Revisiting ancestral Y chromosome haplogroups DE-YAP+ and D lineages

Forearlier posts on this topic, see previous post: “Divergence of East Asians and Europeans Estimated Using Male- and Female-Specific Genetic Markers” by Yoshio TATENO et al. Genome Biol Evol (2014) 6 (3): 466-473. doi: 10.1093/gbe/evu027

A Rare Deep-Rooting D0 African Y-Chromosomal Haplogroup and Its Implications for the Expansion of Modern Humans Out of Africa

Marc Haber,Abigail L. Jones,Bruce A. Connell,Asan,Elena Arciero,Huanming Yang,Mark G. Thomas,Yali XueandChris Tyler-Smith


今天的人类在非洲以外下降主要是from a single expansion out ∼50,000–70,000 years ago, but many details of this expansion remain unclear, including the history of the male-specific Y chromosome at this time. Here, we reinvestigate a rare deep-rooting African Y-chromosomal lineage by sequencing the whole genomes of three Nigerian men described in 2003 as carrying haplogroup DE* Y chromosomes, and analyzing them in the context of a calibrated worldwide Y-chromosomal phylogeny. We confirm that these three chromosomes do represent a deep-rooting DE lineage, branching close to the DE bifurcation, but place them on the D branch as an outgroup to all other known D chromosomes, and designate the new lineage D0. We consider three models for the expansion of Y lineages out of Africa ∼50,000–100,000 years ago, incorporating migration back to Africa where necessary to explain present-day Y-lineage distributions. Considering both the Y-chromosomal phylogenetic structure incorporating the D0 lineage, and published evidence for modern humans outside Africa, the most favored model involves an origin of the DE lineage within Africa with D0 and E remaining there, and migration out of the three lineages (C, D, and FT) that now form the vast majority of non-African Y chromosomes. The exit took place 50,300–81,000 years ago (latest date for FT lineage expansion outside Africa – earliest date for the D/D0 lineage split inside Africa), and most likely 50,300–59,400 years ago (considering Neanderthal admixture). This work resolves a long-running debate about Y-chromosomal out-of-Africa/back-to-Africa migrations, and provides insights into the out-of-Africa expansion more generally.

HUMANS outside Africa derive most of their genetic ancestry from a single migration event 50,000–70,000 years ago, according to the current model supported by genetic data from genome-wide (Mallicket al.2016;Paganiet al.2016), mitochondrial DNA (mtDNA) (van Oven and Kayser 2009), and Y-chromosomal (Weiet al.2013;Hallastet al.2015;Karminet al.2015;Pozniket al.2016) analyses. The migrating population carried only a small subset of African genetic diversity, particularly strikingly for the nonrecombining mtDNA and Y chromosome where robust calibrated high-resolution phylogenies can be constructed, and in each case all non-African lineages descend from a single African lineage, L3 for mtDNA or CT-M168 for the Y chromosome. Yet there has been a long-running debate about the early spread of Y-chromosomal lineages because their current distributions do not fit a simple phylogeographical model. The CT-M168 branch diverged within a short time interval into three lineages (C-M130, DE-M145, and FT-M89), and just a few thousand years later the lineage DE-M145 further split into D-M174 and E-M96 (Pozniket al.2016), illustrated in Supplemental Material, Figure S1. Thus, around the time of the expansion out of Africa, between one (CT-M168) and four (C-M130, D-M174, E-M96, and F-M89) of the known extant non-African lineages were in existence (plus additional African lineages). The complexity arises because three of these four early lineages (C-M130, D-M174, and FT-M89) are exclusively non-African, apart from those entering Africa through recent gene flow; while the fourth lineage (E-M96) is largely African, where it constitutes the major lineage in most African populations. The debate began in the absence of reliable calibration, and these distributions were interpreted as arising in two contrasting ways: (1) an Asian origin of DE-M145 (also known as the YAP+ lineage), implying migration of CT-M168 out of Africa followed by divergence into the four lineages outside Africa and then migration of E-M96 back to Africa (Altheide and Hammer 1997;Hammeret al.1998;Braviet al.2000), or (2) an African origin of DE-M145, implying divergence of CT-M168 within Africa followed by migration of C-M130, D-M174, and FT-M89 out (Underhill and Roseman 2001;Underhillet al.2001).The first scenario requires two intercontinental lineage migrations, while the second requires three and is thus slightly less parsimonious.

