Thursday, 17 September 2020

SAPDA, a new (and better?) admixture inference analysis

It is widely accepted in the genetic genealogy community that admixture analyses are currently still in their infancy, and most useful at the global, continental level. The SAPDA admixture inference analysis is such a global analysis. It was developed by Dilawer Khan and released in 2019, and is available on the website GenePlaza for a fee of approximately 14 USD. SAPDA is short for Shared Ancestral Population Defining Alleles.

Disclaimer 1: I am not a geneticist. This is a layperson's review written from a the perspective of a genealogist, and with a genealogist's interests in mind.

Disclaimer 2: I am not affiliated with Dilawer Khan, GenePlaza or Eurasian Genetics in any way, nor with any other DNA-related company or organisation.

What is SAPDA, and how is it different?

The SAPDA analysis uses a new and specially designed method to pinpoint East Eurasian and Siberian genetic admixture in West Eurasians – and vice versa – which is often overlooked by commercial admixture analyses (such as 23andMe) and other calculators based on the ADMIXTURE software. In other words, the SAPDA analysis presents itself as a superior alternative to ADMIXTURE, and, as such, purports to paint a truer picture of a person’s ancient ancestry than most other available analysis tools. As Dilawer Khan puts it, “SAPDA seeks to address some of the limitations of currently ancestry inference programs”.

Read Khan's own introduction to SAPDA on the Eurasian Genetics blog. The blog post also explains the datasets and reference populations used.

Here is a compilation of screenshots from my report (results calculated on 17 September 2020):

In the following, I will go through various aspects of my results and discuss them in depth.

West Eurasia

Most of my ancestry is European, and the high percentage of West Eurasian DNA is as expected. In every testing company analysis, my highest European region is Scandinavia, followed by the British Isles. My Finnish estimate varies, peaking at 10% (FTDNA myOrigins v1.0) and 10.5% (MyHeritage upload). SAPDA does not go into subregions, however.

East Asia and Siberia

SAPDA is able to pinpoint a substantial genetic contribution from East Asia and Siberia/the Americas (3.6% + 6.9% = 10.5% in total). Some Siberian and East Asian is expected, due to my Forest Finn and Sami ancestors (see e.g. this blog post); however, my score is unexpectedly high. My East Eurasian lies within the Northwest European average range, but my Siberian-American is distinctly elevated, even relative to Finns (compare references here). Besides Forest Finns and Sami, some of my Siberian and East Asian DNA might come from Romani, Malagasy, or Taíno ancestors, or from my Russian lineage, or from ancestors who were part of the Mongol and/or Turkic migrations into Europe – if not several (or all) of the above.

SAPDA separates admixture into “Ancestral” (more recent) and “Deep Ancestral” (more ancient) levels, based on analysis of older versus more recent shared mutations. Interestingly, my Ancestral graphs above show less West Eurasian and more East Asian and Siberian than the Deep Ancestral graphs. Dilawer Khan himself, commenting on similar results, has stated that “The slight increase in E Eurasian admixture in the more recent past is unsurprising for Europeans in light of the migrations west by Turkic and Iron Age Steppe peoples.” In other words, this admixture is real, and may be relatively recent.

East Asian DNA also shows up for me in testing company results. In an earlier version of 23andMe’s Ancestry Composition, I had 0.1% Broadly East Asian. In FTDNA’s myOrigins, I have <1% Northeast Asian. This Asian admixture (read by FTDNA as a combination of Northeast and Southeast Asian) is also carried by my father and his father, which is the line where my Sami ancestry comes in.

East Africa

Although I have some African (West African and Malagasy) ancestors born in the 1600s and early 1700s, recombination seems to have watered down my African genetic component to nearly zero, which is consistent with my results from other analyses, including my testing company results. Interestingly, however, SAPDA shows that I am homozygous for (having two copies of) 23 East African signature alleles in a cluster on chromosome 16, which makes it very likely that I do indeed have real East African DNA (according to Dilawer Khan, having two copies of an allele “would have a much lower probabllity of match due to a random event”). East African does show up in my Deep Ancestral graphs above, at 0.50% in admixture proportions and 0.40% in GSI.

Chromosome painting showing the locations where I have one or two copies of alleles defined as East African. SAPDA provides such chromosome paintings for each geographical region. Mousing over the red and blue marker gives additional specific information about the individual locations.

Visualisation of all my East African alleles and their frequencies in populations. Some of them seem to have a 0% frequency outside Africa, and are thus presumably informative of African ancestry. Similar visualisations are given for all the regions.

I know of one (possible) East African ancestor on my mother’s side – the Ethiopian woman Qirwerne who lived in the 1100s. If David Hughes' genealogical theory is correct (see previous link), my mother descends from Qirwerne through 8 known lines. My father is also possibly descended from Qirwerne, and if so, it would provide one possible explanation for why I am homozygous for the East African alleles.

