search this blog

Thursday, October 26, 2017

Ancient Guanches genetically most similar to modern-day Berbers (Rodríguez-Varela et al. 2017)

Over at Current Biology at this LINK. Emphasis is mine:

Summary: The origins and genetic affinity of the aboriginal inhabitants of the Canary Islands, commonly known as Guanches, are poorly understood. Though radiocarbon dates on archaeological remains such as charcoal, seeds, and domestic animal bones suggest that people have inhabited the islands since the 5th century BCE [1, 2, 3], it remains unclear how many times, and by whom, the islands were first settled [4, 5]. Previously published ancient DNA analyses of uniparental genetic markers have shown that the Guanches carried common North African Y chromosome markers (E-M81, E-M78, and J-M267) and mitochondrial lineages such as U6b, in addition to common Eurasian haplogroups [6, 7, 8]. These results are in agreement with some linguistic, archaeological, and anthropological data indicating an origin from a North African Berber-like population [1, 4, 9]. However, to date there are no published Guanche autosomal genomes to help elucidate and directly test this hypothesis. To resolve this, we generated the first genome-wide sequence data and mitochondrial genomes from eleven archaeological Guanche individuals originating from Gran Canaria and Tenerife. Five of the individuals (directly radiocarbon dated to a time transect spanning the 7th–11th centuries CE) yielded sufficient autosomal genome coverage (0.21× to 3.93×) for population genomic analysis. Our results show that the Guanches were genetically similar over time and that they display the greatest genetic affinity to extant Northwest Africans, strongly supporting the hypothesis of a Berber-like origin. We also estimate that the Guanches have contributed 16%–31% autosomal ancestry to modern Canary Islanders, here represented by two individuals from Gran Canaria.

Rodríguez-Varela et al., Genomic Analyses of Pre-European Conquest Human Remains from the Canary Islands Reveal Close Affinity to Modern North Africans, Current Biology (2017),


ghostnorris said...

They seem to be about as SSA admixed as modern Tunisians and Mozabite.

bellbeakerblogger said...

The individual from Gran Canaria with elevated HG ancestry, and previous individuals from El Hierro and La Gomera shows a possibility that the islands were colonized twice, or at least in waves with the bulk of Berberization being concentrated in Tenerife and the more distant islands retaining a more European like character, and for this individual more WHG ancestry.

They hint around by noting the differences in unique mtdna across the islands.

It would be interesting to see a direct comparison between Gran Canarians and pre-Beaker Kehf el Baroud. If the Gomerans and Hierrans come available, them as well.

bellbeakerblogger said...

Kehf el Baroud was in Fregel et al, 2017

capra internetensis said...


Yes, I wonder how KEB (and IAM) would come out. No mt haplogroup sharing with them, but T2c1d-152C shared with El Toro and Sardinians, J1c3 Neolithic Sardinian, and both at Els Trocs. J1c3f supposedly at Afalou, but that was dubious HVS-I only stuff.

I wish they had done some actual analysis comparing them to different Sub-Saharan Africans, all of these North African papers are leaving that hanging.

Have you heard something about Gomero or just wishing?

bellbeakerblogger said...

I was referring to modern distribution in Gomera. Here are several paper on the other islands.

A paper Co-authored by Bernard Secher, Fregel and others (2014)

Also La Palma (2009)

El Hierro (2016)

capra internetensis said...


Doh completely mixing up my place names.

bellbeakerblogger said...

This is an interesting paper I missed by Hernandez et al, 2017:
"The distribution of mitochondrial DNA haplogroup H in southern Iberia indicates ancient human genetic exchanges along the western edge of the Mediterranean"

They breakdown those H lineages in Southern Iberia and Morocco from the most recent DNA. Whatever it was, it seems to have been spreading North and South by the Middle Neolithic. They also suggest an origin in Central or Eastern Europe, which really squares both Hervella and Brotherton I think.

Ahmed nazari said...

And this is why i thank the lord baal hammon everyday that we have the technology and numbers to resist the byzantines and russians. Perphaps if the guanches had been in contact with their berber kindred. They would have survived.

André de Vasconcelos said...

Doesn't this seem to hint that SSA was present in north Africa before the Arab slave trade?

idurar said...

