DNA and Family Trees

Traditionally, the history of families rested with a grandmother and that history was passed down word of mouth as certainly before WW1 much of the Lika Serb population was illiterate. Together with many records being destroyed over many wars it has left a huge gap in trying to document Lika family histories.

The advent of relatively cheap DNA testing has breathed new life into discovering Lika family histories. If nothing else it has encouraged people to document their family tree and make publicly available through DNA testing companies websites. It is however unfortunate that increasingly these family trees are being made private and therefore inaccessible to other family historians.

As part of a new initiative for the website, there’s an opportunity to share simplified family trees through the website and to post information requests. To begin with this will be controlled through the website to ensure that family trees are of a minimum standard and information requests are clear and concise. To help kick it off follow this link for a sample page consisting of the authors ancestry with indicative formats and how to send information to the website for family trees to be created.

The initial DNA testing impetus was in the Lika diaspora, particularly in the USA, followed by the other English speaking counties, but more and more people now living in Serbia and of Lika descent are testing and there’s a few who are testing the old folks left in Lika.

DNA Intro

The Lika Serb population was small, perhaps 125,000 at peak in the late 19th century, and until WW2 people mainly married within the Orthodox community and mainly from the same or neighbouring villages. In spite of ensuring that at least third cousins or closer did not marry, in the 400 years since settlement in Croatia this relatively small population has developed some endogamous attributes. This has implications for understanding DNA test results.

Multiple lines of relationship

Where a person is fully descended from Lika stock (both mother and father), the DNA results are typically skewed for their Lika matches as they are likely linked through both their parents. The testing companies are simply unable identify these multiple lineages and simply aggregate the values e.g. a ‘full Lika heritage’ person matching another ‘full Lika heritage’ person could be reported as being 3rd cousins but are likely to be 4th or 5th cousins through each parent. Where one parent has also been tested there is a better chance of making sense of DNA results and this improves as more family members are tested, e.g. a child inherits DNA equally from both parents and so where a child matches a third person, but the tested parent is not a match, the child’s DNA match must have been inherited through the other parent.

For Lika folk, the story is usually more complex, e.g. if the child and parent both match the third person, but the child’s match is the same or a greater measure than the tested parent, then the child is related to the third person through both their parents. Where a child and parent have both tested a useful rule of thumb to understand if there are multiple lineages at play is to compare the value (centimorgans cM – see below) of the match e.g. if the parent’s match with a third person is around double the value of the child’s match to the same third person then this indicates a single lineage match.

To improve the chances of discovering family through DNA there are a number actions that can be taken; i) Document your family history as much as possible (include village of birth) and share this information with your DNA matches, ii) Test as many known family members as possible through as many different strands of your family tree, iii) Focus on DNA matches that have limited Lika heritage, e.g. where a DNA match has only one grandparent or better yet one great grandparent from Lika the chances of multiple family links are greatly reduced.

DNA test types, providers and tips

In trying to rebuild a family history the ‘Autosomal’ DNA test offered by the main testing companies (Ancestry, 23andme, FamilytreeDNA, Myheritage and other testers) is the only choice. The ‘Autosomal’ DNA test and analysis provides the possibility of discovering DNA matches amongst the tens of millions of other testers to within 5 to 6 generations at most, which for the high majority of people is more than they will know about their family history. Beyond 1st cousins the results are increasingly only indicative of relationships e.g. a match estimated to be a 2nd cousin is usually accurate, however a knowledge of family history is the best way to confirm the DNA match estimate. Beyond 2nd cousin it becomes increasingly imperative to know your family history.

There are variety of reasons to choose one testing company over another and the following sections try to give some insight in to which one is the best for you. It is important to remember that DNA testing and analysis is evolving both in understanding DNA, the technology used and analysis techniques; in 10 years time it will be very different.

Spitting isn’t for everyone

FamilytreeDNA and Myheritage (uses FamilytreeDNA’s technology) use a swab test, the others require spittle. When testing older people the swab test is by far the best DNA collection method. Testing the oldest people is a priority as each generation loses part of its family DNA lineage context.

Database size and regional focus

The value of a DNA test is directly relevant to the size of the testing company’s database. Ancestry is still the leader with 23andme in 2nd place. Myheritage has increased substantially particularly outside the USA and is likely now larger than FamilytreeDNA.

Your test is compared against the company’s database to find DNA matches, the larger the database the more matches will likely be found. In trying to compete with Ancestry and 23andme on database size, FamilytreeDNA and Myheritage allow upload of DNA results from other test companies and provide a list of matches against their database free of charge, although there is a charge for their other reports. Unfortunately, Ancestry and 23andme have decided against offering this functionality.

All four of the main companies are looking to expand the volume of testers across different parts of the world and ethnicities from their initial western European focus. Anecdotally, Myheritage seems to be attracting more slavic testers particularly those living in European countries. As more people with Lika heritage test the more information will be generated about this community.

