Friday, July 18, 2014

On the mythology of natural selection: Part IV. Functional selection

Not all evolutionary change has to be based on overpopulation and resulting vicious competition for scarce resources in a cruel and bloody Nature.  Yesterday, we suggested that organismal selection, a kind of inverted natural selection, is another means by which adaptive complex traits could arise.  But there are others.

Functional selection
Multicellular organisms develop from single cells. A cell is a very complex structure with all sorts of components.  Genomes code for hundreds to thousands and more different forms of functional RNA and protein.  These must interact for the organism to produce the structures it has evolved to need.  These ideas apply to single-cell species and to 'higher' organisms.

Cellular organelles.  From Weiss and Buchanan, 2009, The Mermaid's Tale

In multicellular organisms, each individual arises from a single initial cell (say, a fertilized egg) that then divides into specialized cells, forming different tissues and organ systems.  These largely are formed by processes of embryogenesis, before the organism is 'presented' to the environment as a free-standing individual--a new baby, say.  This is true of plants and animals.

The cells in an embryo are not in any serious sense competing for scarce resources and they don't overpopulate, say, relative to their their resource supply in the Darwinian concept of evolution.  But if their molecules fail to interact in ways needed for their particular species, the embryo will fail to develop into a viable adult.  Or, some favorable variants might lead to a greater chance of successful hatching.  These phenomena can be called functional selection.

While one can cite some rather post hoc ways to force this to seem like Darwinian selection (for example, in utero competition among pups in a litter), functional selection within given conceptuses can remove harmful variants in a way that doesn't really involve any sort of competition among individuals. The result will appear in DNA sequence comparisons as greatly reduced variation in important regions of the genome--such as is routinely found in protein-coding genes, and referred to as purifying selection.

Frog development.  From Weiss and Buchanan, The Mermaid's Tale, 2009

The usual explanation for sequence conservation is Darwinian competition or, rather, a kind of inverted Darwinism.  Darwin was trying to explain the positive evolution of new complex traits over time, leading to the world's diversity of differently adapted plants and animals.  The faster fox got the rabbit, the slower ones died.  Genetic variants that led to faster feet rose in frequency, and are what we see today in fox genomes.  Darwin's idea was that the environmental mix is always changing, always and relentlessly forcing this sort of adaptation due to competition for limited resources.

But what we mainly see in sequence comparisons is conservation, with known functional regions of the genome varying less than areas with less, unknown, or no function.  This is so widely observed that it has, perhaps somewhat incorrectly been taken as one of, if not the criterion for asserting function to some area of the genome.

So even here, even to the extent that Darwinian ideas of relentless competition are right, the effect is an inversion of the relentless race to be different.  Instead, it's a relentless guardian of the status quo!  And this raises a serious problem.  Conservation is so widespread in genomes, that one wonders how competitive purifying selection could actually work, because so many individuals would have inherited mutations, mostly harmful, that hardly any could survive the conservative screening of selection.  Too much genetic load, as it's called.

Functional selection perhaps provides a way out of this problem.  Organisms typically produce scads more gametes (pollen, sperm, eggs) than they need or will ever actually use.  Genetic variants expressed in the development of these cells will be purged easily and with low cost because the organism will sill produce enough viable gametes.  Similarly, most species produce vastly more zygotes (e.g., fertilized eggs) than ever need hatch as developed embryos.  Most genes are used during gamete or embryo formation, so that very low- or no-cost screening by functional selection can generate the observed kinds of patterns of sequence conservation, without the need for very expensive and extensive natural selection by competition among adults for limited food or mates.

Genetic variants that lead to, say, healthier embryonic development or gamete formation may tend to have a better chance at appearing as an adult and hence being transferred to later generations.  These would be relevant to some sorts of adaptive change (though not to those that do, in fact, require inter-individual competition). So functional selection could also contribute to positive adaptive evolution and new traits.  Again, this would have to be very gradual, just as Darwin said and everybody basically agrees is the case with complex traits.

This is not just an off-the-wall idea.  Decades ago, based on comparative phenotype data, CH Waddington made somewhat similar observations about the apparent high conservation of developmental pathways and how that constrained (he called in 'canalization') development.  He was a quirky nonconformist and was rather derided for it, and his idea was based on standard natural selection; but with modern DNA data and developmental genetic technologies, we see high conservation in developmental mechanisms (many examples are now  well known, perhaps exemplified by the Hox genes and body segment development in animals).  If such genes are not properly expressed in properly combinations and so on during development, the embryo may not form in a viable way even to be born (or, if in the germline, a gamete not to be formed).

And as with organismal selection, while functional selection is not Darwinian natural selection, it isn't mystical or strange, and doesn't provide any sort of challenge to the idea that complex traits and their genomic basis do evolve by normal, natural historical processes.  It's just that evolution is more nuanced, and less about ruthless competition, than is so widely and, we think, uncritically assumed.


Manoj Samanta said...

I presume this set of posts is leading to Nei's paper and book.

"Decades ago, based on comparative phenotype data, CH Waddington made somewhat similar observations about the apparent high conservation of developmental pathways and how that constrained (he called in 'canalization') development. He was a quirky nonconformist and was rather derided for it, and his idea was based on standard natural selection."

In 2006, Eric Davidson and Erwin published a Science paper on similar ideas, but evolutionary Coyne attacked them saying they were bringing back the discarded theories of macromutation. If you are familiar with the exchange, it is worth checking.

Ken Weiss said...

Reply to Manoj
No, we aren't explicitly leading to Nei's ideas, which basically stress the importance of mutation in setting the stage--but that traits are then still molded by selection. I don't recall Davidson & Erwin's particular point except again I think it was about the source of variation that later won the selection battle.

Nei is, of course, not a strong selectionist even regarding his view about the importance of mutation. For decades he's been one who has seen the weakness in dogmatic selection arguments.

Our point is not about the nature or structure of ultimate new variation, but about the reasons for subsequent differential proliferation. Or have I missed your point?