Saturday, 26 August 2017

Genetic proximity: what is it, and how much is needed?

This is a topic that has been barely covered here before, but is surely of interest for many of my readers especially since the 2015 article on the genetic studies involved in the Tauros Project was published by Rewilding Europe. The reason why I have covered neither the article nor the issue of genetic proximity here as such is that it is a lot to write and explain, and I haven’t had the time previously. However, this summer I have the time and inspiration to finally cover this topic appropriately.

By genetic proximity, we of course refer to the genetic proximity of living domestic cattle to its wild progenitor, the Eurasian aurochs Bos primigenius primigenius. But there are several ways in which domestic cattle can be “close to their ancestor”, several ways how genetic proximity is defined or measured, and what it implicates for “breeding-back”. However, as I always admit, I myself am a layman in the field of genetics so please point me to mistakes if I made some.

What was especially striking when Rewilding Europe published their article on genetics was the chart of Nei genetic distance of a number of cattle breeds to the aurochs. The most obvious result is that there is seemingly no clear correlation between a less-derived phenotype and “genetic proximity” (see down below), which led some people to conclude that “genetics don’t matter”. But genetics do matter of course: if you insert an aurochs’ genome into a Holstein zygote, you get an aurochs. But why this discrepancy? To solve this question and to find out what it means for “breeding-back”, we have to look at what “genetic proximity” means in this case.
Nei distance analysis published here
How genetic proximity is usually measured

Of course it would be best and most comprehensive to compare the full length of two genomes against each other. However, as genomes are huge molecules with a lot of information (in the case of cattle, the genome includes 3 billion base pairs and 22.000 genes), this would be very effortful. And apart from that, only a very small fraction of the whole genome codes for the defining differences between closely related species, and an even smaller portion codes for individual variation. So in order to make it easier, geneticists often rely on easily identifiable, short DNA sequences such as marker genes, and haplotypes of the mitochondrial genome or Y chromosomes. The advantage of molecular marker genes, such as cytochrome c and others, is that there is a constant average mutation rate per generation and that they are barely effected by selection as they have minor influence on the individual phenotypic variation, especially because many variations are neutral. This also makes them useful for determining the time of splitting up between evolutionary lines (“molecular clock”, the markers being used are called molecular chronometers). While variations on the mitochondrial genome, Y chromosome and their haplogroups as well as other marker genes are often used to determine the time of divergence or degree of genetic proximity under the assumption of a constant mutation rate and that they are barely affected by phenotypic selection, their influence on the actual phenotype and thus the nature of the animals is comparably minor. Mitochondrial genes mostly serve functions in the mitochondria themselves, and there are only a few genes on the Y chromosome that are actually relevant (lying on the sex-determining region of the chromosome). Haplotypes (variations passed on by only one parental lineage) are often used as an indicator of relatedness, but they actually are just an accumulation of more or less neutral variation on haplotypes and their influence on the organism as a whole is really minor. That is why the variation on those markers usually studied are barely affected by selection, what makes them in turn good markers. But their influence on the genetic architecture of the animals is meagre. They can indicate relatedness, but do not guarantee that the rest of the genome, especially those regions coding for the particular differences between the taxa compared, will be similar too.

As an example: let us assume that we take the genome of an ordinary Holstein-Frisian and exchange all sequences of its marker genes, haplotypes and even the complete mitochondrial genome with those sequences of an original aurochs and let that embryo develop. The result will still be a Holstein-Frisian, because the genetic material exchanged are mostly more or less neutral variations on genes that have a minor influence on the individual variation within a species or even between related species and not the regions that define the differences between a Holstein-Frisian and an aurochs. A Holstein cow with aurochs mitochondria would still be a Holstein cow in our perception, because the influence of mitochondrial DNA of the organism as a whole is comparably small. Furthermore, haplotypic variation does neither determine the identity of an individual or a species.