Another company, LivingDNA.com, is generating a broader database from its initial focus on the UK and it also allows DNA data uploads from other testers. iGENEA is an affiliate of FamilytreeDNA focused on Europe and seemingly Germany. This would be useful as many Lika people settled in Germany, but the company’ prices are nearly 3 times the cost of basic autosomal tests. iGenea’s results appear within the FamilytreeDNA database, but only list a person’s initials and the contact email is generic, which makes it near useless.

To have the least cost exposure to all four of the largest databases you will need to test with both Ancestry and 23andme and then upload the data to the other two testing companies. ‘Gedmatch.com‘ is not a testing company but accepts DNA data from all testing companies and in the experience of this website provides the best overall analysis free of charge. It does have chargeable elements for its more detailed reports.

DNA information provided

All the companies follow essentially the same process, although they use different technologies (FamilytreeDNA and Myheritage use the same) and differing algorithms for analysis. Ancestry is a standout in not providing chromosome level data for matches, which all the other testers do. It is surprising Ancestry continue to exclude this information which is readily available from its analysis process.

Testing company analysis of your DNA varies from company to company (due to technology, algorithms, database), although over the last 5 years the differences have been reducing. However, it is still relatively common that two people who match at a 3rd to 4th cousin level on for example 23andme do not match each other at all on Myheritage or Ancestry. At various times the companies have run into issues with their analysis and there are many examples of seemingly arbitrary results in the matching process (particularly for more distant matches), which when questioned through customer support the companies are unable to provide logical explanation. If nothing else these issues point to the fact that DNA understanding and analysis is still evolving. When the complexity of Lika Serb history and ethnicity is thrown in to the mix it is bound to result in some outliers and possibly a few inexplicable matches.

An interesting case is the occurrence of Ukrainian matches for Lika people. Although relatively low level matches, the number of such matches requires some explanation. Assuming the matches are correct, a possible reason is that an Austrian regiment which near exclusively recruited from Lika was deployed in present day south west Ukraine in 1915 during World War 1. The soldiers would have been living off the land between battles and skirmishes and some settled there after the war when released as POWs.

See also the below section on ‘Understanding DNA results’.

Autosomal vs Y-DNA & mtDNA tests

The vast majority of people purchase the basic Autosomal test as it provides the possibility of finding relatives within 4 to 5 generations of separation. The ‘Y-DNA’ (males only) and ‘mitochondrial (MT)-DNA’ tests offered by FamilytreeDNA and other specialist testers are, for the high majority of people, of little value in trying to discover relatives as these tests indicate deep historical relationships based on an individual’s Y-DNA & mtDNA Haplogroups. Open the page Understanding Haplogroups in a new window.

In certain circumstances it is useful to confirm the Y-DNA haplogroup e.g. where people share the same surname, confirming their ‘Y’ haplogroups will indicate whether there is also a shared ancestor many hundreds or thousands of years ago. The website https://www.yseq.net provides the means to test for a specific ‘Y’ haplogroup at a low cost (around US$20 in 2021 ). The result from this test is either ‘yes or no’ for a match to the specific haplogroup. Given there are thousands of haplogroups you need to be fairly certain that your specific ‘Y’ haplogroup is within a related group, although ‘Yseq’ do offer more expensive tests to narrow down the choice. In this case, compare the cost with the FamilytreeDNA ‘Y’ tests. As it rarely changes across thousands of years, the mtDNA Haplogroup provides no value for people looking to discover family connections.

The standard 23andme autosomal test identifies intermediate Haplogroups therefore providing the best introduction to this aspect of DNA testing. FamilytreeDNA charges extra to its autosomal test for the exact same information and the other testers do not provide this information at all. At significantly more cost FamilytreeDNA (and ‘Yseq’) also offer analysis to discover what is called a ‘terminal haplogroup’.

Women can discover their paternal ‘Y’ haplogroup by testing their father, brother, uncle (father’s brother), male cousin ( uncle’s son) or any other male that is known to have descended from the same male line as their father.

Irrespective of which tester is used the most important activity is to document your family tree with names, dates and village names within the testing company’s website. Without this information the chances of resolving DNA matches, other than close family, is relatively small. Secondly, interacting with your DNA matches directly via messaging on the testing companies websites will increase the chances of resolving ancestral relationships that DNA tests have high-lighted.

Understanding DNA results

The below table outlines the major elements of information provided by testing companies and the relevance for people with Lika heritage. As testing companies use differing algorithms for analysis of your DNA there are variances in results between testing companies. The makeup of each testing company’s database also impacts its results. DNA analysis is still in the early stages of development and as it evolves over time current generally accepted views will also evolve.