Another possible explanation is that the East African DNA comes from my parents’ Romani ancestors, some of whom were probably originally Siddi people, i.e. descendants of Bantu-speaking East Africans who came to India as merchants, sailors, indentured servants, slaves, and mercenaries beginning in the 7th century AD. Romani scholar Ian Hancock argues that “From the very beginning, the [Romani] population has been a composite one […] Evidence points to Dravidian, Scythian, and even East African (Siddhi) input into the early mix of militia and camp followers” (my emphasis).

East African DNA also shows up for me in testing company results. In an earlier version of 23andMe’s Ancestry Composition, I had <0.1% East African (on chromosome 6). GEDmatch analysis showed this segment to be inherited from my mother.

South Asia

One of SAPDA’s drawbacks is that it does not report South Asia (the Indian subcontinent) as a separate category; it seems to be read as part of the East Eurasian category. Another calculator designed by Dilawer Khan (and with a similar purpose as SAPDA), the Eurasia K5, is able to pinpoint my South Eurasian ancestry at 1.0%. This can be explained as coming from my Romani lineage, in which South Asian DNA has been mixing with West Asian DNA since the 1000s and with European DNA since the 1200s. The low South Eurasian percentage is as expected, since the Scandinavian Romani people are a highly European-admixed population. Viewed in the context of my total amount of Scandinavian Romani ancestry (1/8 + 1/64 = 14.06%), my result is consistent with the testing company results and GEDmatch results of other people of Norwegian and Swedish Romani descent.

A person with a Romani great-great-grandmother got rather similar results to me in SAPDA, except that he had elevated East African (5% admixture, and 1.50%/1.60% in GSI Deep Ancestral/Ancestral).



I'm very happy with SAPDA. I especially like the fact that it shows (in the report's own words) "graphical evidence of genes shared with other populations in support of admixture percentage calculations". More generally, SAPDA gives support and plausible context to a number of trace regions I've been receiving at testing companies as well as from calculators on GEDmatch and elsewhere. As mentioned, the smaller percentages I get from SAPDA are consistent with the smaller trace regions I've been getting at 23andMe and FTDNA. And while my non-European percentages at SAPDA are higher than I expected, they do make sense from the historical context, and the Asian percentage levels are actually not that far off some of my other third-party results. All in all, I am now considering SAPDA to be probably the most accurate estimate of my continental-level ancient and semi-ancient ancestry, and yes, that is high praise.

I find SAPDA to be a very good analysis for those (like me) who are specifically interested in their continental ancestry in an ancient and semi-ancient timeframe. The only drawback is the lack of certain regions, such as South Asia. Although the report is a bit technical and demands some basic knowledge of genetics from the user, it has better structure and more features and interactivity than any other ancient calculator I have tried. SAPDA promises no more than it delivers, and it delivers exactly what it promises.

Updates on my genealogical research in 2020!

Dear all,

It's been a while since I've blogged. The past six months have been rough, with South Africa having one of the world's strictest Covid lockdowns, and myself being focused on finishing my PhD thesis. While I have taken time to do genealogical research, I haven't had the energy and capacity to share the results of my research on this blog. In this post, I'll try to do some catching up.

View from Signal Hill in early March, just before lockdown started. Photo by me.

My grandmother's father, again

The most important piece of news from 2020 concerns my now 15-year-long inquiry into the identity of my maternal grandmother's biological father. While the "official" father isn't completely discounted, he is unlikely as a candidate, and it seems very likely (based on DNA as well as historical circumstantial evidence) that our family story, which claims that my grandmother's father was a Romani man, is correct.

Last year, I found the identity of the man who was my great-grandmother's boyfriend (and, reportedly, fiancé) during World War 2, and who is said by many in our family to have been my grandmother's biological father. I strongly believed I had found the answer to the mystery. This man, Finn, was of part Romani descent, and as I described in my blog post, many of his Romani ancestors made sense in the context of my mother's autosomal matches.

However, it seems I was a bit too hopeful. My mother also has another match, to a woman with the same story as my grandmother: She is supposedly the daughter of an unknown Romani man, and comparisons with other people in her close family leads me to conclude that my mother matches her on her paternal (Romani) side. This woman is one of my mother's closest autosomal matches. However, the match does not fit with the theory that Finn is my grandmother's biological father! It is too large/close a match to be overlooked or explained away, and it suggests that our Romani connection is much closer than it would be if it goes via Finn.

Additionally, a closer look at my mother's Romani matches (people who are ethnically Romani, or people who at least know, by name, who their Romani ancestors were) reveals that some of them do not fit with the Finn theory, as Finn is not related to them - at least not within any plausible timeframe.

Taken together, all of this suggests that the family story is more literally true than I had imagined: It seems that my grandmother was in fact the daughter of a man who was not only of Romani descent (however distantly), but actually Romani, ethnically and culturally.

This fits with a lesser-known family rumour that my grandmother's father may not have been either of the two men I've researched so far, but a mysterious third man.

After a lot of hard, nitpicky work sifting through matches and comparing family trees, I arrived at the conclusion that my biological great-grandfather must have been one of two sets of brothers, sons of two brothers who married two sisters. Closer scrutiny favours one specific candidate, and he is the one I currently believe to be my grandmother's biological father. He was a man of fully Romani ethnicity (although like all Norwegian Romanies, his lineage is very mixed when you follow it back in time). He belonged to an old and famous Norwegian Romani family who lived the travelling lifestyle well into the 20th century. The man - my probable biological great-grandfather - died when I was a child, and I could theoretically have met and known him. For reasons of privacy, I will not divulge the man's identity, or the identities of his parents, on this blog.