Most of the identity of an organism is defined by the nuclear genome. If you want to determine truly influential differences between aurochs and cattle, you would have to look there. Which is why the geneticists cited in the Rewilding Europe article have had a look at nuclear autosomal SNPs (Single nucleotid polymorphisms). SNPs, as long as they lie on coding regions, do have an influence on the phenotype. Many mutations causing cancer in humans, for example, are SNPs. For the study, 770.000 SNPs have been investigated and compared between one aurochs individual and 35 cattle breeds. They calculated the Nei genetic distance and presented the results (shown above).

The genetic difference between wild aurochs and domestic cattle

As far as my understanding goes, the last word is surely not spoken with that. It is of course state of the art to measure genetic distance using SNPS, haplotypes, marker genes et cetera, but we have again to look at the relevant genetic differences between aurochs and cattle. Just as the aurochs was neither defined via its mitochondria or haplogroups, it also was not defined by a couple of thousand SNPs. I imagine that the truly vital genetic differences between an aurochs and domestic cattle concern the following biological aspects:

- neurology
- endocrinology
- developmental regulation: timing et cetera
- sexual dimorphism
- metabolism
- morphologic aspects
- immunology

As outlined in a number of previous posts where you also find relevant literature (here and here), a lot of the differences between domestic and wild morphology are probably caused by pleiotropic effects and developmental cascades and therefore concern the upper five points; probably only a few novel morphological mutations appeared (alleles such as those causing scurred horns, new colour variants, extreme cases of dachshund-leggedness etc.).
I think those seven aspects are where we should look for the defining genetic differences between cattle and Eurasian aurochs that have a large biological impact since they are probably those regions that determine whether we have to deal with an aurochs or domestic cattle. Presenting the full genome sequencing of a British aurochs, Park et al. 2015 noted that “important questions remain unanswered, including […] which genomic regions were subject to selection processes during and after domestication. […] Finally, the functions of genes showing evidence for positive selection in B. taurus are enriched for neurobiology, growth, metabolism and immunobiology, suggesting that these biological processes have been important in the domestication of cattle1, what fully supports my view.

It is therefore my suspicion that we really have to identify the regions coding the neurobiological, developmental, endocrinologic, morphologic, metabolic and immunologic differences between aurochs and domestic cattle if we want to make genetic comparisons that truly matter on a wider biological basis concerning the nature of the animals and not coincidental variations that have been accumulated on regions that are barely effected by selection.
It is therefore not surprising that standard genetic methods that are usually used to determine simple “relatedness” that look at marker sequences, be it haplotypes, mitochondrial genes, SNPs or molecular chronometers that are not deeply effected by phenotypical selection do not show a clear correlation between a less derived and a strongly derived phenotype, although the latter clearly implicates a lot of strong directive selection and mutations.
There are multiple angles to look at a genome, and if we would look at those regions named above that are probably where we should look for matches with the aurochs, we would probably receive a different picture because those regions are directly and highly involved in the shape, form and function of the organism and directly and highly affected by phenotypic selection.

Furthermore, and to come back to the chart presented by Rewilding Europe, I am not sure if the Nei distance is the right tool to compare aurochs and cattle. The Nei distance was developed to look at the divergence of populations via mutation and genetic drift. In the case of cattle we do not have nice clean cladogenesis, but at first we have a dramatic bottleneck, then strong selective pressure during domestication, then very likely also local introgression on multiple regions by different regional variants of aurochs, and, not to forget, very unequal selective pressure on the different populations/lines/breeds of domestic cattle we see today.

Implications for “breeding-back”