An additional very rare haplogroup, DE*, carrying variants that define DE but none of those that define D or E individually, added to this complexity. First identified in 5 out of 1247 Nigerians within a worldwide study of >8000 men (Wealeet al.2003), DE* chromosomes were subsequently reported in a single man among 282 from Guinea-Bissau in West Africa (Rosaet al.2007) and in 2 out of 722 Tibetans within a study of 5783 East Asians (Shiet al.2008).While the phylogeographic significance of these rare lineages was immediately recognized, their interpretation was hindered by the incomplete resolution of the phylogenetic branching pattern and the possibility that they might originate from back-mutations at the small numbers of variants used to define the key D and E haplogroups, or genotyping errors rather than representing deeply divergent lineages, plus the lack of a robust timescale. Large-scale sequencing of Y chromosomes has now provided both the phylogenetic resolution and the timescale needed (Weiet al.2013;Hallastet al.2015;Karminet al.2015;Pozniket al.2016), so we have therefore reinvestigated the original Nigerian DE* chromosomes using whole-genome sequencing to clarify their phylogenetic position. We then consider the implications for the out-of-Africa/back-to-Africa debate related to Y-chromosomal lineages, and the expansion out of Africa more generally. …


Construction of a calibrated Y-chromosomal phylogeny

We constructed a series of phylogenetic trees based on all the Y-chromosomal sequences in our data set, or subsets of them. All showed a consistent structure, in which the Nigerian DE* chromosomes formed a clade branching from the DE lineage close to the divergence of D and E chromosomes (Figure 1A) in comparison with a set of Y chromosomes representing most of the world (Figure 1B).The Nigerian chromosomes had 489 derived SNPs exclusive to their branch in addition to a large deletion spanning ∼118,000 bp (Y:28,457,736–28,576,276). All DE-M154 chromosomes shared 29 SNPs. The Nigerian chromosomes shared seven SNPs with other D chromosomes, one SNP with E chromosomes, one SNP with C1b2a chromosomes, and one SNP with an F2 chromosome (Table S1). The reads overlapping these SNPs were visually investigated using the Integrative Genomics Viewer (IGV) version 2.4.10 and seen to support the calls. We consider sharing of a single SNP as a recurrent mutation in different lineages and interpret the Nigerian chromosomes as lying on the D lineage, diverging from other D chromosomes at 71,400 years ago (Figure 1C), very soon after its divergence from E at 73,200 years ago. We name the lineage formed by the Nigerian samples D0, to reflect its position on the tree and avoid the need to rename all the other D lineages.

Figure 1

Y Chromosome phylogenetic tree from worldwide samples. (A) A maximum-likelihood tree of 180 Y-chromosome sequences from worldwide populations. Different branch colors and symbols represent different haplogroups assigned based on ISOGG v11.01. The Nigerian chromosomes sequenced in this study are highlighted in blue and assigned to the novel D0 haplogroup. Bootstrap values from 1000 replications are shown on the branches. (B) Map showing location of the studied individuals with colored symbols reflecting the haplogroups assigned in A. The clade consisting of the D0 and D haplogroups is represented by blue squares and is observed in Africa and East Asia. (C) Ages of the nodes leading to haplogroup D0 in the phylogenetic tree (point estimates; branch lengths are not to scale). Haplogroups D0 and D are estimated to have split 71,400 (63,100–81,000) years ago while the D0 individuals in this study coalesced 2500 (2200–2800) years ago.

三个D0染色体不能区分one another, and have a coalescence time of ∼2500 years (Figure 1C), consistent with their collection from different villages, languages, ethnic backgrounds, and paternal birthplaces (Wealeet al.2003).The autosomal genomes of these individuals confirm their genetic ancestry as West Africans (Figure S2).