To sum it up: It currently looks like my grandmother has one "official" father (who married her mother while she was pregnant), one rumoured father (who was her mother's boyfriend/fiancé during the war), and one actual biological father (the third man).

It has been a tortuous journey, and even though the connection is not 100% proven, I am happy that I've gotten even this far.

The Russian connection

Another breakthrough this year is that I've been able to corroborate my mother's father's family story of Russian ancestry with DNA evidence in the form of autosomal matches as well as admixture. An Ossetian-American genealogist has given me much assistance looking into these matches and to construct a plausible scenario through which we may be related to them. The connection is not just to Russia, but to a specific region of Russia, namely the Kuban area in the southern part of the country. Kuban is part of the Northern Caucasus region, and home to the Adyge people (also known as the Circassians). If my theory is correct, I have an Adyge 4th-great-grandmother and a Cossack 4th-great-grandfather.

Interestingly, a ship from this general region of Russia arrived in Oslo exactly at the time when my great-great-grandmother (the one who is said to have Russian ancestry) was conceived. This was the only Russian ship that arrived in Oslo during the period when she was conceived. My great-great-grandmother's mother worked in a hotel near the harbour, and it seems very likely that she might have come into contact with a passing Russian sailor. It is not proven that my Russian ancestor was aboard this ship, but it seems like a very good possibility. I have also been advised that he may have been a man who was stationed in Norway in some military capacity.

More Forest Finns

Another result of this year's genealogical research is that I've found several additional Forest Finn ancestors on my maternal grandmother's mother's side, in the Aurskog/Blaker area. It seems that my closest "fully" (culturally) Forest Finn ancestor was a 5th-great-grandfather born as late as 1801, which pulls the Forest Finn heritage considerably closer to me in terms of time (and generations). The cumulative Forest Finn lineages in my ancestry now add up to almost 3% of my ancestral background.

These are the main breakthroughs of 2020. I will try to blog more often from now on, but I can't promise regular posts. I'll be moving back to Norway in December, a process which will demand some time and effort.

I hope you're all doing as well as you can in these difficult times. Stay safe.

Friday, 14 February 2020

The diversity of an Afrikaner genealogy

I've received permission to write a blog post about the genealogy of a friend of mine, a South African woman who comes from a family of Afrikaner (also known as "Boer") heritage. The Afrikaners are primarily descended from Dutch colonial settlers in South Africa, and speak Afrikaans, which is often said to be a daughter language of Dutch. During Apartheid, the Afrikaners were classified as White.

My friend has ancestors who were farmers, artisans, traders, slaves and soldiers. Her family has been in South Africa for hundreds of years, and some of her ancestors took part in seminal events in South African history. Some worked for the Dutch East India Company. One was killed in a conflict with the Zulu people in 1840. One of my friend's 6th great-grandparents, German-born Peter Bekker (1673-1745), once escaped from the prison on Robben Island. Another 6th-great-grandfather, Johannes Bockelenberg (born in Germany in 1668) was married to a granddaughter of the famous Krotoa, the Khoi woman who worked as Jan van Riebeeck's interpreter and whose life has been portrayed in a recent movie. My friend, however, is not herself a descendant of Krotoa (as far as I know); her ancestor was married twice, and Krotoa's granddaughter was the first wife, while my friend descends from the second wife.

Several of my friend's ancestors were Frenchmen who established themselves as wine farmers in the valley of Franschhoek - "French Corner" - which is still famous for its wineries. I've visited the valley myself, and enjoyed some of its excellent wines.

In terms of ethnic origins, my friend's ancestral background looks like this:

Afrikaans/Dutch (presumed): 1422/2048 = 69.43 %
German: 404/2048 = 19.73 % (closest ancestor born in 1770)
French: 148/2048 = 7.23 % = (closest ancestor born in 1692)
Indian: c. 50/2048 = 2.44 % = (closest ancestor probably born in 1777)
Latvian: 8/2048 = 0.39 % = (closest ancestor born c. 1720)
West African: 5/2048 = 0.24 % = (closest ancestor born c. 1650)
Norwegian: 4/2048 = 0.20 % = (closest ancestor born c. 1705)
Chinese: 2/2048 = 0.10 % = (closest ancestor born in the 1600s)
Indonesian: 2/2048 = 0.10 % = (closest ancestor born c. 1660)
Unknown non-European: 2/2048 = 0.10 %
Malagasy: 1/2048 = 0.05 % = (closest ancestor born in the early 1600s)

Her total non-European ancestry adds up to 4.79%, of which the Asian/Oceanian part is 4.50% and the African part is 0.29%. It goes to show that even in South Africa, where racial segregation used to be law, the line between "White" and "non-White" was always less than clear-cut.

