8 Inferring Human Demographic History from Genetic Data
195
coalescent times generally lead to greater divergence between sequences. Li and
Durbin (2011) use a hidden Markov model to estimate the coalescent times between
two haploid sequences sequentially across the genome (the times vary across the
genome because of recombination). The distribution of times is then used to estimate
the trajectory of past population sizes in the history of the population containing the
sample. While currently limited to analyses of a single genome, the PSMC provides
a novel way of utilizing genome-wide data to make inferences about human history,
and the results so far are consistent with previous findings of a population bottleneck
in non-Africans and recent growth in all populations (Li and Durbin 2011). More
recent work has used the sequentially Markovian assumption in other theoretical
frameworks, allowing for the analyses of larger sample sizes and population splits
(Sheehan et al. 2013; Schiffels and Durbin 2014; Terhorst et al. 2017; Steinrücken
et al. 2019).
8.7
Ancient Admixture
After modern humans evolved in Africa 150–200 Kya, they quickly expanded to
colonize the rest of the inhabitable world. As they did, they encountered other
hominin groups that already occupied the rest of Africa and Eurasia. These other
groups, often called “archaic” humans, included Neanderthals, Denisovans and
Homo erectus in Eurasia, H. floresiensis in island Southeast Asia, and several
unnamed groups within sub-Saharan Africa (Klein 2000; Trinkaus 2005; Rightmire
2009). The extent to which the expanding modern humans interacted and interbred
with the various archaic human groups is still unclear, though some interbreeding
must have occurred (see below). We first discuss how ancient DNA from archaic
hominins has changed our perspective on this issue. Then we describe other indirect
methods for inferring the existence of ancient admixture.
Direct Evidence for Ancient Admixture The isolation and sequencing of a portion
of the mtDNA hypervariable region from the Neanderthal-type specimen opened
up a new avenue of research for human evolutionary studies (Krings et al. 1997).
Subsequent work has generated whole mtDNA sequences from several Neanderthals
(Briggs et al. 2009) and a putative H. heidelbergensis individual (Meyer et al.
2014), low-coverage draft genomes from a Neanderthal (Green et al. 2010) and
a Denisovan (Reich et al. 2010), and high-coverage genomes from a Denisovan
(Meyer et al. 2012) and two Neanderthals (Prüfer et al. 2014, 2017). Denisovans
were a group of archaic humans whose only remains have been found in a single
cave in Southern Siberia. They are known almost exclusively from their DNA, with
very little morphological information available from the limited fossil remains that
have been found (Bennett et al. 2019; Viola et al. 2019). Analyses of the Denisovan
genome have shown that they are distant cousins of Neanderthals (Reich et al. 2010).
Studies of Neanderthal mtDNA found that the Neanderthal sequence was
outside of the range of normal human variation, suggesting that any Neanderthal
contribution to the modern human gene pool was limited (Krings et al. 1997; Serre
195
coalescent times generally lead to greater divergence between sequences. Li and
Durbin (2011) use a hidden Markov model to estimate the coalescent times between
two haploid sequences sequentially across the genome (the times vary across the
genome because of recombination). The distribution of times is then used to estimate
the trajectory of past population sizes in the history of the population containing the
sample. While currently limited to analyses of a single genome, the PSMC provides
a novel way of utilizing genome-wide data to make inferences about human history,
and the results so far are consistent with previous findings of a population bottleneck
in non-Africans and recent growth in all populations (Li and Durbin 2011). More
recent work has used the sequentially Markovian assumption in other theoretical
frameworks, allowing for the analyses of larger sample sizes and population splits
(Sheehan et al. 2013; Schiffels and Durbin 2014; Terhorst et al. 2017; Steinrücken
et al. 2019).
8.7
Ancient Admixture
After modern humans evolved in Africa 150–200 Kya, they quickly expanded to
colonize the rest of the inhabitable world. As they did, they encountered other
hominin groups that already occupied the rest of Africa and Eurasia. These other
groups, often called “archaic” humans, included Neanderthals, Denisovans and
Homo erectus in Eurasia, H. floresiensis in island Southeast Asia, and several
unnamed groups within sub-Saharan Africa (Klein 2000; Trinkaus 2005; Rightmire
2009). The extent to which the expanding modern humans interacted and interbred
with the various archaic human groups is still unclear, though some interbreeding
must have occurred (see below). We first discuss how ancient DNA from archaic
hominins has changed our perspective on this issue. Then we describe other indirect
methods for inferring the existence of ancient admixture.
Direct Evidence for Ancient Admixture The isolation and sequencing of a portion
of the mtDNA hypervariable region from the Neanderthal-type specimen opened
up a new avenue of research for human evolutionary studies (Krings et al. 1997).
Subsequent work has generated whole mtDNA sequences from several Neanderthals
(Briggs et al. 2009) and a putative H. heidelbergensis individual (Meyer et al.
2014), low-coverage draft genomes from a Neanderthal (Green et al. 2010) and
a Denisovan (Reich et al. 2010), and high-coverage genomes from a Denisovan
(Meyer et al. 2012) and two Neanderthals (Prüfer et al. 2014, 2017). Denisovans
were a group of archaic humans whose only remains have been found in a single
cave in Southern Siberia. They are known almost exclusively from their DNA, with
very little morphological information available from the limited fossil remains that
have been found (Bennett et al. 2019; Viola et al. 2019). Analyses of the Denisovan
genome have shown that they are distant cousins of Neanderthals (Reich et al. 2010).
Studies of Neanderthal mtDNA found that the Neanderthal sequence was
outside of the range of normal human variation, suggesting that any Neanderthal
contribution to the modern human gene pool was limited (Krings et al. 1997; Serre