^ yes, but the focus is always on the African/SSA for some reason, despite the fact other things too make Berbers distinct in the modern West Eurasian variation (less ANE-like for instance). Anyway, while these samples aren't very old, they strongly suggest remarkable continuity in the region. That was already obvious with autosomal data from modern Northwest Africans which does not show - outliers excluded - a lot of diversity. It's actually a pretty monolithic population for a such a huge territory.

Labayu said...

@André de Vasconcelos

Not seeing that as all. In fact the opposite:

"However, outgroup f3 statistics [19] suggest that the Guanches share more genetic drift with non-African test populations than with African test populations, including Northwest African populations of Berber origin (Data S1, sheet 2). This observation is inconsistent with the PCA and the uniparental genetic marker data, indicating that the outgroup f3 statistic may be misleading, possibly due to the complex history of recent sub-Saharan admixture events in North African populations [12, 21]"

The reason being Guanches don't have sub-Saharan ancestry. Also see Figure 3.

Matt said...

@Labayu: The point in that section is that Both outgroup f3 statistics and average pairwise differences identify Sardinians as the population sharing the most drift with both modern North Africans and Guanches (Table S2). This is in stark contrast to our expectation that North Africans would cluster with other North Africans..

However, it replicates recent findings [21] showing that both ancient and modern Egyptians share more genetic drift with European populations than they do with other North African groups, while still showing evidence of strong genetic affinity to one another in other types of analyses. This behavior might be due to varying degrees of admixture from highly divergent sources (e.g., sub-Saharan populations) into different populations, which can affect the interpretation of f3 statistics [22].

Therefore, we conclude that these statistics are not suitable to identify the modern population most similar to Guanches in the sense that we intend it for this study

One inference - for Guanches to share more with Eurasians via f3 statistic does not imply they lack SSA ancestry... unless this also implies the same for present day North Africans.

See - Table S2:

@Andre: Taking Davidski's Global10, there's definitely a signal there that NW Africa has a specific signal of sharing on the axis 10 which splits Gambia / Sierra Leone from East African Bantu population.

Re: this paper, I don't have a lot to say about it, but I think it's good to see them looking at the issues of fst and f2 measures being discordant against f3 and direct measures of pairwise differences (see the supplement). I think there's a strong need to understand these issues

(This paper is a big review, but it's a relatively complicated spatial / algebraic subject - ).

Labayu said...


Thanks, although I didn’t understand that alone as an indication that Guanches do not have sub-Saharan ancestry. I caught that the explanation for the differences between the measures may in fact be that the various North African populations and Guanches are closest to each other (for the most part) but have minority ancestries from highly divergent sources. But judging by the various measures and the ADMIXTURE plot, it seem Guanches don’t have exactly the same admixture as modern North Africans. Specifically, the Guanches are missing the Yoruba affinity in the ADMIXTURE plot and aren’t close by FST to people from the Shaigi tribe like the modern North Africans are. The Shaigi tribe are Arabs of Sudan of mixed Arabian and Nubian ancestry.

I assume the question could be addressed much better than the data here speaks to, but that’s what I meant by seeing some evidence of the opposite.

Matt said...

Labayu: I assume the question could be addressed much better than the data here speaks to, but that’s what I meant by seeing some evidence of the opposite.

Yeah, this is true and it's not totally as clarified as it could be.

But judging by the various measures and the ADMIXTURE plot, it seem Guanches don’t have exactly the same admixture as modern North Africans.

What I do see there is it looks like there's a cline between the North African populations, with Egyptians on one end and Burbur and Mozabites on the other. The Guanches seem roughly intermediate the Burbur on this cline, though they do seem to have a bit more Europe EN/MN related ancestry than would be strictly on line...

However, the Guanches do take large amounts of the Mozabite component, which will contains SSA, and in the K2-K4 (in the supplement) when there is a simple blue African component and a pink Eurasian component, there is a roughly similar amount to of blue to what we see in the Egyptian samples, same kind of "ball park", though, yeah, possibly very slightly less than present day NW Africans. Their Shaigi sample also seems to have about the same amount of blue as Guanches. Though K2 can be misleading in some situations, this looks like it roughly would match what more formal tests could typically find for people in that region...

Lukasz M said...

There is some plink or geno dataset, or only fastq?