I find it fascinating to compare my friend's ancestral background to my own. My friend and I share several ancestral ethnicities and regions (Dutch, German, French, West African, Norwegian and Malagasy). Her non-European ancestry is connected to the Dutch slave-trade. Most of the slaves taken to the Cape Colony were from Asia, including the islands of Indonesia, which is reflected in my friend's genealogy. African slaves were few, and I actually seem to have more African ancestry than my friend does (my total being 0.49%, via Saint-Domingue/Haiti in the 1700s). However, it is not inconceivable that there could be some unknown additional African ancestors hiding in my friend's family tree.

Interestingly, my friend's Norwegian ancestor came from Fredrikstad, which is the city where my mother, stepfather and brother currently live. It's a small world indeed!

A beach outside Cape Town. Most of my friend's ancestors came to these shores on ships from Europe and Asia.

Monday, 27 January 2020

LM Genetics, version 2: Review of my grandmother's brother's new report

Lukasz Macuga of LM Genetics recently contacted me, telling me that he had upgraded his Ancestry Report service, and offering me a free report if I agreed to write a review of it. I accepted his offer.

Besides the fact that I was given a free report to review, I have no affiliation with LM Genetics.

Lukasz's reports are based on a person's Eurogenes K36 results, and I sent him the results of my grandmother's brother. The K36 result is a fully anonymous set of numbers, impossible to trace back to the tester, which means that Lukasz's analysis is absolutely safe in terms of privacy if you only send him this information. However, if you do not know your K36 results (i.e. if you are not on GEDmatch), you will have to send Lukasz your autosomal raw data file for him to analyse.

As long-time readers of this blog may remember, I did a review of the original version of Lukasz' report back in 2018, using K36 results from my mother and my grandmother's brother. You can read that review here.

Like the previous version, the upgraded report contains a mixture of various analyses, each contributing to the overall picture of a person's geographic ancestral origins. In the following, I will go through the main components of the report, comparing them to what we know about my great-uncle's ancestry.

Admixture estimates

My general impression from all the different parts of the report is that the main regions match my great-uncle's known ancestral heritage very well. He is mainly Scandinavian with significant admixture from Continental Europe (especially the North Sea region), as well as some from the British Isles and some from Haiti (a mix of African, European and probably Taíno). My great-uncle's LM Genetics report reflects this background, giving him Northwestern European regions as his top categories, with Scandinavia consistently coming out as his main region.

One of the strengths of the LM Genetics report (which was also there in the previous version) is that it breaks down Scandinavia into regions. In the nMonte3 oracle, my great-uncle's main Scandinavian regions are Denmark, Norway, and the southern Swedish province of Skåne. This is extremely interesting, as my great-uncle's ancestry is mainly from Norway, while he also has significant Danish ancestry further back, as well as a great-grandfather from Skåne. The Scandinavian subregional breakdown thus seems to reflect his paper-trail very well, except that the percentages are somewhat off. The nMonte3 oracle gives him 32.6 % Denmark, which is far too high and suggests that the algorithm is unable to fully distinguish between Denmark and southern Norway. This, however, is as expected, due to shared ancient and recent history. Interestingly, my great-uncle's Skåne percentage is 15.8% which is very much in line with his great-grandfather coming from that region.

In addition to the Scandinavian regions, my great-uncle scores small but significant percentages of German, which is in line with his semi-distant known German ancestry. Several of his German regions are in North and Northeast Germany, which is exactly where many of his known German ancestors came from. My great-uncle also gets 1.4% Yorkshire & Humber. This is very interesting, since his only known English ancestor was from Yorkshire. However, this ancestor was born in the late 1500s, and while the regional assignment is correct, 1.4% is far too high.

In addition to the regions that fit his known ancestry, my great-uncle also gets percentages of regions where he has no known paper-trail ancestors, such as Finland, Estonia and Ireland, and the north of Sweden. I'm not sure if these are misreadings, statistical noise, or indications of actual ancestors who are yet to be discovered.

Besides the "regular" nMonte3 oracle, the LM Genetics report contains targeted nMonte3 oracles which remove certain references (the ones closest to the target person's majority ancestry) in order to provide a closer look at the remaining minority admixture. Some "new" regions pop up in these targeted analyses. For instance, in one of them, my great-uncle scores 0.2% "FR_Finistere", which is interesting to me since one of his ancestral couples were indeed from Brittany, either from Finistère itself or from the next-door region of Côtes-d'Armor.

In the previous version of the report, my great-uncle scored 0.2% Sindhi in one of the oracles. South/Central Asian categories show up in the new report too, at 0.2% read variously as Gujarati, Punjabi or Tajik. My great-uncle also gets <2% Central Asian at FTDNA. It may point to distant origins from somewhere around northern India, which, to me, suggests Romani ancestry. In my review of the previous version of the report, I speculated on the possibility that the South/Central Asian should be added to the Eastern European percentages in order to find out the Romani ancestors' total genetic contribution. Just as before, this total seems to be in the range of 1-2% in the targeted nMonte3 oracles, suggesting a 3rd or 4th great-grandparent (if the person was a German or Scandinavian Romani and had German or Scandinavian DNA as well, then the total would be even higher and the ancestor would be even closer. Definitely a trail worth pursuing in the written records). My great-uncle still gets Eastern European in the new version, but this time the regions have been narrowed down even more, and seems to be strongly focused on Poland. Perhaps I am now one step closer to finding my great-uncle's elusive Romani ancestor? A connection through his great-grandfather from Skåne does not seem implausible, since Skåne lies just across the sea from Poland.

