Post by Admin on Feb 24, 2021 0:36:45 GMT
Cave bear nuclear and mitochondrial relationships are
highly incongruent
Phylogenetic relationships among cave bear taxa have been
largely guided by analysis of their mitochondrial DNA.15,16 Our
previous study on nuclear genomes did reveal one instance of
mito-nuclear discordance among three Late Pleistocene
taxa,10 but the limited sampling of this study precluded a broader
assessment of cave bear nuclear relationships. To investigate
this further, we analyzed the Middle Pleistocene praekudarensis
genome dataset alongside novel genome datasets of Late Pleistocene
cave bears generated from their petrous bones, representing
the taxa rossicus (Kizel Cave, Ural Mountains, Russia),
kanivetz (Medvezhiya Cave, Ural Mountains, Russia), and
kudarensis (Hovk 1 Cave, Armenia). We also included published
datasets from the taxa spelaeus (Eiros Cave, Spain),10 eremus
(Windischkopf Cave, Austia),10 ingressus (Gamssulzen Cave,
Austria),11 and a second kudarensis individual from Hovk 1
Cave10 in addition to modern Georgian and Late Pleistocene
Austrian brown bears,10 two modern polar bears,17 and a modern
Asiatic black bear18 as outgroup (see Tables S1 and S2).
We investigated relationships among the cave bear nuclear
genomes using Principal Components Analysis (PCA). This
involved sampling a single mapped nucleotide from each individual
at each position of the reference genome, which provided
data from a total of 487,747 variable transversion sites after strict
filtering (STAR methods). PCA suggested three major groups
(Figure 1A) comprising: the Caucasus cave bears praekudarensis
and kudarensis, which cluster together as predicted by
morphology; a second, geographically widespread and broadly
European group including spelaeus, ingressus, eremus, and
kanivetz from the Urals; and finally, the Urals cave bear rossicus,
which is distinct from all other cave bears and may represent a
Urals-specific group. PCA further suggests a hierarchy of relationships
within these groups, with each successive PC separating
different taxa from one another (Figure 1B).
Figure 1
Cave bear relationships
The three major groups identified by the PCA deviate from expectations
based on mitochondrial DNA, which instead supports
two major clades comprising the Caucasus cave bears and all
European and Urals cave bears, respectively, with rossicus
nested within the latter (Figure 1C). We further investigated these
contrasting nuclear relationships using phylogenetic analysis,
including representative brown bears, polar bears, and the
Asiatic black bear outgroup. Palaeogenomic datasets are
generally associated with high rates of error, which can distort
estimates of phylogenetic branch lengths when only a single
read is sampled.11 We therefore applied a recently developed
method, Consensify,11 which calls the majority base from a
random sample of three mapped nucleotides. This method provided
a single high-quality allele from each individual for a total of
4,318,414 genomic sites, of which 39,122 were variable.
Maximum likelihood phylogenetic analysis of this dataset (Figure
1D) supported the expected relationships between polar
bears, brown bears, and the cave bear clade, as well as the position
of the Caucasus cave bears within the latter. Among the
sampled cave bears from Europe and the Urals, however, there
is not a single sister-group relationship that agrees between the
mitochondrial and nuclear phylogenies.
highly incongruent
Phylogenetic relationships among cave bear taxa have been
largely guided by analysis of their mitochondrial DNA.15,16 Our
previous study on nuclear genomes did reveal one instance of
mito-nuclear discordance among three Late Pleistocene
taxa,10 but the limited sampling of this study precluded a broader
assessment of cave bear nuclear relationships. To investigate
this further, we analyzed the Middle Pleistocene praekudarensis
genome dataset alongside novel genome datasets of Late Pleistocene
cave bears generated from their petrous bones, representing
the taxa rossicus (Kizel Cave, Ural Mountains, Russia),
kanivetz (Medvezhiya Cave, Ural Mountains, Russia), and
kudarensis (Hovk 1 Cave, Armenia). We also included published
datasets from the taxa spelaeus (Eiros Cave, Spain),10 eremus
(Windischkopf Cave, Austia),10 ingressus (Gamssulzen Cave,
Austria),11 and a second kudarensis individual from Hovk 1
Cave10 in addition to modern Georgian and Late Pleistocene
Austrian brown bears,10 two modern polar bears,17 and a modern
Asiatic black bear18 as outgroup (see Tables S1 and S2).
We investigated relationships among the cave bear nuclear
genomes using Principal Components Analysis (PCA). This
involved sampling a single mapped nucleotide from each individual
at each position of the reference genome, which provided
data from a total of 487,747 variable transversion sites after strict
filtering (STAR methods). PCA suggested three major groups
(Figure 1A) comprising: the Caucasus cave bears praekudarensis
and kudarensis, which cluster together as predicted by
morphology; a second, geographically widespread and broadly
European group including spelaeus, ingressus, eremus, and
kanivetz from the Urals; and finally, the Urals cave bear rossicus,
which is distinct from all other cave bears and may represent a
Urals-specific group. PCA further suggests a hierarchy of relationships
within these groups, with each successive PC separating
different taxa from one another (Figure 1B).
Figure 1
Cave bear relationships
The three major groups identified by the PCA deviate from expectations
based on mitochondrial DNA, which instead supports
two major clades comprising the Caucasus cave bears and all
European and Urals cave bears, respectively, with rossicus
nested within the latter (Figure 1C). We further investigated these
contrasting nuclear relationships using phylogenetic analysis,
including representative brown bears, polar bears, and the
Asiatic black bear outgroup. Palaeogenomic datasets are
generally associated with high rates of error, which can distort
estimates of phylogenetic branch lengths when only a single
read is sampled.11 We therefore applied a recently developed
method, Consensify,11 which calls the majority base from a
random sample of three mapped nucleotides. This method provided
a single high-quality allele from each individual for a total of
4,318,414 genomic sites, of which 39,122 were variable.
Maximum likelihood phylogenetic analysis of this dataset (Figure
1D) supported the expected relationships between polar
bears, brown bears, and the cave bear clade, as well as the position
of the Caucasus cave bears within the latter. Among the
sampled cave bears from Europe and the Urals, however, there
is not a single sister-group relationship that agrees between the
mitochondrial and nuclear phylogenies.