Models for the expansion of Y-chromosomal lineages out of Africa

The updated phylogeny including the D0 lineage adds two key pieces of information to the debate about the phylogeography of the Y lineages ∼50,000–100,000 years ago and the mode of expansion out of Africa. First, it increases the number of relevant lineages at this early time period from four to five, and second, it provides a reliable timescale for the branching times of these lineages, and thus for the lineages in existence at any particular time point.

In the phylogeny (Figure 1A), the DE lineage now contains three, rather than two, early sublineages: one exclusively African (D0), one mainly African (E), and one exclusively non-African (D). We therefore consider the implications of this revised structure for interpreting the present-day Y phylogeography as the result of male movements at different times between 28,000 and 100,000 years ago (Figure 2).要做到这一点,我们需要校准的发展史,and for this use the ancient-DNA-based mutation rate (Fuet al.2014), which has been widely adopted (e.g.,Pozniket al.2016); we consider in theDiscussionthe implications of alternative mutation rates and some of the other simplifying assumptions we make here.

Figure 2

模型的早期运动Y-chromosomal公顷plogroups out of Africa and back. (A) Simplified Y-chromosomal phylogeny showing the key lineages, including D0. Lineages currently located in Africa are colored yellow; those currently outside Africa are blue. Triangle widths are not meaningful, except that they show that E and FT are the predominant lineages inside and outside Africa, respectively. The small orange triangle in FT represents the R1b-V88 back-to-Africa migration that took place after the time period considered here. (B–D) Models for lineage movements that could lead to the present-day African or non-African distributions of lineages, using point estimates of dates derived from the phylogeny (see Table S2). The three models represent migrations out of Africa at different time intervals, indicated by the purple, brown, and green shading in A. Arrows in B–D indicate intercontinental movements and their direction, but do not represent particular locations or routes. The first colored arrow(s) represent the lineage(s) that migrated out, during the time period shown at the top of the maps. Additional uncolored arrows represent subsequent migrations and their time intervals needed to produce the present-day distributions.

We consider three scenarios based on our split-time point estimates of the Y-chromosomal lineages (Table S2). First, between 101,000 years ago (divergence of the B and CT lineages) and 77,000 years ago (divergence of the DE and CF lineages) only one lineage with present-day non-African descendants is present in the phylogeny (CT;Figure 2A), so present-day Y-lineage distributions could be explained by migration of the single lineage CT out of Africa, followed by back-migration of the D0 and E lineages between 71,000 years ago (origin of D0) and 59,000 years ago (divergence of E within Africa) (Figure 2B).This and all other scenarios require migration out of E-M35 after 47,000 years ago (its origin) and before 28,500 years ago (its divergence) to explain its presence outside Africa (Figure 2, B–D).其次,在76000年前(散度之间tween C and FT) and 73,000 years ago (divergence between D and E), three relevant lineages are present (the C, DE, and FT lineages,Figure 2A), so migration out of these three followed by back-migration of D0 and E as above (Figure 2C) would explain the distributions. Third, between 71,000 years ago (split of D and of D0) and 57,000 years ago (divergence within FT), five relevant lineages are present, and migration out of three of these (C, D, and FT) would explain the present-day distributions without requiring back-migration (Figure 2D).For simplicity, we do not include the short intervals between these three scenarios of 500 years and 1800 years (Figure 2Aand Table S2).


The new D0 data presented in this work are based on just three Y chromosomes, but have far-ranging implications for the structure of the Y-chromosomal phylogeny and hence male movements and migration out of Africa more generally. Our phylogenetic results are consistent with three scenarios (Figure 2, B–D), and we now consider some of the complexities associated with these, and how they fit with nongenetic data.