An interesting tidbit, which I do not believe was in the previous version of the report, is that my great-uncle now scores a tiny bit of Native American. This is in line with his FTDNA myOrigins admixture estimate. This DNA is a mystery - it may be either real Native American (and if so, probably Taíno) or a reflection of ancient Siberian origins through Sami or Finnish ancestors. It does, however, seem to reflect some kind of real genetic signal.


Since I have no background in statistics, there are some parts of the report that are a bit too advanced for me, specifically the MDS plot and the dendrogram. The Correspondence Analysis Plot seems to basically be the same as the PCA plots we knew e.g. from the old 23andMe ancestry reports.

Just like the previous version, the report includes a correlation map which "compares [the target person's] results to different regional averages around the world". The map visualises a person's likely ancestral origins through geographical gradients, which I think is a nice and honest way of presenting the data. My great-uncle's new map is almost identical to the old version, but the new version breaks down more regions in Norway, showing - correctly - that my great-uncle's ancestry comes from the southern and western parts of the country (his genetic affinity to Hedmark county is especially evident, which is correct, since he has no known recent ancestry from that area). The new version of the report also includes a second map, a Euclidean distances map, which is a different way of showing essentially the same data. For my great-uncle, the two maps are more or less identical, although not fully.

Both maps are a very good fit with his known ancestry within the last 500 years.

General conclusion

My impression with the new version of the LM Genetics report is very favourable. Indeed, I have very little to add to my previous conclusion: I am still very happy with the report. It is detailed, and seems to be at least as accurate - even at the European subregional level - as some of the testing-company admixture estimates (although it seems to overestimate minority admixture regions that are close to the target person's majority ancestry). However, it is also unfiltered and rather technical, which means that it takes some skill and previous knowledge to interpret it correctly. I would therefore recommend it only to advanced users.

In my review of the previous report, I stated that "My only disappointment with the report is that there are no micro-level regional breakdowns within Norway". This has now been remedied, which means that I really don't have any complaints left. I find LM Genetics to be a very good admixture estimate service.

Friday, 24 January 2020

What are you? Genealogy and identity

"What are you?" What are the labels you use to describe yourself? More often than not, the answer to this question has to do with a combination of many different things: Where you were born and raised, what environment you grew up in, and the choices you've made in your life. Not least, it usually has a lot to do with you genealogy.

What am I?
It's a question I've often asked myself, and one to which the answer is more complex than one might assume.

I was born in Norway and spent most of my childhood there, except for one year in the United States. I've grown up with mainstream Norwegian culture and know it very deeply. I've always assumed that "Norwegian" must be my primary ethnic identity, because it has to be. But does it really?

My known ancestral background is almost exactly 3/4 ethnically Norwegian (i.e. coming from the ethnic majority group known in Norwegian as etniske nordmenn). However, my relationship with Norwegianness is difficult and ambiguous: Having moved around a lot in my childhood, I lack the geographical groundedness which most other Norwegians seem to take for granted, and which tie them to a particular local brand of Norwegian culture. Additionally, my family has been city-dwellers for several generations on both sides (longer than usual among most Norwegian families I know of), which has separated us from the rural life which "primordial Norwegianness" is so often seen as rooted in. For me, my Norwegianness pertains mainly to my citizenship, the landscape and mainstream cultural environment I grew up in, and my native language (Norwegian). While there are times when I do feel a connection to a more rooted and ethnic Norwegianness, for example when playing certain forms of traditional music on the jaw harp (my favourite instrument to play), my relationship to this kind of Norwegianness is usually distant and tenuous and somewhat less important to me than might be expected from my large amount of ethnically Norwegian ancestry.

Perhaps the most important factor in this is the fact that I've often been made to feel I don't belong in Norway. There are stories to be told about my failed attempts at fitting in and attaining ownership of what was supposed to be "my" culture, but these stories go way beyond the topic of genealogy.

Lom stave church, from the 12th century. This was the parish church of some of my ancestors.
Photo by me (2019)

Being active in the online genealogy and genetic genealogy community, I often interact with Americans. These Americans sometimes comment on my family tree by saying "oh, wow, you've got such a deep connection to your country and culture" or something to that effect. It's meant as a compliment, of course, but more often than not, it makes me think "yeah... I wish I felt that connection!" It's something that I hope will change with time.

