Introduction
In today’s technology-rich society, we often times forget that technology is still imperfect at best. This reality applies to Direct-to-Consumer DNA testing we use for genealogy. Despite our advancements in DNA testing over the last couple of decades, its underlying technologies are still limited and are continuously evolving. While many genetic genealogists, especially those who are new to DNA testing, are frustrated with these limitations, there are still ways to work with the results to help inform and advance your genealogy research. Part 3 of this blog series is not meant to scare people away from DNA testing. Its purpose is to help inform your understanding of the current state of DNA testing and to help manage your expectations. The good news is most of the major testing companies are working with their partners in the scientific and technology communities to build better testing equipment. Thus, overtime, many of these limitations will likely be improved upon or outright eliminated.
Here are some of the current limitations that exist with today’s DNA testing
1. Relationship estimates. As part of your test results, testing companies provide you a list of genetic matches with whom you share DNA. These genetic matches are people who have also conducted DNA testing with the same company, such as 23andMe or AncestryDNA, or transferred results from one company to another (such as transferring a copy of their AncestryDNA results to Family Tree DNA). The companies compare your DNA with other customers in their databases. If you meet their matching thresholds (more on this topic below), then the companies will add you to your matches’ lists and your matches to your list.
Along with the list of genetic matches, the testing companies provide an estimate or prediction of the type of relationship that you have with all of your matches based on the amount of DNA you share with them. Companies can only provide an estimate of the relationship because it is nearly impossible for companies to state conclusively the exact relationship between two DNA matches for a couple of reasons.
First, when it comes to autosomal DNA (the type of DNA that AncestryDNA or FTDNA’s Family Finder test), the amount of DNA you share with a match could be indicative for several different types of relationships. For instance, if you share 25% autosomal DNA with another person, that could indicate either a grandparent-grandchild or aunt/uncle-niece/nephew relationship. If you share 1% DNA with a match, such an amount could be indicative of around two dozen or more types of relationships (different degrees of cousins essentially). Y-DNA test results are similar in that companies can only provide a modest estimate of the possible distance two men shared a common paternal ancestor based on the difference (mutations) found in their respective Y-chromosomes. The estimates for how many generations back the shared paternal ancestor lived can vary significantly.
Second, DNA testing companies do not know the actual relationships among their customers when testing DNA samples. They simply see a person’s DNA sample as an individual sample that requires testing. They are not testing the person’s sample with the view point of “this sample is from X Cosgrove’s first cousin.” Thus, they do not factor in people’s actual relationships when looking for DNA matches within their databases nor when providing their relationship estimates. Personally, I am absolutely fine with this because I want the objectivity in the reported results based upon the amount of DNA I share with my matches, not on any prior knowledge a company may have about my matches and I.
2. Test results themselves are imperfect. While we have made significant strides in DNA testing technology in recent years, it is still imperfect, especially when analyzing small amounts of DNA shared between two people. There are known false positives in test results, resulting in a company reporting two people as DNA matches when in fact they are not related at all. There are also known instances when a false negative is reported between two people. A false negative is when the testing company commits an error by either failing to accurately match two genetic relatives with one another or by omitting DNA matches from each other’s match list. Both false positives and false negatives tend to occur when the shared DNA between the genetic relatives is very low.
Each company has its own matching threshold that they use to determine when they assess two people to be valid DNA matches with one another. They do this to limit the number of false positives their customers receive. Unfortunately, their efforts to eliminate false positives do result in occasional exclusion of valid DNA matches from your match list. While this frustrates many genetic genealogists, we need to keep in mind that the amount of DNA we share with our excluded matches are relatively small amounts so it is not like companies are omitting parents off of children’s lists or grandchildren off of their grandparents’ lists. You should feel relatively confident that people predicted to be 4th cousins or closer are likely valid genetic relatives, whether or not the predicted or estimated relationships are correct.
Many experts in the field of genetic genealogy suggest that false positives are likely to occur in the 10 centimorgan or lower range among matches (centimorgan is a level of measurement of DNA; 10 centimorgans is a very small amount). Other experts suggest that you should be cautious when analyzing matches that you share as much as 15 centimorgan of DNA. And nearly all experts agree to ignore any matches that you share less than 7 centimorgan of DNA. However, exceptions do exist. I have a couple of distant cousins who I share around 6.5 centimorgan of DNA with who I had documented in my family tree based upon evidence we collected through on our shared genealogy research. We took our DNA tests a year after I had added them to my family tree. While we appeared on each other’s DNA match lists, the amount of DNA we shared were below most experts’ recommended thresholds.
Even though I have a couple of valid exceptions to the general rule of ignoring matches below 7-10 cM range, I can state with a high degree of confidence that many of my lower matches are either false positives or so distantly related that it is highly unlikely I will ever be able to prove how we are actually related.
3. Ethnicity estimates are just that……estimates. The technology does not exist (yet) that can provide a person their exact ethnic breakdown with 100% accuracy. Hence, this is why all of the companies provide ethnic ESTIMATES as part of their autosomal DNA test results packages. Some companies will claim to be more accurate than others, but to me, this is a subjective statement made by those companies for marketing purposes. Since their reference population panels differ from company to company, your ethnicity results will differ if you test with multiple companies. I don’t know of anyone whose estimates are similar from across multiple companies.
It is important to understand that to build their reference population panels, companies had to collect DNA from people who allegedly have deep ancestral roots in those specific areas of the world. Its highly probable that each company used different people to build their own unique reference populations. For example, 23andMe and AncestryDNA likely used different people to build their reference population for Ireland. Not only did they recruit different people, they likely recruited people from different parts of Ireland as well. Thus, the genetic make-up of the people within their respective reference panels are genetically unique and differ enough that your results will also differ between the two companies.
