DNA: Genghis Khan gives DNA tests a black eye

Bennett Greenspan’s talk included information on a New York Times article about Genghis Khan’s descendant being a mild-mannered guy named Tim Robinson in Miami.

Luckily, a production company that had planned to fly Robinson to Mongolia to meet his “cousins,” called Family Tree DNA for a second opinion before they purchased plane tickets.

Robinson tested at Family Tree DNA in 2003, and told them to call Greenspan for verification. Checking results, it was determined that if he were a Genghis descendant, his haplogroup would be C3, at one end of the tree of mankind. However, Robinson was an R1a, at the other end of the tree.

Bennett called Robinson to see if he could run a deeper test of ancestry called a SNP test. Robinson asked how fast the test could be run, since he was supposed to have lunch with the Mongolian ambassador.

The test was run quickly, confirming that Robinson was not Mongolian. “Fifteen minutes of fame went up in smoke,” says Greenspan, and the lunch was cancelled.

The New York Times printed a retraction.

Greenspan delights in asking audiences how many people read the original story: hands go up throughout the room. He then asks who read the retraction. Fewer hands, even in this DNA-smart audience.

The entire Mongolia incident makes Greenspan uneasy because an inaccurate story received huge coverage, with much less coverage for the retraction after the proper ancestry was confirmed.

“If the industry is going to predict such things, there must be accuracy.” He felt the incident was bad for the company that made the error, but also for the field as a whole. DNA test-takers need to have confidence in the results.

In another story, journalist Howard Wolinsky of the Chicago Sun-Times submitted samples to two genetic genealogy companies. He had already tested at Family Tree DNA. Amazingly, each of the two other companies made a different haplogroup prediction. Wolinsky was confused.

After retesting and reconfirmation, it was evident that Family Tree DNA’s results were accurate and the others were not. One company retested Wolinsky’s sample and confirmed Family Tree DNA’s original prediction. As of the conference date there was no news on the other company’s actions.

How many people test with multiple companies? It’s not known, but an error in labeling a customer can send him on a wild goose chase over several years looking for genealogical connections that don’t exist, such as a possible Native American ancestry prediction.

Although his company's 2003 prediction for Wolinsky was correct, Greenspan feels it is now time to raise the bar for haplogroup testing for genetic genealogy. His company is now providing a SNP assurance program. “If our system doesn’t provide a clear and unambiguous result for any Y-DNA sample, we will upgrade the sample to a confirmed haplogroup for free.”

“We don’t know what other companies will do when presented with the information that Howard has shared with all the principal DNA testing companies, but we feel the time has come to raise the bar for the benefit of all test takers,” he added.

The moral: Tell it correctly, or don’t tell it.