My 23andMe Ancestry Composition (October 2019), with overwhelming Scandinavian and only a trace of South Asian DNA from our Romani ancestors

Another identity that's been important is my life is Romani. Romani heritage was part of my maternal grandmother's identity. It was talked about as a matter of fact, and with pride, during my childhood. We even have some family traditions that might come from Romani ancestors. Feeling an affinity to a culture where travel and movement seemed to be very central, I became interested in Romani topics more generally, and started learning romani rakripa - the Scandoromani language - from Ludvig Karlsen's dictionary at the age of 14. Later, the Romani connection lead me to take an interest in Roma integration, the topic of my Master's thesis. I was also a paying member of a Romani community organisation for a while, in order to learn about the culture and language (a never-ending endeavour, of course, although that specific organisation is now defunct). If my research is correct, our closest "fully Romani" ancestors are a 3rd-great-grandfather - my maternal grandmother's great-grandfather - who lived from 1869 to 1947. Further research has suggested that I also have a Romani 4th-great-grandmother on my father's side, who lived from 1798 to 1834.

This cottage, named Furua ("the pine"), was the birthplace of Ludvig Karlsen and is considered an important heritage site for the Norwegian Romani community. Photo by me (2015), taken on an outing with my Romani organisation.

Thirdly, I've felt strongly connected to the Forest Finns ever since I found out that my family is descended from them. Our closest fully Forest Finn ancestor was born as far back as in 1704, but at least three of my great-grandparents have this ancestry (some DNA analyses estimate me to be as much as 10% genetically Finnish, and my autosomal match list is full of Finns). I grew up within the Forest Finn settlement area, and our primary school teachers took us out into the forest to see the remains of Forest Finn homesteads. As a teenager, I taught myself some Finnish. I've been active in the Forest Finn online community for years, and I've even taken part in some real-life political activism on behalf of the Forest Finns. The community itself acknowledges as a Forest Finn anyone who has an ancestral link and self-identifies as such.

MyHeritage gives me 10.5 % Finnish, more than any other testing-company admixture analysis

Typical taiga landscape near Gjøvik, right at the edge of the original Forest Finn settlement area.
An old fence is being eaten up by the swamp. Photo by me (2019)

Finally, my mother's father's family has believed that one of my great-great-grandmothers was of Russian descent. As a child, I internalised this as part of my identity, started teaching myself some Russian and developed a very strong interest in the history and anthropology of Russia. Then, after I started properly researching my great-grandmother's family history, I started thinking that the Russian story was just a myth, since there seemed to be no sign of any Russians in our family whatsoever. However, I have recently (February 2020) found out, with the very good help of Russian-American genealogist Luba Tabolova, that the story is probably true after all, and that my great-great-grandmother's biological father - who is not named in paper records - was most likely a Russian man. The finding is based on genetic evidence, and I will write a separate blog post about it at some point. It explains my mother's 5% Eastern European DNA showing at FTDNA as well as her many Russian autosomal matches. The ancestor in question, my 3rd-great-grandfather, must have been from the Northern Caucasus region, probably of part Kuban Cossack and part Adyge descent. The Kuban Cossacks are a mixed people of part Slavic, Turkic and Caucasian origin, while the Adyge are an indigenous Caucasian people. I have loved Adyge music for many years (click here for a sample).

There are many other ethnicities in my ancestral background, but I feel less personally connected to them, even though some of these connections are actually closer in time and generations than the Romani, Forest Finn and Russian ones. This goes to show that identity is about much, much more than just "blood quantum", DNA admixture, and so forth. To be clear: I acknowledge and celebrate all parts of my ancestry, including, for example, my 18th-century Black Haitian roots, but that does not mean I feel it would be right to call or consider myself Black or Haitian. Being of a certain descent is different from actually being that something and embodying it in one's everyday life and interactions. As any adoptee will tell you, ancestral background and personal identity are not the same thing.

In terms of my own identity and "being", I feel Norwegian, Romani and Forest Finn. These are the labels that have the most meaning to me; they are the labels with which I have an actual cultural connection, and which I feel say something about me as a person. Norwegian + Romani + Forest Finn is a very Eastern Norwegian mix, which actually does give me a certain geographical grounding in the general area where I spent most of my childhood.

Of all those labels, the Romani identity has always had a special place in my heart. I've been advised that I should claim it, in order to help de-stigmatise a stigmatised identity. I've decided that I will claim it, mainly because connecting to those roots is something that feels right.

After a couple of weeks of intense research in late 2019, I realised that I also probably have several relatively close Sami ancestors in addition to the ones I already knew about. The Sami are the indigenous people of central and northern Scandinavia and are divided into several linguistic and cultural subgroups. If the lines I've found are all correct, they amount to the equivalent of one Sami 4th great-grandparent. It all goes through one specific great-great-grandmother of mine, and the people in question were all poor, sedentary "village Sami" from the South Sami area, specifically Innherred, Nord-Trøndelag, Norway as well as Jämtland and Ångermanland in Sweden.

My closest ancestor who grew up (probably) fully within Sami culture seems to have been my 4th great-grandmother, who lived 1805-1891 in modern-day Steinkjer. She was the daughter of a single mother who seems to have been of fully or mainly Sami ancestry. The absent father was mixed Swedish and Sami.

