DNA path may not lead to ancestors
Having your DNA tested for information about your ancestry seems to be the latest fad. Buyer beware. It is not very useful for ancestry research except as supplementary information to confirm what you already know.
It is very accurate for close relationships such as parent-child, grandparent-grandchild and first cousins, but most people already know the identity of these relatives. It is useful and accurate for adopted people who don’t know the identity of their birth parents and want help to find out. They will need tailormade DNA tests and preferably some prior knowledge of likely parents from whom a DNA sample can be obtained.
The standard DNA test offered by ancestry research companies is the autosomal test, which examines chromosomes 1-22 (the autosomes) and the X-chromosome. This gives you percentages of your ancestry by broad geographical regions. These state the approximate percentages of your DNA markers inherited from each region by comparing the frequency of your autosomal DNA markers with those of others in the various population groups.
This is done by computer searches for matches between your autosomal DNA and the autosomal DNA of millions of others on the database. Given that most individuals on the database are people still alive, this is information about distant cousins, not distant ancestors.
First cousins share grandparents, second cousins share great-grandparents, third cousins share great-greatgrandparents, and so on. Nearly all people from a defined geographical location alive today are about 13th cousins.
The best that can be said about these DNA ancestry percentages by geographical location is that they give you information about your ancestry indirectly from information about your cousins, and only in approximate terms. The percentages don’t tell you how distant is the identified common ancestry from the stated region.
Autosomal DNA matches cannot identify individual ancestors by name in the remote past. They can identify recent ancestors and close cousins individually, but most of us already know their identity. Autosomal percentages by geographic region can also provide information about ethnicity, with the same limitations.
The percentages of your DNA inherited from geographical regions tell you little more than you could have guessed. For example, if your grandparents or greatgrandparents came from UK/Ireland it is highly likely some forebears came from Europe and Scandinavia because of regular travel for thousands of years. A result such as, “30 per cent French/ German, 20 per cent British/Irish, 10 per cent Scandinavian” can’t be surprising.
The percentages will be less accurate the more distant the ancestry. This is because you inherit about half your DNA markers from each parent, a quarter from your four grandparents, an eighth from your eight great-grandparents, and so on.
As little as eight generations prior to yourself the proportion of your DNA received from those ancestors will be minuscule and possibly missed in the database search for matches. The markers from that distance may not even exist in your DNA because inheritance is more random from more distant ancestors.
Only 10 generations back, you have more than 1000 direct ancestors from whom you get a little of your DNA. Percentages by geographical region are not enlightening about all these ancestors, and pretty useless beyond 10 generations.
The autosomal test results by geographical region are said to be too inaccurate beyond eight generations prior to yourself. Other types of test can be more accurate than this for some purposes. For example, the mitochondrial DNA test is said to be accurate to 50 generations prior to yourself, but only along the female lines because this DNA passes from mother to child with few modifications.
The autosomal tests cannot identify markers of inherited diseases or specific inherited physical traits. For these you need other test types that cost more.
Specific tests for adopted people to confirm parent-child relationships are usually very accurate, giving 95 per cent probability that two individuals are parent and child, or 95 per cent probability they are not. These tests preferably need a DNA sample from both and examine the matching markers in the X- and Y-chromosomes or the mitochondrial DNA in females.
This feature of DNA testing — its accuracy for close relationships — is also useful for matching DNA at a crime scene with a suspect.
DNA tests for ancestry are expensive. You would be better to spend that money on a subscription to the ancestry research company’s service giving online access to a large range of civil and parish records. These, however, are only as accurate as the information provided by the original informant, and tracing a lineage can become error-prone the more distant the ancestry. But at least you get a person’s name that you could verify by other means. And they will be more use to your ancestry research than approximate percentages by geographic regions that tell you what you already know, or have guessed. Companies promoting DNA tests for ancestry seldom explain the limitations. They want your money.