The genome as ecosystem: Clarifying a novel approach to the study of “jumping genes”

Most of us are accustomed to thinking about the genome as a blueprint or program: its primary function is to control the development of an organism. However, only a small portion of the DNA in most species (4% in humans) is involved in coding for proteins affecting organismal phenotypes. Instead, most plant and animal genomes are largely composed of a class of genetic elements called transposons or “jumping genes.” Such transposable elements (TEs) are capable of autonomously self-replicating and reinserting into new chromosomal locations, and are thus subject to intra-genomic selection for differential replication. TEs can be harmful to the organism when they disrupt a gene; but they can also advantageously modulate gene expression when landing near a regulatory region. Although scientists have known about TEs since the 1950s, their evolutionary significance and precise relationship to the organism remains in dispute.

The prevailing way of thinking about TEs regards them as genomic parasites. However, in recent years this model has come under scrutiny. Some critics argue that TEs are simply too abundant in eukaryotic genomes to qualify as genuine parasites. It is also argued that the parasite/host model fails to explain the vast differences in TE abundance among species, i.e. the “C-value paradox.”

In this paper, I explore the recent proposal that an ecological approach to the genome is superior to the parasite model. In a nutshell, genome-level ecology views TEs as organism-like entities inhabiting an intra genomic ecosystem. Families of TEs are thus regarded as akin to ecological communities, while non-mobile features of the genome are regarded as their local environment. Models and concepts borrowed from the discipline of community ecology are then used to explain patterns of TE abundance and distribution.

This proposal sounds intriguing, in part, because it raises fundamental questions about what it means to take an ecological perspective towards some entity, and how exactly such an approach differs from a co-evolutionary (host/parasite) framework.

I attempt to answer these questions by presenting a general distinction between “strictly ecological” and “strictly evolutionary” modes of explanation. These are best understood as alternative idealization strategies, or, as alternative sets of simplifying assumptions about the system under investigation. On this view, a strictly ecological mode of explanation ignores variation among members of the focal population, treating them instead as an idealized type with fixed functional properties. Changes in the abundance and distribution of those entities are then explained in terms of their functional relationships to specific features of the environment. By contrast, a strictly evolutionary mode of explanation makes the opposite set of idealizing assumptions. The focal entity is a variable population. Influences of particular ecological factors are ignored (for simplicity). Instead, changes in population structure are explained in terms of “internal” properties such as effective population size or selection coefficients. Interestingly, on this reading a “complete” natural selection explanation (in Robert Brandon’s sense) qualifies as a sort of hybrid, where one attempts to explain how the two types of factor interact over successive generations. Such a combined eco-evo explanation is inherently more complicated than either of the more idealized approaches. For this reason, a strictly ecological or a strictly evolutionary mode of explanation can sometimes emerge as the preferred alternative, even when we know that, in reality, both evolutionary and ecological processes are operating –at least to some extent– on the system in question.

The payoff of this perspective is that it allows us to precisely measure the extent to which a given pattern calls for a particular type of explanation. This is achieved by simply partitioning the variance among the two types of factor. For example, to the extent that differences in TE abundance and distribution co-vary with structural properties of different genomes, an ecological explanation is called for. By contrast, to the extent that those differences co-vary with different lineages of “host” organism, a co-evolutionary explanation is called for. Alternatively, if the variance in TE abundance and distribution is correlated with both types of factor, then a combined eco-evo explanation is warranted.

I conclude with a proof of concept, showing how this framework has been successfully applied to communities of genetic elements in both closely related and more distantly related species. Our central finding is that differences among TE communities are explainable entirely by ecological factors among the closely related (drosophila) genomes, whereas similar differences are entirely explained by evolutionary factors among more distantly related (animal) genomes. In our sample, there was no need for a more onerous eco-evo explanation.