Complexities arise because although the phylogenetic structure, including the branching order, is very robust (Weiet al.2013;Hallastet al.2015;Karminet al.2015;Pozniket al.2016), its calibration depends entirely on the mutation rate used. The mutation rate chosen above, based on the number of mutations “missing” in a 45,000-year-old Siberian Y chromosome (Fuet al.2014), has been widely adopted (Pozniket al.2016;Balanovsky 2017),但冰岛谱系的大规模的研究encompassing the last few centuries suggested a rate ∼14% faster (Helgasonet al.2015).This faster mutation rate would translate directly into 14% more recent time estimates so that, for example, the Y-chromosome movements out of Africa in the three scenarios presented above would be 87,000–66,000, 65,000–63,000, and 61,000–49,000 years ago, respectively. These differences between mutation rates inferred in different ways should be seen within the context of a wider debate about human mutation rates, previously based largely on autosomal data (Scally and Durbin 2012).Each mutation rate is also accompanied by its own uncertainty, leading to the 95% confidence intervals in Table S2, which include the mutation rate uncertainty. We also assume that the mutation rate is constant over time and does not differ between lineages. The first assumption is very reasonable for the time period of most interest here, 50,000–60,000 years, when the mutation rate averaged over 45,000 years (Fuet al.2014) is used. A flexible mutation rate that assumed a real increase in recent times would have little influence on these estimates since the Fuet al.rate already includes the last few centuries. Differences in mutation rate between lineages need further investigation, but would not be sufficient to affect the scenarios presented inFigure 2. For these reasons, we believe that the Fuet al.rate, averaged over 45,000 years, is the appropriate one to use for the times of interest here.

These genetic times can be compared with dates from nongenetic sources for modern humans outside Africa. The 45,000-year-old Siberian fossil (Fuet al.2014) was reliably dated using carbon-14, while a ∼43,000-year-old fragment of human maxilla from the Kent’s Cavern site in the UK was dated using Bayesian modeling of stratigraphic, chronological, and archaeological data (Highamet al.2011).Archaeological deposits at Boodie Cave in Australia were dated to ∼50,000 years ago using optically stimulated luminescence (Vethet al.2017).Thus, there is strong support for the widespread presence of modern humans outside Africa 45,000–50,000 years ago. Earlier dates have also been reported, for example the Madjedbebe rock shelter in northern Australia dated by optically stimulated luminescence to at least 65,000 years ago (Clarksonet al.2017), a modern human cranium from Tam Pa Ling, Laos was dated by Uranium-Thorium to ∼63,000 years ago (Demeteret al.2012), and 80 teeth from Fuyan Cave in southern China dated using the same method to 80,000–120,000 years ago (Liuet al.2015), raising the possibility of a substantially earlier exit (Baeet al.2017).Such early archaeological dates also, however, raise the question of whether or not the humans associated with them contributed genetically to present-day populations (Mallicket al.2016;Paganiet al.2016).Archaeological data alone therefore do not provide an unequivocal date for the migration of the ancestors of present-day humans out of Africa.

All non-Africans carry ∼2% Neanderthal DNA in their genomes (Greenet al.2010), and Neanderthal fossils have only been reported outside Africa. The geographical distribution of Neanderthals thus suggests that mixing probably occurred outside Africa, and the ubiquitous presence of Neanderthal DNA in present-day non-Africans is most easily explained if the mixing took place once, soon after the migration out. This mixing has been dated with some precision using the length of the introgressed segments in the 45,000-year-old (43,210–46,880 years) Siberian male (Ust’-Ishim) to 232–430 generations before he lived,i.e., 49,900–59,400 years ago assuming a generation time of 29 years (Fuet al.2014).If this date represented the time of the migration out of Africa, it would exclude the first two scenarios (Figure 2, B and C).Thus, the combination of Y phylogenetic structure and dating of the out-of-Africa migration based on the 45,000-year-old Siberian fossil (Fuet al.2014) favors the third scenario (Figure 2D) involving the migration out of C, D, and FT between 50,300 years ago (lower bound of the FT diversification, Table S2) and 59,400 years ago (upper bound of the introgression; seeFigure 3), which is in accordance with suggested models incorporating an African origin of the DE lineages (Underhill and Roseman 2001;Underhillet al.2001).According to this interpretation, the reported Tibetan DE* chromosomes (Shiet al.2008) would most likely represent back-mutations or genotyping errors at the one SNP used to define haplogroup D, but require further investigation.