So we would actually have to clearly determine those particular regions determining the defining differences between aurochs and cattle and get an overview over the allelic differences there. I do not think that it is a problem that we have only one full aurochs genome yet, because those traits defining the aurochs will, unsurprisingly, be universal among aurochs. However, I do not think that the results will be that enchanting or provide an additional directive for “breeding-back” projects, because of two reasons: 1) the fundamental differences between aurochs and domestic cattle concerning the seven factors above will probably be more or less universal among domestic cattle because all of them show the typical traits of domestication to a more or less clear extent. It would surprise me if we could still find the genetic make-up for all defining wild aurochs traits split up and distributed among the cattle of this world. This could be expected if domestication was an uncoordinated process where each of the factors was modified separately, which was certainly not the case as an organism functions as a whole and domestication was a coordinated, conscious process with a clear objective 2) certainly some breeds will be closer to the aurochs than others, but the question is how much and if the extent is relevant at all. For example, if the phenotypically most primitive cattle on this world show a 1,01% match to the aurochs on those seven crucial factors (just a symbolic number) while the most domesticated cattle show a 1,009% match, then the difference has to be considered negligible despite the primitive cattle having a number of alleles for a primitive morphology. I don’t think the difference would be that small, but I also do not expect any miraculous surprises. I think that all domestic cattle on this world are pretty similar in being domestic, with the derived examples of course carrying it to the extreme concerning a few morphological and other traits.

The claim that “no aurochs genes were lost, but just split up and distributed among domestic cattle” that has been floating around in the web for a while not only has no empirical support but is also to be considered unlikely for the reasons outlined above. The dramatic genetic effect of domestication – a narrow genetic bottleneck at first, strong selection on a large number of genes that have a dramatic influence on fundamental aspects of the organism – probably irreversibly and universally purged off a lot of defining wildtype alleles and therefore aurochs alleles from the domestic cattle gene pool, creating modern domestic cattle. The occasional local introgression of aurochs may have left traces in the modern domestic cattle pool (especially in the form of immunologic adaptions etc.) but certainly it did not alter their domestic nature. And even if all the defining aurochs genes were indeed still present in the modern domestic gene pool, and even if we had already identified and tracked them down in the populations, one should not make illusions over uniting them by conventional breeding in an anywhere near future because we are probably talking about at least hundreds of gene loci here. Body size alone is determined by over 50 loci (in humans)2, and you already see how longsome it is to achieve the right colour setting although we are dealing only with a couple of genes there. Therefore, truly genetically reconstructing the aurochs by selective breeding would be a centuries-long project, and many of the key genes are probably lost anyway. It would probably be way faster, cheaper and endlessly more effective to genetically reconstruct an aurochs via either cloning or CRISPR-Cas9.

Therefore, studies on genetic markers and SNPs are nice, but we are certainly not looking at the key genes that are relevant for true genetic identity of the animal that the aurochs originally was. Furthermore, there are unfortunately good reasons to assume that many of those defining key genes are lost due to the dramatic process of domestication that all of our modern day cattle went through. That is why you read that few on “genetic proximity” on this blog. We don’t know what truly matters yet, and it is likely that we also don’t have it anymore. Thus, I fear that I have to say that claiming “we are genetically breeding-back the aurochs” is the same kind of simplicity once practised by the Heck brothers, just carried onto the next level – even if you back it up with marker or SNP analyses. Please do not get me wrong, this is just the personal opinion of a biology student sitting in front of its laptop, therefore I am open for any critique.


Literature

1 Park et al. 2015: Genome sequencing of the extinct Eurasian wild aurochs, Bos primigenius, illuminates the phylogeography and evolution of cattle.

2 Visscher 2008: Sizing up human height variation.

1 comment:

  1. One could argue that even the marsupials had a sabertooth-model, and so genetic proximity is obsolete.
    On the other hand i think it would be odd if some rare breeds that seem to have some Auroch-related genes others don't have could become extinct while there are breeding-back programms. This doesn't mean that i think that anybody should get focused exclusivly on rare spanish breeds, but if no one does...someone should.
    If taken seriously then there are only two breeds of interest on this list : Pajuna and Sayaguesa. I think it would make sense if Axarquia gets analysed also, and maybe Corriente. And there are a few Longhorns that could be interesting.
    So why mix these breeds at all ? To get a self-sustainable herd with a low inbreeding-coeffivient.

    There could be other breeding-projects like keeping up genes from nearly extinct basal breeds, and maybe adding some african cattle to widen the genetic pool...maybe to get something that resembles about the first domesticated cattle.
    I think if just anything gets mixed up, then from a genetic point of view maybe everything could get lost or watered down.

    ReplyDelete