Ethnicity estimate
Testing companies change their ‘ethnicity’ estimates as more people test with them.
Given the relatively small number of testers with Lika heritage the ‘ethnicity’ data remains very high level with Lika folk variously identified as one or a mix of; East European, Croatian, Bosnian, Serbian, North East Italy, Balkan & Greek, South East European and other variants of these labels. As more Lika people test it will improve.

The most useful aspect of this information is to roughly estimate the earliest Lika generation of your DNA match e.g. if your match has around 25% of any of the above ethnicity types then it’s likely that one of their grandparents was from Lika and if around 12%, then it’s one of their great grandparent (assuming that only one parent has Lika ancestry).
Centimorgans (cM)
For close family to 1st cousin the cM value and the age of a DNA match can generally confirm the family relationship. The cM range for 2nd cousin is quite wide although testing company estimates for 2nd cousin are fairly accurate. Beyond 2nd cousin the variability in cM means you need to refer to family trees to resolve the DNA match, although some assumptions can be made to test relationship scenarios against other DNA matches.
It is generally accepted that bits of matching DNA (segments) of less than 7cM should not be used as indicator that there is a DNA link, however where there are multiple matching segments of over 7cM these smaller bits can be useful.
This chart provides an estimate of possible relationships based solely on the cM value with your match and another similar chart is provided here.
cMs, number of matching segments & chromosomes
The endogamous-like attributes of Lika folk is in part reflected in multiple bits of DNA (segments) with low cM values e.g. a 30cM match across 4 segments likely represents historic bits of DNA that have been passed down through multiple generations intact. So what seems a possible third cousin match may in fact be a fifth cousin. A 30cM match across 1 segment is far more significant as it’s unlikely that this biggish chunk of DNA has survived intact across more than 3 or 4 generations.
However, there are also issues with single large segment matches. As DNAPainter.com (see bottom of page) usefully documents, some areas of certain chromosomes are a wasteland of DNA bits and therefore matching DNA segments in these areas can be ignored unless you have multiple other matching segments with your DNA match. Myheritage.com in particular has suffered from this issue reporting single segment matches of around 40cM that clearly did not make sense and typically for these matches there were less than 10 listed matches. When this 40cM segment was plotted in DNAPainter it appeared within this aforementioned ‘wasteland’.
Ancestry continues to be the only company not to provide the chromosome and segment data; see the bottom of the page for one method to get around this.
Estimated relationship
This is the testing companies’ best guess based on the cM value and the number of segments. Beyond 2nd cousin, the estimates are difficult to interpret, which is reflected in the broad estimate range e.g. 3rd – 5th cousin, 4th to distant cousin etc. This is particularly difficult for Lika folk due to the aforementioned endogamous attributes of the community.
Shared matches
The implication of a ‘shared match’ can be mistaken for shared ancestor.
In Ancestry, it means that a third person is a match for you and separately is also a match for the match you are reviewing i.e. it does not indicate that all three persons match each other. However, some general assumptions can be made in context of other ‘shared matches’. 23andme has the same concept but they do indicate if there is 3-way match on a particular segment and Myheritage provides the same indicator. This 3-way match or ‘triangulation’ is indicative that all three people share a common ancestor.
Deeper understanding of DNA matches
Independent of endogamy there are a number of other considerations that impact understanding DNA matches. These are identified as: Identical by Descent (IBD), Identical by Population (IDP) and Identical by Chance (IBC).
IBD are DNA matches through shared ancestors. IDP are matches through shard population groups, that is, being part of a relatively isolated population group will tend to result in shared bits of DNA but no discernible shared ancestor. IBC is somewhat confusing, but in general suggests that small segments can be the same through pure chance. For a deeper dive on these types please go to this webpage.
In summary, testing across an extended family and multiple generations will help assert IBD matches and identify those matches that are IBP or IBC.

Other websites for analysis

In the experience of the author, www.gedmatch.com provides the best analysis of DNA. It accepts DNA data from all testing companies and has built a significant database and is typically used by people who want to dig deeper in to their DNA. Gedmatch provides a set of free reports and a set of ‘paid for’ reports many of which go well beyond the functionality offered by the testing companies.

Another useful website is www.dnapainter.com, which provides visualisation of your matches using a chromosome level view of your DNA matches although you will have to enter the matching segments for each match. By default Ancestry.com is excluded as it does not provide this information to customers. It sounds complicated but is in fact straight forward to use. The simple chromosome charts enable easy viewing of possible groups of matches. Myheritage and Gedmatch provide similar but auto-generated charts which they call ‘clusters’. In general, these charts are difficult to interpret as to what the results actually mean and it is limited to their databases.

It is disappointing that Ancestry.com continue to withhold chromosome information (as of May 2022) from its customers. Ask your Ancestry.com matches to upload their DNA data onto Familytreedna, Myheritage and Gedmatch to reveal the chromosome level detail.

Design a site like this with WordPress.com
Get started