The biggest surprise is that if all the lines are correct, my amount of Sami ancestry is as significant as my Romani ancestry, and more significant than my Forest Finn ancestry. I'm not sure if the Sami heritage in our family was simply forgotten with time, or if it was deliberately hushed up, as was common in many families due to anti-Sami prejudice (and indeed straight-out racism) within mainstream Norwegian society. Norwegian Sami experienced state-sponsored forced assimilation, known as "Norwegianisation" (fornorskning). The Sami connection must certainly have been known by my great-great-grandmother who was born in 1864. Her daughter, my great-grandmother, was born in 1905 and died in 1990, the year I was born. It is possible that she, too, knew. If she had been alive today, and if my theories are correct, then my great-grandmother would have been eligible to join the Sami Parliament's electoral roll based on the fact that her great-grandmother had Sami as her home language. Of course, this might never be provable on paper, and it is all very theoretical - but it is still interesting.

I have been interested in Sami language and culture my whole life, and I taught myself South Sami (great coincidence!) at a relatively high level as a teenager, with the help of a teacher named Helen Blind Brandsfjell. However, nobody has ever talked about anything Sami in my own family. Because of this, I am uncertain about how to relate to the newly rediscovered Sami part of my family. Many Norwegians are rediscovering and reclaiming lost Sami heritage. I, too, wish to acknowledge my family's Sami-ness as part of my own ancestral heritage, but since it is relatively distant in time and generations, and since we had no previous family stories of Sami ancestry, I am unsure how best to do it.

The best thing, I guess, is to do what I've always done: To acknowledge, remember and honour my ancestors, no matter who they were and what group they belonged to. Because that's what they were, after all: Human beings, like ourselves.

Thursday, 23 January 2020

Genealogical breakthroughs of 2019

As you have no doubt noticed, my blog has been on a hiatus during most of 2019. During this period, I have made several genealogical breakthroughs, some of which overturn previous findings. Here is a summary of the most important discoveries I've made in the year 2019, and they are significant!

My "new" great-grandfather Finn

The first and perhaps most important breakthrough (as it pertains to a very recent ancestor) is that I have found the likely identity of the man - named Finn - who is reputed to be my maternal grandmother's biological father. After discussing with family members and other genealogists, I have come to the conclusion that the story is probably true, and that this man is probably my biological great-grandfather. This Finn lived until the year 2000, and was thus alive during my own lifetime.

Finn's ancestry is an interesting mix of Norwegian, Swedish, Romani, and various European ethnicities (such as Walloons, Dutchmen, Baltic Germans, and even one man who seems to have been of Portuguese Jewish background). The Romani connection makes Finn fit very well into my grandmother's story of Romani ancestry. Indeed, although my grandmother's paternity is not fully proven, both candidates have Romani ancestry, which means that my grandmother's story is true regardless of who her biological father really was.

Finn's Romani connection goes via my 5th great-grandmother, a woman born in Sweden in 1822 and baptised as Amalia Victoria (no surname). Amalia Victoria died in 1902. According to her baptism record, she was the daughter of "travelling woman Anna Maria", and she grew up with non-Romani foster parents who later adopted her.

After a lot of painstaking elimination work in Swedish sources (and correspondence with several Romani genealogists), I have concluded that Anna Maria must be identical with Anna Maria Holm, wife of travelling horse-salesman Wilhelm Roos, to whom she had been married for many years by that time. In other words, my ancestor Amalia Victoria was the daughter - and youngest child - of Wilhelm Roos and Anna Maria Holm. This couple descends from some of Sweden's oldest Romani families. Our Holm family, for instance, is a branch of the Frantzwagner family, which has a very long history in Sweden and even has its own lineage society. Our Roos family descends from Erik Rosch, who claimed to be "king" of the Romani community in Finland and in 1787 petitioned the King of Sweden to be officially recognised as such. Erik Rosch is my 8th-great-grandfather. I will probably write a separate blog post about our newfound Romani ancestors at a later stage.

My "new" great-great-grandfather Marcus

Another important breakthrough is the discovery that my direct paternal great-grandfather Fredrik (born in Trondheim in 1904) was probably not the son of his "official" father, Martin Pedersen. This revelation comes from a combination of Y-DNA and autosomal DNA matching:

(1) Our Y-DNA clearly shows that we descend from the Scottish-Norwegian Matheson family. This does not fit the paper trail, and leads one to conclude that there must have been an NPE somewhere along our direct paternal line.
(2) Several distant relatives of Martin Pedersen have tested, but do not show up as autosomal matches to my paternal grandfather. This is not conclusive proof that Martin is not our ancestor, but combined with the Y-DNA evidence, it is cause for doubt.
(3) We do match several relatives of Martin's wife, my great-great-grandmother Oline; she is without doubt our genetic ancestor.
(4) We do match a close-ish relative of a Matheson man who lived in Trondheim at the time of my great-grandfather's conception in 1904.

Taken together, this leads me to conclude that Oline most likely had my great-grandfather Fredrik with someone other than Martin Pedersen - someone who belonged to the Matheson family. I have discussed this with a prominent Norwegian genealogist, who agrees with my assessment. My grandparents, the oldest living generation on this branch of my family tree, have been informed of this development.

