Khoe-San matrilineages and prehistory

A most interesting study has just been published that reconstructs the prehistory of the Khoe-San peoples of Southern Africa primarily using mitochondrial DNA analysis but with very important reliance on archaeological data as well.

Karina M. Schlabusch et al., MtDNA control region variation affirms diversity and deep sub-structure in populations from Southern Africa. BMC Evolutionary Biology 2013. Open access → LINK [doi:10.1186/1471-2148-13-56]

Abstract (provisional)

Background

The current San and Khoe populations are remnant groups of a much larger and widely dispersed population of hunter-gatherers and pastoralists, who had exclusive occupation of southern Africa before the influx of Bantu-speakers from 2 ka (ka = kilo annum [thousand years] old/ago) and sea-borne immigrants within the last 350 years. Here we use mitochondrial DNA (mtDNA) to examine the population structure of various San and Khoe groups, including seven different Khoe-San groups (Ju/'hoansi, !Xun, /Gui+//Gana, Khwe, =Khomani, Nama and Karretjie People), three different Coloured groups and seven other comparative groups. MtDNA hyper variable segments I and II (HVS I and HVS II) together with selected mtDNA coding region SNPs were used to assign 538 individuals to 18 haplogroups encompassing 245 unique haplotypes. Data were further analyzed to assess haplogroup histories and the genetic affinities of the various San, Khoe and Coloured populations. Where possible, we tentatively contextualize the genetic trends through time against key trends known from the archaeological record.

Results

The most striking observation from this study was the high frequencies of the oldest mtDNA haplogroups (L0d and L0k) that can be traced back in time to ~100 ka, found at high frequencies in Khoe-San and sampled Coloured groups. Furthermore, the L0d/k sub-haplogroups were differentially distributed in the different Khoe-San and Coloured groups and had different signals of expansion, which suggested different associated demographic histories. When populations were compared to each other, San groups from the northern parts of southern Africa (Ju speaking: !Xun, Ju/'hoansi and Khoe-speaking: /Gui+//Gana) grouped together and southern groups (historically Tuu speaking: =Khomani and Karretjie People and some Coloured groups) grouped together. The Khoe group (Nama) clustered with the southern Khoe-San and Coloured groups. The Khwe mtDNA profile was very different from other Khoe-San groups with high proportions of Bantu-speaking admixture but also unique distributions of other mtDNA lineages.

Conclusions

On the whole, the research reported here presented new insights into the multifaceted demographic history that shaped the existing genetic landscape of the Khoe-San and Coloured populations of southern Africa.

From the reading of the paper, I gather the following chronology (which should be always taken with some caution because of the uncertainty of "molecular clock" methods but in this case they seem reasonably backed from the material/cultural evidence record):

1. L0d coalescence time estimate may correlate with the arrival of MSA to the region c. 100 Ka ago (I estimated once ~90 Ka, so it is consistent with my thought).
2. Its sublineage L0d1'2 might have expanded c. 50 Ka ago (I would rather think of a more ancient chronology, soon after the L0d node - they can't correlate it properly with any obvious archaeological pattern, so it might be, I guess, more related to the apogee of MSA c. 75 Ka ago).
3. Some L0d1 subclades (notably L0d1a, L0d1b) would have expanded with the transition to LSA (40-20 Ka ago).
4. L0d2a shows an star-like expansion that they estimate to have happened c. 7-8 Ka ago and would be related to an Epipaleolithic (with microlithic industry) that is also notable for the increase of the density of archaeological findings in South Africa and Lesotho. This lineage also shows secondary expansion with pastoralism later on.
5. The introduction of herding c. 2000 years ago may have affected the correlations between the various L0d lineages. However most lineages show signs of expansion in this period. The main exception is L0d1a (decrease instead) and to some extent L0d1c (first decrease, later increase probably related to the !Xun late adoption of pastoralism, affecting especially to L0d1c1).
1. L0d3 is too old to have expanded with pastoralism, so the authors reject  Tatiana Karafet's hypothesis that it expanded in this period and that it could be related (to most unlikely) linguistic relation between Sandawe and Khoe-San. Instead they suggest (as I did in the past) that L0d3 had an East African distribution instead with only minor spreading to the Khoe-San in relation with pastoralism.
6. The recent Iron Age (last millennium) arrival of Bantu-speakers absorbed primarily L0d2a, which is the most common lineage of Khoe-San peoples (including Coloureds, with the partial exception of Cape Coloured, where it is second to L0d2b).

The paper only briefly mentions L0k1, which is most concentrated towards Katanga (D.R. Congo) and may therefore have arrived to Southern Africa only with Bantu or pastoralist flows.

Frequencies and estimated timelines of major Southern African L0d and L0k lineages (from fig. 4):

See also:

• PhyloTree: subtree L(xL3)
• At my old discontinued blog Leherensuge:
• The early expansion of H. sapiens in Africa (mtDNA)
• Reviewing the mtDNA L lineages (notes): L0
• At this blog:
• Khoisan autosomal genetics
• L0k lineages found in Zambia
• The Neolithic site of Leopard Cave (Namibia)
• African diversity and the possible origins of Humankind