Thank you! This was really illuminating. Makes sense that a cytoskeleton & transport would develop first, making symbiosis possible. Great research design to figure out something so cryptic.
They looked for duplication events (and dated them using computer models of genetic drift) — where the whole genome is duplicated — which set species up for rapid evolution because now one copy can change while one remains unchanged and functional.
Very useful in agricultural breeding as well. For example wheat has undergone one or more recent whole-genome duplications that have allowed it to diverge from its wild ancestors.
> Oh! I see. Wow, is full genome duplication very common?
Yes. And in more complex organisms, you also have the intermediate step of chromosomal duplication.
In humans, for example, besides the common XX and XY variants, there are also XXX, XXY, XXYY, and so on. Note that sex chromosome duplications in particular are not inheritable, but there are many others that are.
Meanwhile, chromosomal deletions and fusions are also possible. For instance, chimpanzees, gorillas, and orangutans have 48 chromosomes, while humans have 46 (usually, anyway). In the human branch after we split off from chimps two chromosomes fused to create chromosome 2 (this is a simplification, as there was a lot of swapping and patching that occurred before the fusion at various points in our common ancestry).
It's probably important to realize that the mechanisms of evolution itself have evolved to both allow and leverage this sort of accidental cut and paste modification without the organism simply failing outright. That isn't to say that such changes are likely to be beneficial, but they are often neutral in the immediate aftermath, and therefore eventually a possible source of generativity. This is very different from how source code works, and also very different from how genetic algorithms typically work.