Several members of the Matheson family lived in Trondheim in 1904, but I have eliminated all but three as potential candidates for Fredrik's biological father. Further digging and weighing of evidence leads me to believe that Fredrik's biological father was, in all likelihood, a man named Marcus Fredrik Matheson (1836-1905), a master baker by profession and a relatively old man at the time of Fredrik's conception. He is the man whose close-ish relative my grandfather matches. I may write a separate blog post about this at a later stage, detailing exactly how I came to the conclusion that Marcus is my probable ancestor. If any descendant or close relative of Marcus reads this, please contact me!

Interestingly, this discovery means that I am not, in fact, a genetic Pedersen. The surname refers to Martin Pedersen's father Peder, who, if my theory is correct, is not my genetic ancestor at all.

South Sami ancestors

One huge genealogical discovery came right at the end of 2019, namely the fact that my paternal grandfather likely has a significant amount of South Sami ancestry (around 1/16), within a relatively recent timeframe. This deserves its very own blog post, and soon, so I will not elaborate on it here.


Finally, I want to mention that my previous theory about my Haitian connection is sort of validated. It seems like we are probably connected to the same family I thought we were connected to - but on a slightly different branch of our family tree, and one generation further back in our matches' family tree! Things are still iffy, but it looks very likely that a certain 4th great-grandmother of mine was the illegitimate daughter of Haitian sailor Pierre Étienne Lévêque (whose mother was Jeanne Galbo, the former slave - which means we are probably descended from her after all!). There are a couple of other father candidates too, but Pierre Étienne is the most likely one, because he is specifically said to have been a sailor. I will probably write a separate blog post about this too, later on.

Thursday, 16 January 2020

Ancient "ethnicity" estimates from MyTrueAncestry

Hello everyone!

My blog has lain dormant for almost a year. This has been due to a combination of work pressure, personal issues, and the general busy-ness of life. I have, however, made some significant genealogical discoveries during 2019, and as part of revitalising this blog, I will soon be posting about them here. In the meantime, there have been some new arrivals on the third-party autosomal analysis tool scene, and I will begin 2020 with a couple of reviews of such tools.

First out is the company MyTrueAncestry, which has started providing admixture estimates based on ancient samples. This is part of their free service, and lately I've seen many people sharing their results on social media.

My result looks like this:

Celts: 24.2 %
Vikings (Swedish): 18.1 %
Danish Vikings: 17 %
Longobards: 13.3 %
Vikings (Icelandic): 8.82 %
Saxons: 7.26 %
Norwegian Vikings: 5.72 %
Franks: 2.37 %
Vandals: 1.49 %
Ostrogoths: 0.83 %
Visigoths: 0.68 %

The percentages are not actually listed; you have to mouse over the sections of the graph to see them.

While the estimate looks interesting at first glance, it also raises many questions in my mind: (1) The terminology (Vikings are not an ethnic group); (2) Anachronisms (Vandals and Vikings lived hundreds of years apart); (3) Geography (why do they differentiate between Norwegian and Icelandic Vikings?); (4) Information value: What is the relevance of estimating ancestry from populations so far back in time that basically everyone descends from them anyway?

Most importantly, I question whether it is really possible to distinguish genetically between closely related ancient peoples, such as the Ostrogoths, the Visigoths and the Vandals, who were all East Germanic tribes sharing a relatively recent common linguistic and cultural origin. I would have to see a detailed white paper (which MyTrueAncestry, to my knowledge, does not provide) to be convinced that my 1.49 % "Vandal" result means that this part of my DNA actually came from ancestors who were part of the Vandal tribe. Of course, it is a statistical certainty that I descend from Vandals, just like every other European person does; the question is whether or not the specific segments labelled "Vandal" by MyTrueAncestry can be attributed to those Vandal ancestors.

What this estimate really tells me is that I'm basically European, a mix of Celtic and Germanic (mainly North Germanic) peoples. My Germanic total is about 3/4 of my genetic makeup, while my Celtic total is about 1/4. The Germanic-Celtic proportions are obscured by the fact that everything "Celtic" is baked into one large umbrella category, while the Germanic components are spread out across a number of different categories. This makes my "Celtic" component look much more significant than is actually the case.

While the estimate is probably not wrong per se (3/4 Germanic and 1/4 Celtic seems reasonable for a Norwegian with known British and Continental European ancestors), I could have guessed it beforehand based on my schoolbook knowledge of European history. The analysis missed my Uralic (Sami, Forest Finn) and non-European contributions, which shows that there is a margin of error, possibly quite a significant one.

My conclusion about this ancient "ethnicity calculator" is that it may have some value for those - like myself - who are interested in ancient ancestry; however, value is lost because the information is presented in an unhelpful way (i.e. the breakdown of Germanic categories versus the all-encompassing "Celtic" category") and due to the problems of terminology mentioned above (the term "Vikings", anachronistic juxtapositions of ethnonyms, as well as unhelpful geographical divisions).

It also seems to give false negatives for small amounts of minority admixture, as none of my small Uralic or non-European components show up.

My verdict about this tool: Use it for entertainment purposes if you must, but interpret it critically, and take the group labels with a whole spoonful of salt.