People like dogs. They're man's best friend, after all. There are tons of different breeds, many of which could be classified as unique species if we didn't know better. Our interest in dogs has led to lots of studies into dog breeding, figuring out which genes gave rise to the different morphologies, and determining the evolutionary history of dogs.
The history we're most familiar with is that of domesticated wolves. That is, thousands of years ago, human populations tamed wild wolves, and those animals eventually gave rise to the dogs we know today. I'll admit, I thought that's where the story began and ended. Well, it's not. Darren Naish provides an excellent review of an alternative explanation for the origins of the domesticated dog. The hypothesis Naish describes posits that domesticated dogs are remnants of a lineage that branched off from wolves prior to the domestication event. This hypothesis is supported by morphological, behavioral, and genetic evidence.
Molecular markers are commonly used for reconstructing evolutionary relationships -- on both long and short timescales. Ideally, one can draw upon many independent (ie, unlinked) markers when making inferences. That's why mitochondrial DNA (mtDNA) and Y chromosomes are flawed; they behave as a single locus because of the lack of (or minimal) recombination. A single instance of natural selection should leave a pattern across the entire molecule (see here for some background), swamping out the signature of recent demographic patterns. For these reasons, mtDNA and Y chromosomes are poor markers for reconstructing short term demographic patterns.
That brings us to this paper reporting some sequences generated from Y chromosomes sampled from multiple dog breeds. Here is how the authors justify their project:
"It has been shown for humans that individuals may have Y-chromosomal haplotypes that are identical based on a large number of microsatellite markers even though the haplotypes as defined by biallelic markers are different, a contradiction which is due to recurrent mutations of the microsatellites. This shows that Y-chromosomal biallelic markers are imperative for phylogenetic studies where they can serve as a backbone for the Y-chromosome phylogeny upon which the more detailed microsatellite variation can be imposed. Furthermore, for the dating of ancient population genetic events, analysis of DNA sequence variation offers a better tool than do microsatellites. There is thus a need for DNA sequence information for the dog Y chromosome."
I won't disagree that DNA sequence variation is better than microsatellites for population genetic analysis. Though I doubt that Y chromosome polymorphisms will be all that useful "for the dating of ancient population genetic events." Any signature of selection at one locus will be lost when the next selection event occurs on the Y. Also, any demographic signature will not survive due to subsequent selective events. The only evidence of population genetic events that remain will be that of the most recent event. Using neutral markers to identify demographic or selection events requires recombination between haplotypes. No recombination means you can't perform the same analyses you would on autosomes.
So, what did the authors find? Here is a summary of the polymorphism data:
"Among the 10 dogs, 9 haplotypes were found which were defined by a total of 14 polymorphic positions, 13 substitutions and 1 indel. As expected for haploid sequence, none of the individuals were heterozygous for any nucleotide position. The average number of differences between two individuals was 4.2, giving a π value of 2.91 x 10-4. A π value for the wolf Y chromosome of 0.4 x 10-4 has earlier been reported, seven times lower than the here reported value for dogs. The reason for this discrepancy is not clear. One explanation might be that the here analysed dog sequences probably mostly derive from non-genic regions while the wolf sequences were mostly intronic sequence, possibly giving different substitution rates for the two data sets. Another explanation might be a geographical bias for the wolf sample (30/36 animals were from Russia (unspecified region)/the Baltics/Scandinavia, which because of the great mobility of wolves may be quite closely related populations). It can be noted that for the human Y chromosome π =1.51 x 10-4 (2.3 Mb of the NRY sequenced), which is in best agreement with the value for the dogs in this study."
Keep in mind that estimates of π (the average pairwise differences between all of the sequences in the sample) aren't all that useful without a comparison with the number of polymorphic sites in the sample. That said, there aren't enough individuals in this sample and not enough polymorphic sites to make any real population genetics inferences. But that wasn't the purpose of the study -- they were trying to identify SNPs that could be used in subsequent research. So why throw in these hand waving explanations for patterns of Y chromosome polymorphism?
Our take home messages are twofold: the evolution of domesticated dogs may not have occurred as most people believe, and we shouldn't study that evolution using only Y chromosomes.
- Log in to post comments