This becomes critical as many companies attempt to refine the specific geographic locations from which your ancestors originated. Thus, while 23andMe and Ancestry may both provide estimates on your Irish ethnicity in the same general range as the other (say 23andMe reports 30% and Ancestry reports 50%), they may differ where within Ireland they estimate your ancestral DNA originated based on the composition of their own unique Irish reference panels. In my case, Ancestry reports a significant Ulster DNA signature in my DNA profile whereas 23andMe has suggested I have a more balanced profile from across all of Ireland. In essence, both are actually accurate when you look at where my various Irish family lines originated in Ireland (my paternal grandfather’s Irish lines come from across all parts of Ireland while my paternal grandmother’s lines are mostly Ulster Irish in origins.
4. Our understanding of ethnicity is tainted by today’s world map and population migrations over the centuries. Something that the majority of people overlook is that your ethnicity results are looking REALLY FAR back into the time capsule that your DNA truly is…..over 500 years back in time in fact. There are countless people who have voiced their frustration with their ethnicity results on Facebook forums and blogs. A person may state that their paperwork trail (genealogy speak for research findings) shows they are 100% Irish, but their ethnicity results show 50% Irish, 25% German, and 25% Scandinavian. While there is likely an element of inaccuracy in their ethnicity results as I stated above, what people don’t realize is the real possibility that some of their ancestors may have actually migrated to Ireland long before records were produced and maintained. Thus, the paperwork trail either no longer exists or never existed that showed their merchant ancestor from Germany decided one day to stay in Dublin as opposed to head back to Germany where he was born and raised. Or a brief relationship occurred between a transit foreigner with a local person that resulted in a child several months later. It happened. A lot.
People have to keep in mind that many of today’s nation-states did not exist 500 years ago……and for those that did, their national boundaries differ today than they did back then. Nationalities didn’t exist in many parts of the world as we know them today. Thus, we tend to look at our ethnicity results through today’s lens as opposed to how the world looked 500+ years ago, which would be more relevant and accurate when analyzing our ethnicity estimates. Wars, famines, epidemics, and economic crises caused millions of people to migrate from their ancestral lands to new lands over the centuries. Plus, people then, as they are today, were curious and adventurous. They may have been traders, adventurers, or soldiers of fortune who fell in love with a new land and decided to stay. All of these events were common in our shared human history, though many genealogists tend to forget about such occurrences and their frequencies, which if they took the time to study the actual history of their ancestral lands, could see a viable explanation in front of them for why their ethnicity results may differ from their “modern” paper trails.
5. Autosomal DNA itself has its limitations. The amount of DNA we inherited from our ancestors decreases by 50% with each passing generation in your tree. For instance, I share 50% of DNA with each of my parents; 25% with each of my grandparents; 12.5% with my great-grandparents, etc. and less than 1% with each of my 5th great-grandparents. I share even less DNA with my distant cousins than I do with the ancestors who we have in common, if I share any DNA with them at all.
As you can see by this brief example, you share less DNA and likely very little to no DNA with distant ancestors the further you go back in your tree. Thus, if you are trying to break through a brick wall involving a set of your 6th or 7th great-grandparents, the odds of you being able to do so by using autosomal DNA test results are low. This limitation is not due to limitations in existing technology; its a limitation due to the biological processes involved in how we humans transmit DNA from parents to children from generation to generation.
Keeping things in perspective
I will leave my readers with a few recommendations to keep in mind to help overcome our known limitations with today’s testing technologies:
- NEVER make a conclusion on a relationship between two or more people based on your DNA results alone. DNA results provide additional data points that must be used in conjunction with other credible and vetted evidence produced through traditional genealogical research methods.
- Recognize that 50% of fourth cousins (cousins who share a pair of 3rd great-grandparents) do not share DNA with one another. Yes, the actual relationship is not that distant, yet half of fourth cousins will not share DNA with each other. Having said that, the amount of DNA you could share with a fourth cousin could range from a moderate amount (~100 cM) to very small amounts. Thus, just because you and a 4th cousin, who you discovered through your genealogy research, do not share DNA does not mean you are not genealogically related or that you made errors in your research. It just means that you are not genetically related. Having said that, I would take the time to re-verify your existing research findings and try to find additional evidence to strengthen your conclusion that the two of you are indeed genealogically related.
- Don’t put too much weight into your ethnicity estimates. All of the companies periodically update their reference panels in hopes of improving both the accuracy and precision of the ethnicity estimates. Some people feel the updated results are a better reflection of their known ethnicity while others feel that the newer results are far less accurate than their original results. Personally, I think there is value in ethnicity results for people who are starting from ground zero, such as adoptees, who may not have any understanding and knowledge about their biological families. In these cases, the ethnicity results may help the genealogists with little knowledge of their family background to focus their research into certain records collections. Outside of these cases, I would not invest much time into one’s ethnicity estimate, especially if you have well documented research and comprehensive family histories..
- Keep in mind that the technology used for DNA testing is constantly evolving for the better. Some of the limitations that exist today may not exist in 5 to 10 years from now. Thankfully, many of the companies update their databases and test results and introduce new tools periodically, most of which are free for existing customers. I like to tell people that DNA testing is an investment that continues to generate dividends overtime. We just need to practice a bit of patience.
Our next blog entry in the series will focus on some of the more common frustrations people experience with DNA testing.