Blind men and a brown bear

An old Indian tale tells of a group of blind men who create a mental image of an elephant by touching it. Because each of them feels a different part of the body, they come to disagree. I was reminded of this story while studying another large mammal, the brown bear.
Blind men and a brown bear
Like

Aiming to uncover the phylogeographic history of the brown bear, we had sequenced the genomes of nearly 100 individuals. Combined with publicly available data from other individuals, this resulted in a dataset of 128 brown bears spanning the entire species range. However, when we mapped the data against genomic regions with different inheritance properties – i.e., the autosomes, the X chromosome, the Y chromosome, and the mitochondrial genome – each of the four datasets revealed different population structures.  

The autosomal and X-chromosomal data agreed, apart from some minor inconsistencies, that brown bear population structure is generally consistent with the geographical location and connectivity of extant populations. This contrasted sharply with the distribution of mtDNA haplotypes, which in brown bears is famously characterised by abrupt discontinuities between neighbouring populations and by haplotype sharing between populations separated in time and space. The geographic distribution of Y-chromosomal haplotypes proved even less predictable, with individuals from within the same population often having less in common with each other than with individuals thousands of kilometers away. What could have caused all these discrepancies?

One of the fundamental differences between the four datasets was that the mtDNA and Y-chromosomal data sets contain information for a single unlinked locus only, while the autosomal and X-chromosomal datasets contain information for thousands of unlinked loci. Because single loci are known to be susceptible to stochastic processes (i.e., random genetic drift of ancestral variation), it initially seemed reasonable to assume that the observed discordant signals of the mitogenome and Y chromosome were simply noise resulting from chance effects.

However, a more careful comparison of our research findings eventually led us to propose an alternative explanation. The new hypothesis we came up with was: partial or complete non-overlapping temporal resolution. Just as the blind men felt different parts of the body and therefore perceived seemingly contrasting descriptions of an elephant, the four genomic regions zoomed in on different time intervals of the brown bear’s phylogeographic history. Specifically, we hypothesised that the autosomal dataset reflects contemporary population structure, whereas the non-recombining mtDNA and Y-chromosomal datasets, and to a lesser extent the X-chromosomal dataset, would be better at retaining signals of a past population structure.

Consider, for example, the brown bear populations of Kamchatka and southwest Alaska, on opposite sides of the Bering Strait. Until approximately 11,000 years ago, these populations were connected by the Beringian land bridge. This shared ancestry is still evident from sex chromosomal and mtDNA data sets, all of which identify Kamchatka bears and southwest Alaska bears as a monophyletic clade. Autosomal data, in contrast, suggest that Kamchatka bears cluster with Russian Far East populations, and all Alaska bears with North American populations – consistent with current population connectivity. 

If, in our analogy, the elephant represents the phylogeographic history of brown bears, the autosomal data set provides a detailed and accurate description of the front of the elephant but no information about the back. This information about brown bear population structure in the distant past would be missing if it were not for the three other blind men, who examine and describe these remaining body parts: the mitogenome, the X chromosome and the Y chromosome. In the tale, the blind men can reconstruct the complete picture of the elephant if they work together and combine their independent findings. Similarly, a comprehensive overview of the demographic history of the brown bear is achieved by integrating the inferences from the four different datasets.  

 

A window to the past

The reason that mtDNA and the sex chromosomes, unlike autosomes, are able to retain signals of past population structure is that they are less influenced by gene flow. This is particularly true for mtDNA and the Y chromosome, both of which are non-recombining. When gene flow events introduce allochthonous mtDNA or Y-chromosomal haplotypes into a population, these haplotypes will start to circulate in the gene pool as non-dissolving agents. The allochthonous haplotypes do not exchange information with the autochthonous haplotypes, as autosomes and the X-chromosomes do. This means that gene flow cannot diminish the genetic distance between these haplotypes. It can only change their geographical distribution. Thus, while their non-recombining nature makes these loci susceptible to the confounding affects of random genetic drift, on the other hand it allows them to preserve the signatures of past population structure, and therewith to provide a window to the past.

For all the similarities between the mitogenome and the Y-chromosome – i.e., single-locus, uniparental inheritance, haploid, non-recombining – there is one important difference: one is inherited from the mother, the other from the father. When gene flow is sex-biased, as is the case for brown bears, this leads to different distribution patterns.

The discordance between brown bear mtDNA and Y-chromosomal phylogeographic patterns indicate that gene flow in brown bears is primarily male-mediated. Long distance dispersal by males has resulted in overlapping ranges of neighbouring Y-chromosomal clades, evidenced by paraphyletic clustering. Female philopatry, on the other hand, has led to non-overlapping ranges of mtDNA haplotypes. These ranges meet at sharp boundaries, which often coincide with a geographical barrier such as a lowland or a body of water.

While non-recombining loci can reveal population structure in the distant past, the flip side is that they are less suitable for detecting recent events. This is especially true for the mitogenome, due to its relatively short length. Assuming a total length of 16000 base pairs and a mutation rate of 2∙10-8 mutations per base per year, any pair of mtDNA haplotypes whose most recent common ancestor lived around  10.000 years ago, are expected to have independently accumulated on average a meagre three mutations each, and therefore to differ by six sites only. As a result, recent phylogeographic breaks may remain obscure from mtDNA data analyses.

Brown bears in the Eurasian taiga illustrate this type of mitonuclear discordance. The multi-locus datasets revealed a distinct genomic discontinuity in central Russia, which divides a western and eastern clade. Brown bear phylogeography has been interrogated by numerous mtDNA-studies, but none reported this conspicuous phylogeographic break before.               

Falling behind

Different considerations apply to the X chromosome. Despite being recombining, this region can retain signals of past population structure due to the mode of inheritance of sex chromosomes. Immigrant males pass on their X chromosomes only to female offspring, which effectively reduces the migration rate. Therefore, male-mediated gene flow affects X-chromosomal loci less than autosomal loci. As a consequence, X-chromosomal population structure falls behind autosomal population structure, and resembles past population structure longer.

This hypothetical ‘X-chromosomal lagging effect’ could explain why a binary tree constructed from the X-chromosomal dataset still identifies Kamchatkan bears and southwestern Alaskan brown bears as a monophyletic clade, whereas the autosomal data does not. If Kamchatkan bears and southwestern Alaskan bears are descended from a common ancestral Beringian population, male immigrants from other populations could have driven a wedge between the sister populations following the flooding of the Bering Strait, but less so in terms of X-chromosomal loci.

Additional information

The two competing explanations for the observed inconsistencies between our datasets, and the approach we used to determine for each separate case which hypothesis likely holds true, is described in more detail on: https://github.com/mennodejong1986/PopMSC 

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in

Subscribe to the Topic

Ecology
Life Sciences > Biological Sciences > Ecology

Related Collections

With collections, you can get published faster and increase your visibility.

RNA biology and RNA-based applications to medicine

This cross-journal Collection welcomes articles that report on RNA-driven cellular mechanisms, physiologically relevant RNA processes, and the development and characterisation of novel RNA-based medical applications for disease prevention, diagnostic and therapeutic purposes.

Publishing Model: Open Access

Deadline: Dec 27, 2023

Biology of reproduction

For this Collection, we encourage submissions that push forward our understanding of reproduction and its impact on offspring in both model organisms and human studies.

Publishing Model: Open Access

Deadline: Apr 10, 2024