Les Gaulois (2025)

Although of less importance in our research on the making of the Haitian people, we do find the flawed but attempted breakdown of France and western Europe intriguing in Ancestry's 2025 Fall Update. Our own "French" ancestry, estimated at 2% in the previous update, was broken apart into two separate categories: Brittany, France and Acadia (French Canada). We also saw Quebec in our updated results, but since it is attributed to our Hispanic parent's side, we find it more likely to be a result of the update's horrible results for people with roots in the Iberian Peninsula. 

In other words, our "France" results are basically the same. Acadia is basically "France" but Ancestry, to its credit, did not assign any French Canada regions to our Haitian parent. Our ""Northern Wales and North West England" also hails from our Haitian parent, probably a misreading of the minor French ancestry in Haitians. 

Intriguingly, for our Haitian parent, their British region was Southern Wales. If, however, 23andme is accurate, which certainly seems to be the case compared to Ancestry, our Haitian parent's European ancestry leans more heavily toward France and Belgium, the southern Dutch, and Rhinelanders. Since there is undoubtedly genetic overlap between peoples in France and Belgium as well as France and Germany, our guess is our Haitian side's European ancestry is probably derived from people from northern France who had some genetic overlap with these neighboring regions.

From our own "genealogical research" into Saint Domingue's free people of color in Bainet, we the French ancestry was mainly from the northern regions of France. This seems supported by 23andme assigning only small amounts of Spanish and Basque ancestry to our Haitian parent. Ancestry, on the other hand, assigns a mostly Spain ancestry for our Haitian parent's European admixture. Undoubtedly, 23andme is more reliable here, particularly for distinguishing between France and Spain. However, the "Brittany, France" region might be significant in that both of us inherited this region, perhaps pointing to northwestern France as the region from which the European ancestors of Bainet's free people of color often hailed from. We assume our Haitian parent's 0.2% Irish ancestry according to 23andme as well as their Southern Wales estimate via Ancestry is pinpointing to northwestern France, too. 

Ancestry 2025 Update (Africa)

Although we knew Africa was not the focus of the 2025 update for Ancestry, we still thought there might be some changes. Instead, with the exception of a decrease of 1% for Senegal, our sub-Saharan African estimates are essentially the same. For our Haitian parent, nothing changed either except for a minor increase in their Benin & Togo estimate (up by 2%). They retained their minor estimates for various smaller regions while we still have a division of "Nigeria" into Central Nigeria, North-Central Nigeria and Nigeria. 

None of this is too surprising. It looks like the 2025 update is largely unchanged from the 2024 update. And as we already expected, we are more West African than Central African In fact, it would appear that our Haitian parent and I harbor substantial ancestry derived from the Slave Coast and Bight of Biafra, which is also backed by the records of the Trans-Atlantic Slave Trade and Saint Domingue (colonial Haiti). I personally find it intriguing that so little sub-Saharan African ancestry was passed on through the Hispanic side.

The only real change to our African ancestry was the inflated North Africa estimate of 4%. The first time we got our results, the estimate was 2%. Then, following the update, it dropped to 1%. Inexplicably, it shot up to 4% this year. While we did see an estimate of 1% Canary Islands, we doubt we have substantial :Guanche" ancestry that might explain this. It seems like the creation of new areas for Spain and the Iberian peninsula altered or created problems for other nearby regions or categories. Indeed, this might be why we suddenly see a 2% Sephardic Jewish estimate while previous updates never included that. Of course, we do have very distant Sephardic Jewish matches from the eastern Mediterranean or North Africa as well as North African Muslim matches...

23andme Update

Our Haitian relative shared their updated ancestry composition report from 23andme recently. As we expected, little changed for their sub-Saharan African ancestry. The "Broadly West African" category disappeared, but merely boosted their "Nigerian" estimate to 43.2%. They also retained the close Igbo match. As for their other African regions, everything is largely the same. Country matches with Liberia, Sierra Leone ad Ghana, but no specific ethnic groups or regions. Their Congolese and Southern East African results are also similar, with the 0.4% Southern East African estimate minor. Again, hardly anything new. In fact, it confirms their overwhelmingly West African ancestry that appears most rooted in Lower Guinea. From the slave trade data to Saint-Domingue, this suggests almost half of her ancestry is rooted in areas that fed the Slave Coast and Bight of Biafra. 

