often generated by the redeployment of ancestral components of the genome. In this issue of Developmental Cell, Glassford et al. (2015) present a thorough analysis of the co-option of a gene regulatory network in the origin of an evolutionary novelty.

Postimplantation mammalian embryo culture methods have been generally inefficient and limited to brief periods after dissection out of the uterus. Platforms have been recently developed for highly robust and prolonged ex utero culture of mouse embryos from egg-cylinder stages until advanced organogenesis. These platforms enable appropriate and faithful development of pregastrulating embryos (E5.5) until the hind limb formation stage (E11). Late gastrulating embryos (E7.5) are grown in rotating bottles in these settings, while extended culture from pregastrulation stages (E5.5 or E6.5) requires a combination of static and rotating bottle cultures. In addition, sensitive regulation of O2 and CO2 concentration, gas pressure, glucose levels, and the use of a specific ex utero culture medium are critical for proper embryo development. Here, a detailed step-by-step protocol for extended ex utero mouse embryo culture is provided. The ability to grow normal mouse embryos ex utero from gastrulation to organogenesis represents a valuable tool for characterizing the effect of different experimental perturbations during embryonic development.

The mammalian body plan is established shortly after the embryo implants into the maternal uterus, and our understanding of post-implantation developmental processes remains limited. Although pre- and peri-implantation mouse embryos are routinely cultured in vitro, approaches for the robust culture of post-implantation embryos from egg cylinder stages until advanced organogenesis remain to be established. Here we present highly effective platforms for the ex utero culture of post-implantation mouse embryos, which enable the appropriate development of embryos from before gastrulation (embryonic day (E) 5.5) until the hindlimb formation stage (E11). Late gastrulating embryos (E7.5) are grown in three-dimensional rotating bottles, whereas extended culture from pre-gastrulation stages (E5.5 or E6.5) requires a combination of static and rotating bottle culture platforms. Histological, molecular and single-cell RNA sequencing analyses confirm that the ex utero cultured embryos recapitulate in utero development precisely. This culture system is amenable to the introduction of a variety of embryonic perturbations and micro-manipulations, the results of which can be followed ex utero for up to six days. The establishment of a system for robustly growing normal mouse embryos ex utero from pre-gastrulation to advanced organogenesis represents a valuable tool for investigating embryogenesis, as it eliminates the uterine barrier and allows researchers to mechanistically interrogate post-implantation morphogenesis and artificial embryogenesis in mammals.

Learning in deep neural networks is known to depend critically on the knowledge embedded in the initial network weights. However, few theoretical results have precisely linked prior knowledge to learning dynamics. Here we derive exact solutions to the dynamics of learning with rich prior knowledge in deep linear networks by generalising Fukumizu's matrix Riccati solution \citep{fukumizu1998effect}. We obtain explicit expressions for the evolving network function, hidden representational similarity, and neural tangent kernel over training for a broad class of initialisations and tasks. The expressions reveal a class of task-independent initialisations that radically alter learning dynamics from slow non-linear dynamics to fast exponential trajectories while converging to a global optimum with identical representational similarity, dissociating learning trajectories from the structure of initial internal representations. We characterise how network weights dynamically align with task structure, rigorously justifying why previous solutions successfully described learning from small initial weights without incorporating their fine-scale structure. Finally, we discuss the implications of these findings for continual learning, reversal learning and learning of structured knowledge. Taken together, our results provide a mathematical toolkit for understanding the impact of prior knowledge on deep learning.