People ask me why I work on frogs, snails, and strawberries. Do I just like French cuisine? No. I’m interested in, you guessed it, adaptive diversity. And in that regard, there are common themes across these seemingly unrelated projects. Here’s one of the main ones.
Geneticists study associations between traits and DNA. Traits are anything you can measure on an organism: size, shape, number of elbows, etc. The DNA could be a gene that affects the trait, or a proxy sequence close to such a gene. For any given trait, most of the genome has no detectable effect on it. If a geneticist is to find a connection, a gene has to do two things. First, it must produce a chemical involved in the trait’s development, such as the pigment that determines your eye color. Second, it has to vary among individuals in a meaningful way. Even if a gene is essential for your survival, and thus all of your traits, it has no practical consequence if everyone has an equivalent copy. And while most genes come in several discernible versions, these are usually pretty interchangeable. Because of this, the bulk of DNA has little if any effect on naturally occurring variation in any important trait. At least not the big obvious ones: Disease susceptibility. Fecundity. Endurance. Camouflage. Sex.
The genetic variants that do have a major effect are therefore special. They influence both long-term evolutionary trends and the day-to-day business of ecological communities. Geneticists like myself want to find them. Inspect them. Understand them. Where do we look? You might think they would all be clustered together in a centralized location. You know, on the command center chromosome. The genomic HQ. But in fact, the genomes of plants and animals are organized in a very haphazard way. Genes with similar roles are nowhere near each other. You can’t get a coherent picture by reading a chromosome start to finish, any more than you can get a good bedtime story by reading the dictionary cover to cover.
As a result, classical genetics worked on a gene-by-gene basis. In the traditional view, it didn’t matter what genes were near your gene of interest. They probably had no more effect on your gene than any other random gene in the genome. That was certainly my mentality in grad school. I was investigating genes in leopard frogs that produced bacteria-killing substances called antimicrobial peptides. They fascinated me in part because they were so variable. Different frogs produced different peptide combinations. I wanted to know how natural selection maintained such diversity, and how it affected frogs’ abilities to resist disease. But, I didn’t know what chromosomes they were on, what other genes were near them in the genome, or even if they were all close to each other. My committee gave me a PhD anyway. It was only reasonable: figuring out all that would have been an enormous challenge even just ten years ago. Back then, “genomics” was a word reserved for a few model species like mice and fruit flies, not leopard frogs.
But now, when any lab can build a genome-scale dataset for their favorite species, it’s looking more and more like the classical view was too extreme. The fates of genes are tied up with their genomic neighbors. For example, if nature selects a variant at one gene, it can “sweep” through a population, bringing along whatever variants at nearby genes happen to be linked to it. The key factor is recombination, the shuffling of genes. Sufficiently adjacent genes rarely recombine, and thus effectively behave as a single unit. Genes that are very far apart get reshuffled so often that they can’t influence each other. It turns out that in many species, a lot of genes are snuggled up close enough to matter. So what does that mean for which genes are likely to harbor functional variation? They may not be so randomly distributed after all. Some places in the genome might offer a better chance to be an ecologically relevant gene.
To explore this possibility, I ran some simulations. Imagine there are two genes. Each gene comes in two different styles (A or B). For each gene, individuals with two different copies (AB) have a 0.5% higher chance of survival than individuals with two identical copies (AA or BB). Now, the two genes have no interaction with each other. No epistasis, a biologist would say. That is, if you have two A copies at Gene 1, you are no better or worse off whether you have two A copies at Gene 2, or two B copies at Gene 2. But, crucially, if you are AB at both genes, then you can combine the two 0.5% benefits, and you are doubly advantaged.
In a perfect world, both A and B copies at both genes would stick around forever, since that conveys the highest fitness. But, random fluctuations called genetic drift can sometimes remove even beneficial variants from a population. My simulations tested which force would win: natural selection or genetic drift. I simulated evolution over 100,000 generations in populations of 1000 individuals. If I don’t allow the two genes to recombine, then both variants are retained at both genes 82% of the time. If there is infinite recombination, both variants are retained in only 0.1% of the simulations, and just 6% of simulations have even a single gene retain both A and B copies.
In other words, even if the genes have no biochemical relationship, they can give each other a “signal boost” to prevent genetic drift from overwhelming them. A gene is more likely to keep its adaptive variation if it’s closely linked to another gene with similar diversity. It’s a big, scary fitness landscape out there. It helps to have a buddy.
The evidence for this in the real world is merely suggestive at this point, but it opens the doors for fruitful future science. In human populations, this paper from 2009 identified 60 genes where natural selection appears to maintain variation affecting fitness. Remarkably, there were four small genomic sections, together comprising less than 0.04% of the genome, that each contained two or three of these genes. Do these 11 genes, 18% of the targets, use the buddy system? There are several potential causes for this clustering, not necessarily the mutual balancing act described above. Still, it’s an intriguing trend.
In my own research, I’m examining other examples. One species I study is a snail that transmits parasites to humans. The two genes with the largest measured effect on parasite resistance in snails are both in the same section of the same snail chromosome. The genes respond to different parasites, so we know they have two independent roles. That is, we’re not just picking up a signal from the same gene in two different ways.
Here’s another case. Some strawberry species have separate male and female sexes. We have found that the DNA sequence which determines sex in these plants has rapidly jumped around the genome. “Rapidly” in evolutionary time, at least – in less than a million years. What’s the reason for the jumps? Sex genes might have a better chance of survival when linked to the right other genes. There is no obvious benefit to having distinct sexes, and plants might occasionally revert to hermaphroditism. But maybe if the sex gene gets itself into a good neighborhood, la différence is more likely to continue. In fact, sex chromosomes are famous for accruing extreme inequalities. Our own X and Y chromosomes differ more than 2-fold in size, for example. It’s generally thought that these differences have to do with a gene being beneficial in one sex and harmful in another. But maybe in some cases, it just has to do with being linked to any old balanced dichotomy, sexual or not. Diversity fosters diversity.
Different classes of highly variable genes, like immunity genes versus sex genes, are typically studied independently by different research communities. I want to find general rules that pertain to all of them. One set of rules might govern how genomic location affects the propensity to harbor variation. As more labs delve into genomics of myriad species, we are likely to see more examples of this. I bet many biologists already are. I look forward to uncovering and sharing these patterns. There’s so much to explore.