For Europe, however, some changes can be seen. France is finally severed from Germany, and clearly makes up most of her European ancestry. This is hardly surprising for Haitians. The Belgian, Rhinelander, and Southern Dutch is new, but also not too surprising as Belgium neighbors France. We also noticed that our parent and I had distant DNA matches with users from Belgium and the Netherlands on MyHeritage, too. Perhaps 23andme is confirming distant ancestry from that region for both of us. As for the Iberian Peninsula, 23andme's update assigned it to Andalusian, Asturian & Castilian with 0.2% Basque. Very small but there is some overlap with France in the Iberian Peninsula. Haiti, of course, also had people of Spanish origin during the colonial era. Lastly, the Irish ancestry at 0.2% was unexpected. But perhaps to be explained in the overlap of Irish populations with other groups in western Europe?

Fortunately, there is some more precision in the update for our Haitian parent's minor Amerindian ancestry. Like Ancestry DNA, 23andme assigns it to a vaguely South American origin. We would like to think this Central Andean/Amazonian estimate is misreading indigenous Taino ancestry, but our parent does not have any Historical Matches from the indigenous Caribbean. Therefore, if it's not from the indigenous peoples of Haiti or the Greater Antilles, it may be indigenous people who came to Saint Domingue under French rule. Baptismal records indicate Indians from Aruba, today's Colombia, and other regions were in the colony. Perhaps that's the origin of this allegedly South American indigenous ancestry? 

As for trace ancestry, it is now reflected as North African and Northern Indian & Pakistani. The latter appears at all confidence levels, too, while North Africa disappears at 80% confidence level. The North African may be via the minor Spanish ancestry or perhaps West Africa. It is difficult to say with certainty, but North African ancestry is not indicated by this relative's Ancestry DNA results at all. Even for myself, with far more substantial ancestry from the Iberian Peninsula, my North African estimate was only 1% on Ancestry DNA. Thus, is it possible the trace North African ancestry in our Haitian parent is via West Africa? Their South Asian trace ancestry is similarly ambiguous. It is no longer Malayali South Indian but remains an enigma.

African Ancestry in Bainet and 23andme

A very close Haitian relative born in Bainet recently received their 23andme DNA results. In some ways, it's a confirmation of our likely Igbo ancestry. Indeed, the Igbo people were the only specific African ethnic group detected as a very close match in our relative's results. This was not a surprise since our closest African DNA match on Ancestry DNA was with an Igbo Nigerian. Furthermore, the "Ibo" were a consistent and major part of the enslaved African population in Bainet and other areas of Haiti during the colonial period. 

An attempt to show how our relative's African ancestry is broken down in Ancestry's 2024 Update.

Thus, the genetic evidence as well as the historical records (notarial records mentioning slaves by nation, the patterns of the French slave trade, and the smuggling of captives from Jamaica to Haiti's southern coast) point to ancestry among Igbo and/or closely related peoples in southeastern Nigeria. However, just as on Ancestry DNA, our relative also had Yoruba Nigerian DNA matches with various customers. We likely descend from a plethora of individuals from various ethnolinguistic backgrounds across what is Benin, Togo, and Nigeria.

Sadly, the lack of a Benin/Togo category on 23andme leads to a more inflated Ghanaian score for our close relative. The 2024 Update for Ancestry DNA pointed to very strong "Benin & Togo" as well as "Nigeria" scores. Indeed, Ancestry DNA's problematic update even suggested or pointed to ancestry in the East-Central and northern Nigeria for our relative. But 23andme, however, only detected a close match with the Igbo. 23andme's algorithm also gave our relative a higher Senegambian/Upper Guinea score, which we suspect may be due to possible "Bambara" or "Mandingue" ancestry that may be registering as Sierra Leone and Liberia. Lastly, the overall 23andme scores suggest ancestry in West Central Africa is a rather small part of our African origins. And this is in spite of the ubiquitous presence of captives from West Central Africa in Saint Domingue. I guess people with roots in the South really do, on average, harbor less Central African ancestry.

