The evolution of gene expression responses are a critical component of adaptation to variable environments. Predicting how DNA sequence influences expression is challenging because the genotype to phenotype map is not well resolved for cis regulatory elements, transcription factor binding, regulatory interactions, and epigenetic features, not to mention how these factors respond to environment. We tested if flexible machine learning models could learn some of the underlying cis-regulatory genotype to phenotype map. We tested this approach using cold-responsive transcriptome profiles in 5 diverse Arabidopsis thaliana accessions. We first tested for evidence that cis regulation plays a role in environmental response, finding 14 and 15 motifs that were significantly enriched within the up- and down-stream regions of cold-responsive differentially regulated genes (DEGs). We next applied convolutional neural networks (CNNs), which learn de novo cis-regulatory motifs in DNA sequences to predict expression response to environment. We found that CNNs predicted differential expression with moderate accuracy, with evidence that predictions were hindered by biological complexity of regulation and the large potential regulatory code. Overall, DEGs between specific environments can be predicted based on variation in cis-regulatory sequences, although more information needs to be incorporated and better models may be required.

The emergence of infectious agents with pandemic potential present scientific challenges from detection to data interpretation to understanding determinants of risk and forecasts. Mathematical models could play an essential role in how we prepare for future emergent pathogens. Here, we describe core directions for expansion of the existing tools and knowledge base, including: using mathematical models to identify critical directions and paths for strengthening data collection to detect and respond to outbreaks of novel pathogens; expanding basic theory to identify infectious agents and contexts that present the greatest risks, over both the short and longer term; by strengthening estimation tools that make the most use of the likely range and uncertainties in existing data; and by ensuring modelling applications are carefully communicated and developed within diverse and equitable collaborations for increased public health benefit.

In this review we summarise our recent efforts in trying to understand the role of heterogeneity in cancer progression by using neural networks to characterise different aspects of the mapping from a cancer cells genotype and environment to its phenotype. Our central premise is that cancer is an evolving system subject to mutation and selection, and the primary conduit for these processes to occur is the cancer cell whose behaviour is regulated on multiple biological scales. The selection pressure is mainly driven by the microenvironment that the tumour is growing in and this acts directly upon the cell phenotype. In turn, the phenotype is driven by the intracellular pathways that are regulated by the genotype. Integrating all of these processes is a massive undertaking and requires bridging many biological scales (i.e. genotype, pathway, phenotype and environment) that we will only scratch the surface of in this review. We will focus on models that use neural networks as a means of connecting these different biological scales, since they allow us to easily create heterogeneity for selection to act upon and importantly this heterogeneity can be implemented at different biological scales. More specifically, we consider three different neural networks that bridge different aspects of these scales and the dialogue with the micro-environment, (i) the impact of the micro-environment on evolutionary dynamics, (ii) the mapping from genotype to phenotype under drug-induced perturbations and (iii) pathway activity in both normal and cancer cells under different micro-environmental conditions.

Proteins change over the course of evolutionary time. New protein-coding genes and gene families emerge and diversify, ultimately affecting an organism\'s phenotype and interactions with its environment. Here we survey the range of structural protein change observed in plants and review the role these changes have had in the evolution of plant form and function. Verified examples tying evolutionary change in protein structure to phenotypic change remain scarce. We will review the existing examples, as well as draw from investigations into domestication, and quantitative trait locus (QTL) cloning studies searching for the molecular underpinnings of natural variation. The evolutionary significance of many cloned QTL has not been assessed, but all the examples identified so far have begun to reveal the extent of protein structural diversity tolerated in natural systems. This molecular (and phenotypic) diversity could come to represent part of natural selection\'s source material in the adaptive evolution of novel traits. Protein structure and function can change in many distinct ways, but the changes we identified in studies of natural diversity and protein evolution were predicted to fall primarily into one of six categories: altered active and binding sites; altered protein-protein interactions; altered domain content; altered activity as an activator or repressor; altered protein stability; and hypomorphic and hypermorphic alleles. There was also variability in the evolutionary scale at which particular changes were observed. Some changes were detected at both micro- and macroevolutionary timescales, while others were observed primarily at deep or shallow phylogenetic levels. This variation might be used to determine the trajectory of future investigations in structural molecular evolution.