Traditional comparative genomics are core tools for the study of evolution, as changes in the genome may allow one to determine causality with relation to specific phenotypic differences between species. However, one also needs to consider other forces affecting changes on the genome or phenotype level, such as environmental exposures and nutrition. Moreover, in the study of the brain, we need to be cognizant of cell- and circuit-specific factors, as well as emergent properties of human higher cognition and behavior, such as social and cultural buffers that could be acting in concert with changes on other levels (Varki et al., 2008
Animal model organisms play a crucial role in neuroscience. However, one unfortunate trend has been a de-emphasis on studies of human brain. Advances in genomics and genetics and availability of human brain tissue, coupled with non-invasive methods for the study of brain function provide enormous opportunities. The critical challenge is to embrace the evolutionary differences between humans and model organisms, and use them as a platform for discovery, rather than pretending that they don’t exist. This will permit integration of the large amount of data produced in model organisms with an understanding of human brain function, so as to understand the true relevance of these data to humans. Comparative functional genomic investigations will provide a molecular basis for elucidation of brain circuit evolution, and the evolution of human cognition and behavior.
The technical advance of microarrays allowed for unprecedented access to differential gene expression in brain evolution. However, due to the biases and static nature of the information able to be queried using microarrays, this technology is limiting for cross-species comparisons. For comparative genomic purposes, NGS is rapidly making expression array issues of historic interest, but NGS still requires a draft genome for alignment. Therefore, as NGS becomes less expensive and computational resources improve, even individual labs will be able to conduct de novo sequencing of their unique model organism. The focus in evolutionary genomics needs to extend work in traditional experimental organisms such as mouse, rat, drosophila etc., into a deeper appreciation of the myriad of organisms with a CNS, especially those that might have unique brain and behavior functions. For example, recent deep sequencing of small RNAs in marine worms identified ancient microRNAs that may have been essential for the evolution of the CNS (Christodoulou et al., 2010
As always, integration of multiple levels of genomic, regulatory, and RNA and protein expression, while challenging, is a must. Genome-wide transcription factor binding studies (ChIP-seq), sometimes combined with RNA-seq, are just beginning to be conducted (Kim et al., 2010
; Visel et al., 2009
). However, these and other studies are really only beginning to scratch the surface of what NGS can uncover, such as allele specific expression (Meaburn et al., 2010
; Schalkwyk et al., 2010
). Additionally, it will be useful to define molecular pathways based on functional genomic studies that identify gene sets expressed in either a tissue or cell-specific manner (Doyle et al., 2008
), and use these unbiased pathways as a basis for studying genetic pathway associations in normal cognition and disease. Such studies will aid in broadening our acceptance of the role of particular genes in specific tissues.
Phenotype discovery, in general, is a huge area of need within the field of evolutionary genomics. Thus, the use of imaging data analysis to determine phenotypes has been essential for this progress. As imaging is one of the few windows into the brain we can utilize in living patients, it is critical that we harness this technology in combination with genomic data. Perhaps the greatest challenge for the evolutionary neuroscientist is to establish how to combine data from functional genomics studies with behavioral data, again, in particular imaging data, and genome-wide association study information. Such integration relies on being able to determine quantitative behaviorally relevant phenotypes. Recent work in the identification of “phenologs,” or phenotypes that arise from disruption of orthologous genes, is one approach to ascertain such data (McGary et al., 2010
). Through this methodology, phenologs can be detected for diseases such as amyotrophic lateral sclerosis (ALS; human-yeast), autism (human-mouse), and intellectual disability (human-Arabidopsis). Combining this approach with other data from model systems will identify gene networks essential for normal human brain function that can be further explored in lower organisms.
The genomics revolution now permits us to perform key experiments that could rapidly advance our understanding of the evolution of higher cognition. But, one caveat to many of the evolution studies discussed above is the fundamental assumptions that are made in drawing the particular conclusion of each study. In particular, we need to be cautious about over-interpretation of any findings in the setting of uncertain models. Interspecies comparisons as far as human and mouse were important first steps, but some basic assumptions may not have been met and the positive selection of nervous system genes in humans, as initially suggested (Dorus et al., 2004
), is not yet well established. For example, this and other studies need to consider the role of relaxation of constraint, or reducing the restrictions of amino acid changes, rather than only adaptive evolution in the primate brain. One hypothesis is that relaxation of constraint presented the occasion for adaptive evolution to occur. Furthermore, other fundamental characteristics such as mutation rates, and huge differences in population history, as well as gene choice and choice of arbitrary relaxed thresholds for determining protein evolution (Dorus et al. 2004
) may hamper interpretation of these types of comparative studies. In fact, a number of studies have used other metrics in their analyses, and there seems to be little consensus in the field as to whether human brain genes have indeed undergone accelerated evolution or not (Bakewell et al., 2007
; Clark et al., 2003
; Nielsen et al., 2005
; Shi et al., 2006
). It is notable that studies that assess the genome in an unbiased manner, rather than picking only one arbitrary gene set, show no evidence for adaptive evolution in brain genes overall (e.g. (Nielsen et al., 2007
; Shi et al., 2006
Another example is the pathbreaking study of a humanized Foxp2 mouse (Enard et al., 2009
). In this case, are the assumptions about evolution made reasonable, and how can we know? What are the experiments that could even test this? We cannot obviously expect to generate a talking mouse and yet, there is no doubt that many experiments needed to more fully explore FoxP2 function can be done in mouse and other model organisms. The same issues are multiplied when we begin to study gene-gene interactions and molecular circuits. Rather than avoiding this complexity, it will be crucial to embrace the many levels of analyses and cross species comparisons that are necessary if we truly expect to advance to a new level of understanding of human brain evolution.