An attempt at showing the regional breakdown of sub-Saharan African ancestry. The general patterns are similar to those observed via Ancestry's results.

One of the benefits of 23andme for understanding African ancestry is seeing assigned mtDNA haplogroups. Our relative's haplogroup, L1b1a, is common in sub-Saharan Africa. This is no surprise and seems consistent with the African ancestry of Haitians. The maternal ancestry of our people, after all, is undoubtedly an African affair. One only wishes 23andme could have indicated some possible distant matches for West and Central Africa to shed more light on our African ancestry. Perhaps a better idea of the distribution of L1b1a would have also helped here. Ultimately, we find some broad commonalities between Ancestry DNA and 23andme here, and both companies estimated our close relative at about 89-90% sub-Saharan African ancestry, overwhelmingly West African. 

Nos Ancetres les Gaulois?

Although we are mainly interested in the African and Amerindian components in the making of Caribbean peoples, we probably should reflect on European ancestry, too. After all, it undeniably shaped the Caribbean and the making of who we are. However, in our case, the vast majority of our European ancestry is from the Iberian peninsula, Spain and Portugal. This is no surprise for the Spanish Caribbean.. 

Where things get funky and problematic, however, is France. The update assigned us 2% France. The previous estimates did not include France at all, but rather a variety of Northwestern Europe and Sardinia, mainly inherited through our Haitian parent. We suspect Sardinia, which did not appear in her ethnicity estimate at all, may be misread France, just as Scotland, Wales, and England & Northwestern Europe were probably France. The 2024 update seems to have tried to be more accurate with delineating France from other parts of western Europe, with partial success. However, given what we know of our Haitian parent's ancestry, her minor European ancestry is mainly from France. Her original estimates recognized that to a certain degree by assigning her 5% France and then a mix of Northwestern and southwestern European locales (Portugal, Scotland, Wales, England). Strangely, in her 2024 update, our mother's European ancestry was mostly reclassified as Spain, and supposedly 2% of our Spain estimate derived from her. Now, she has only 1% France and I have 2%, which is obviously unlikely since the inheritance feature indicates our France estimate comes from her. How can I have 2% France from her 1%? We suspect the 2% Spain inherited from her and the 1% Sardinia score is also probably from France. 

Needless to say, the update was clunky and inaccurate in some ways, still struggling to distinguish France from other parts of Northwestern and Mediterranean Europe. Spain, in the case of my mother, was vastly exaggerated. Nonetheless, the update at least recognized France in the new estimates for myself. Can we still repeat our bad joke? Nos ancêtres les Gaulois can still be claimed? Bad jokes aside, we are still mainly interested in the African and indigenous background to Caribbean peoples and will continue to focus on such themes. 

Caribbean Ancient DNA...

Although hardly the most sophisticated or reliable method, using Gedmatch's autosomal DNA comparison feature can be entertaining and occasionally useful. In this case, comparing our data to that of around 42 kits of ancient samples was an interesting experience. Indeed, of the 42 ancient samples, we only had matches with 4 individuals, which is significant. All 4 of them, however, were either individuals from Cuba or the Bahamas, surprisingly. We apparently did not share DNA, at least not detectable using Gedmatch, with ancient persons from Puerto Rico and Hispaniola with kits on Gedmatch (that we know of). The dates given for the samples are from undated in one case but the others the woman from Preacher's Cave, in the Bahamas, was the subject of an important article published in 2018 by Hannes Schroeder. According to this study, "Results are consistent with Puerto Ricans and the ancient Taino forming a clade without any significant gene-flow postdivergence." Clearly, the ancient Lucayan woman's genome was similar to that of modern-day Puerto Ricans. Indeed, this study found that Puerto Ricans share large parts of their genome with the ancient "Taino" of the Caribbean. Another study found similarities between the Preacher's Cave sample and precontact Puerto Rico samples in terms of autosomal DNA. These similarities also extend to South American Amazonian groups, such as the Yukpa, Piapoco, Karitiana, and Surui. Furthermore, modern Puerto Ricans harbor mtDNA that is at least partly a reflection of precontact Puerto Rico populations. 

