Hair! An Evolutionary Story

What happened in genomes of mammals who lost their hair?

Hair is characteristic of mammals, and forms part of our advanced sensor suite that so effectively replaces the need for armor, shells, and the like. We can have very soft skin, and use hair for warmth, protection, sensing, display, nesting, and other things. But for some mammals like whales and dolphins, hair became a nuisance and has almost wholly been shed. Why it happened is pretty clear, both functionally and by the theory of natural selection. But figuring out how it happened requires delving into the respective genomes of these animals, including us- who, while hardly hairless, have much less than the common mammal. And the genomes tell an interesting story about how evolution works.

The premise of a recent study that tackled this issue (review) is that convergent evolution, that is the development of the same trait in distinct lineages, usually happens in similar ways, mutating the same genes that are key ingredients or regulators for that trait. So, given the many genomes that have now been sequenced across all the mammals, they asked whether a bunch of hairless mammals share some genetic mechanisms behind hairlessness. While there are several sudden appearances of hairlessness in dog breeds and mice, traceable to specific mutations/genes, the authors were not really interested in the genetics of hair itself, but how during actual evolution, hairlessness came about. 

An example of a large-scale map of syntenic regions that match between human and mouse genomes, by chromosome. Recombination and shuffling between regions and chromosomes happens frequently over long periods of time, in addition to smaller-scale mutations, making it challenging to do this mapping, which is a first step to analyses like those discussed here.

They are looking for signs such as significantly decreased or increased evolutionary rates in particular regions of a genome, which might mean either that that region has escaped selection, perhaps because the need for hair has disappeared, so selection is now indifferent to the maintenance of its genetic ingredients, or because the region is under positive selection, perhaps because the need for hairlessness is urgent, not just a matter of indifference. This method also assumes that corresponding regions of different mammalian genomes (called syntenic regions) can be identified, despite a great deal of rearrangement that will have happened over these long spans of time. 

A rough rundown of the anatomical location of expression of genes which had significant evolutionary speedups in hairless species. The Y axis is a measure of enrichment of that location of expression vs a null hypothesis of expression everywhere, for the set of 27 accelerated genes. Skin and hair expression are clearly favored by genes found in this analysis.

It is the convergence aspect, comparing several different lineages that evolved to similar states, that is hoped to weed out the hair-specific changes from all the other riffraff that happens over the ages. (While they also tried to cancel out the marine-heavy and large size-weighting of their selected species.) Each species has many other challenges, after all, and specific trajectories for each of its genes, and most mutations are meaningless. Selective comparison should help focus on the regions that really matter. The authors found 27 genes with accelerated changes, none of which had signs of positive selection. Half of these genes became defunct- they turned into pseudogenes, which is definitive evidence for relaxation of evolutionary selective constraints, rather than positive selection for new constraints. 

In the following, I give their ranked top genes for evolutionary rate acceleration, divided by those changing in their coding regions, and those changing most in nearby non-coding regions.

== Coding sequence acceleration ==

• FGF11   - fibroblast growth factor, with no previously known role, but is here the top statistical match. FGF5, FGF7, FGF18, and FGF22 are known to participate in hair growth in mice, so this is likely another regulator of hair growth.
• GLRA4 - is one of the pseudo genes, in humans is a glycine receptor and a pseudogene
• KRT2      - keratin
• KRT35    - keratin
• PKP1       - plakophilin, known to be important for skin formation, causes skin fragility when absent- related to intermediate filament system.
• PTPRM - protein tyrosine phosphatase, transmembrane receptor, cell adhesion and related signaling, in hair cells.
• ANXA11  - cell growth / survival regulatory protein causal for ALS disease and epithelial sarcoidosis.
• MYH4      - myosin heavy chain motor protein, known to be involved in hair growth. "... out of 69 KRTs and KRTAPs for which noncoding enrichment could be calculated, 66 showed accelerated evolution in both protein-coding sequence and noncoding regions"

== Noncoding sequence acceleration ==

• ELF3 - transcription regulator, known to be involved in hair development
• FOXC1 - transcription regulator, known to be involved in hair development
• CCDC162-SOHLH2 - readthrough into a transcription regulator
• FAM178B - unknown
• UVSSA - transcription regulator
• OLFM4 - extracellular matrix, promoting cell adhesion
• ADRA1D - hormone receptor and regulator of cell proliferation
• mir205 - translation regulator, known to be involved in hair development 
• .. and numerous other microRNAs

So a theme is emerging here, which is that some of the central players in hair development and hair structure received extra mutations in their coding sequences. Keratins, which make up hair, are an obvious case. The other altered genes have more obscure connections, but it is evident that FGF11 plays some important role in hair development, analogous to its relatives that are all local signaling molecules that instruct cell type and proliferation.

Genes found in this study, spread over a non-coding acceleration vs coding region acceleration in their mutation (evolution) rate. In orange are all the keratin genes, some of which have high rates in their coding sequences, but pretty much all of which have high rates in nearby non-coding DNA. In blue and red are other genes with nearby non-coding rate acceleration, with ones known to participate in hair functions marked in red.

On the other hand, regulators of other genes involved in hair development- a series of transcription regulators and micro-RNAs- were altered in their own regulation, which happens in non-coding areas of genes, but not in their coding sequences. This is because these regulators have many other roles, which would be disrupted by changing their coding sequences. Their special sauce lies in where and when they are expressed, which then leads to the complex combinatorial interactions they have with batteries of other regulators converging at their target genes. A great deal of evolution consists of twiddling with dials, rather than hammering on the machinery, or yanking it out.

So, over the tens of millions of years that mammals have been around, the loss of hair appears to come down to tweeks over many different genes, some of which are thrown out entirely (becoming pseudogenes), others of which are disabled in different ways in their protein sequences (keratins, FGF11), and others that are merely relocated in their expression, so that while most of their roles are untouched, their role in hair development is toned down or eliminated. The genome is an orchestra with a lot of players, whose contributions to the whole is sometimes loud and clear (keratin), but more often indirect and obscure, rewarding deeper forms of music appreciation.

• China is making high level pro-China propaganda.
• But we are as well.
• Conventional economics: wrong!
• Joe Biden, TCB. Or G2G?