Estimation of the time of the out-of-Africa migration incorporating information from Y-chromosomal lineages (green, this work), archaeological dates (brown, Fuet al.2014), and ancient DNA (red, Fuet al.2014).

” data-icon-position=”” data-hide-link-title=”0″>Figure 3

Figure 3

Estimation of the time of the out-of-Africa migration incorporating information from Y-chromosomal lineages (green, this work), archaeological dates (brown,Fuet al.2014), and ancient DNA (red,Fuet al.2014).

mtDNA sequences also provide a robust phylogeny which demonstrates that non-African mtDNAs descend from a single African branch with rapid diversification outside Africa into the M and N lineages and many subsequent branches (Ingmanet al.2000;Devièseet al.2019).Dating using ancient mtDNA suggests a separation of non-African from African lineages after 62,000–95,000 years ago (Fuet al.2013), while an analysis of present-day mtDNAs suggested divergence outside Africa 57,000–65,000 years ago (Fernandeset al.2012).These estimates are based on <1% of the sequence length used from the Y chromosome but are nonetheless very consistent.

This discussion has thus far assumed that present-day distributions of Y haplogroups are relevant to events 50,000–100,000 years ago and thus that Y phylogeography carries information about the major migration out of Africa. Ancient population structure within Africa that separated C, D, and FT from other Y haplogroups beginning after 76,000 years ago with migration out only 50,000–59,000 years ago would also fit the evidence presented above. Present-day Y-chromosomal structure in Africa has been massively shaped by events in the last 10,000 years, including the Bantu-speaker expansion in central and southern Africa (Pozniket al.2016;Patinet al.2017) and entry of Eurasian lineages into northern and central Africa (Haberet al.2016;D’Atanasioet al.2018), and is thus a poor guide to structure before 10,000 years ago. Despite this, it is striking that western central Africa is the location of the deepest-rooting A00 lineage in Cameroon (Mendezet al.2013), a major location of the A0 lineage in Cameroon, The Gambia, and Ghana (Scozzariet al.2014;Pozniket al.2016) and the D0 lineage in Nigeria and Guinea-Bissau (Wealeet al.2003;Rosaet al.2007).This retention of the deepest Y-chromosomal diversity in western central Africa contrasts with the autosomal genetic structure, where the deepest roots have been reported in southern African hunter-gatherers (Gronauet al.2011;Schlebuschet al.2012,2017;Veeramahet al.2012;Mallicket al.2016;Skoglundet al.2017), perhaps supporting the hypothesis of deep population structure (Hennet al.2018;Scerriet al.2018).Analysis of ancient African DNA from 50,000 to 100,000 years ago would provide considerably more information on Y-haplogroup distributions at this time, but is not currently available. In the meantime, further focus on present-day Y-chromosomal lineages in central and western Africa to understand more about deep African lineages seems warranted, and this current study illustrates the broad insights that can sometimes be revealed by very rare lineages.

In conclusion, sequencing of the D0 Y chromosomes and placement of them on a calibrated Y-chromosomal phylogeny identify the most likely model of Y-chromosomal exit from Africa: an origin of the DE lineage inside Africa and expansion out of the C, D, and FT lineages. It suggests an exit time interval that overlaps with the time of Neanderthal admixture estimated from autosomal analyses, and slightly refines it. These findings are consistent with a shared history of Y chromosomes and autosomes, and illustrate how study of Y lineages may lead to general new insights.

This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

See alsomy review of recent literatureevolving views on the Central Asian origins of Tibetan, Sherpa, Nepali and Bhutan (i.e. Himalayan) haplogroup D / YAP+alu polymorphism … these are seen as having East Asian, East Eurasian or Northeast Asian origins.