What we find interesting in our matches is the lack of a Puerto Rico or Hispaniola connection. While this is probably, in part, due to the less sophisticated tools and models via Gedmatch, we did match with another ancient "Taino" sample from the Bahamas. On Gedmatch, it is called Kit WE2232796 (*Bahamas Taino Native), with further indication that its test results are from 23andMe. Is this one of their indigenous Caribbean samples the company uses? This person appears to be the same as the Atunwa Inaru, the name given to this ancient woman from the Bahamas. Anyway, we only shared a total of 3 segments, the longest at 3.6 cM. Our total Half-Match segments was 9.7 cM and a total of 458412 SNPs were used, with 47.467 percent fully identical. As one might expect, shared DNA with someone who lived several centuries ago would be low. Yet, with another kit from the same woman from Preacher's Cave, Atunwa Inaru on Gedmatch, we shared a total Half-Match segments of 14 cM, with the largest segment at 6.7 cM of 3 shared segments. 260057 SNPs were used for the comparison with 38.498 percent identical. We assume this match is at least somewhat reliable since geneticists using more sophisticated matches noted similarities in the ancestry of the Preacher's Cave sample and that of modern and ancient Puerto Ricans. We are not sure which of the two kits for the same person are more accurate or reliable, though we'd bet money on the first kit. 

The other two matches, however, remain more problematic. Both are from Cuba, although the date and exact location are unclear. One, called Brizuela on Gedmatch, is supposedly from the 12th century. With that sample, we share one segment at 4.5 cM, but at a lower SNP density. The other sample from Cuba shared even less with us, only 3.8 cM. But, with this Cuban sample, more SNPs and at a higher density. One wonders if these Cuban samples may reflect a 'ceramic' culture whose ancestors included some of the same groups who migrated into the Bahamas from Hispaniola. Or, alternatively, due to the widespread seafaring abilities of the indigenous Caribbean and the lack of evidence for inbreeding in the sample from Preacher's Cave, perhaps some groups found spouses or significant others from distant islands in the archipelago. What we do not understand, however, is our inability to find ancient samples from Hispaniola with whom we share DNA...

Playing with G25 Coordinates and Indigenous Ancestry...

Again, playing with G25 coordinates (non-simulated ones, this time) and different calculators or Illustrative DNA revealed some degree of genetic continuity from the pre-Columbian Caribbean and modern populations in the region. For example, using my G25 coordinates and Tomenable's calculators for the world in the 6th century BC and 1st century BC, I received the following results: 

Target: Yo Distance: 1.7369% / 0.01736942
46.0	Africa(West)_Niger_Congo_Peoples
22.0	Europe_Northwest
15.2	Europe_Southwest
6.6	America_Cuba_Guanahatabey
5.0	Europe_South_Central
3.2	Africa(Central)_Foragers
0.6	Africa_Garamantians
0.6	America_Ancestral_Puebloans
0.4	America_Amazon_Foragers
0.4	MENA_Judah_Kingdom

Of course, this is only slightly accurate, in that it's using our G25 coordinates to match with select ancient population samples. The sub-Saharan African category is about right, but probably based on mostly modern samples. Europe is probably generically accurate, although most of our European ancestry was probably from southwestern Europe (overwhelmingly the Iberian peninsula, which, over 2000 years ago, had populations genetically related to northwestern European groups). The Garamantian connection is probably a reference to other admixture calculators assigning us some minor degree of North African ancestry (and the Garamantian sample had sub-Saharan ancestry). Intriguingly, the indigenous Americas ancestry, a total of about 7.6 percent, is overwhelmingly linked to this calculator's archaic Cuba sample. Of course, reducing the number of populations to 5 and following other recommendations erases the Guanahatabey. It also reclassifies European ancestry as just Roman Empire-Europe.

Interestingly, Illustrative DNA's Three Way modeling does come slightly close to our ancestry with a best fit of Medieval ancestry as 43.6% Medieval Iberian (Girona), Lucayan Taino (7.2%) and Bantu (49.2%). Obviously, this is only roughly useful, but comes closest to approximating our ancestry as a split of 3 major populations: indigenous Americas, sub-Saharan Africa, and Europe (mainly Iberian peninsula). It's a shame sub-Saharan Africa isn't properly broken down and analyzed, but the difficulty of finding ancient DNA samples makes it somewhat understandable. Regardless, it does seem like our Indigenous ancestry, usually always linked to South American groups, was at least partly based on indigenous populations in the Caribbean. 

Fun with Simulated G25 Coordinates

Although using G25 coordinates from different admixture calculators and then using Vahaduo to run that information against G25 coordinates for various populations, past and present, is fun, it's not very reliable. However, using simulated coordinates from MDLP with a 1st Century BC calculator by Tomenable did match our data with a sample from Hispaniola, from Macorix Ciguayo. Using other simulated scaled G25 coordinates did not detect this, usually tracing our indigenous Americas ancestry to Amazon Farming Tribes and other South American groups. And, if we are being honest, about 2000 years ago was around the time when agriculturalists who likely spoke Arawakan languages were migrating to the Greater Antilles. Furthermore, we have no East African ancestry so the Africa_Cushitic Peoples ancestry is not accurate at all. Needless to say, none of this should be taken too seriously, but interesting to see nonetheless. The other set of G25 coordinates, which only listed Amazon Farming Tribes, was more accurate in other ways since it detected West African ancestry. 

Indigenous Ancestry

Like everyone else, we succumbed to genetic testing. Using AncestryDNA, which is said to sometimes inflate the indigenous ancestry of people with roots in the Spanish Caribbean, we found we do indeed possess "Amerindian" ancestry from the Caribbean. Playing around with GEDMatch has led me to think the "hacked" score of around 7% for Indigenous Puerto Rico may be more accurate than the 9% assigned in the official AncestryDNA results. Overall, this is not surprising. If the average Puerto Rican, according to studies by researchers, possesses around 14 or 15% indigenous ancestry, then someone who is half Puerto Rican will probably inherit some of that. 

We suspect the Bolivia & Peru ancestry detected at 1% might just be from the indigenous Caribbean ancestry. After reading studies on the "Taino" genome and their relations to South America, including the around 14% Ceramic-related ancestry found in Puerto Ricans, we found allusions to some degree of overlap with populations in Andean South America. While the genetic, linguistic, and archaeological evidence points to an origin among Arawakan-speaking groups of the Amazon region, the "Taino" also overlaps with other South American populations. Indeed, one study found links between pre-contact ceramic Puerto Rico and Amazonia and the Andes, although the strongest similarities in terms of mtDNA variation and haplotype frequency was with Eastern Tukanoan groups. What surprised us is that Ancestry reported the Bolivia & Peru was inherited from both parents. Perhaps our Haitian mother does indeed have very distant indigenous ancestry? 

Playing with GEDmatch also pointed to some "Amerindian" ancestry. Various admixture calculations suggest about 6-7% indigenous Americas ancestry, and the "archaic" matches did point to indigenous people in Brazil. Autosomal DNA comparison with the kit for an indigenous woman of the Bahamas who lived around 1000 years ago showed very minor connections, with the longest segment at 6.7 cm for a match. G25 simulated coordinates also included a prehistoric Brazil sample. This all seems about right, although I am sure I do not have mtDNA or Y-DNA haplogroups of indigenous ancestry. Indeed, playing around with raw data suggested J for Y-DNA, which also seems about right. Alas, sub-Saharan African ancestry is never easy to analyze here...