This is an update to my earlier posting, An Avoidable DNA Error:
UPDATE:
I decided to touch base with Ann Turner about this, and I'm glad I did because she shed some new light on the situation. It appears that the article gave a sketchy version of events (and I should know enough to be on the lookout for that) -- and while there was still avoidable error, it was of a slightly different nature than what the article suggests.
Ann was aware of an article from The Journal of the Canadian Dental Association on the research and science behind this particular history mystery and brought it to my attention. According to Ann:
"I'm not sure I'd reconstruct events the way the Nova Scotia article did. I suspect they would have done the HVR2 test at the time if it were possible with the technology available back then. According to the article below, they were aware that two HVR1 results were the same, but they let the dental forensics trump the mtDNA. One sample was completely consumed by the HVR1 test."
http://www.cda-adc.ca/jcda/vol-70/issue-1/24.pdf
If you read the article, you'll see that they essentially had an mtDNA tie between two candidates, so they made the final identification based on dental forensics -- and that's where they apparently took their misstep.
I queried Ann further about the state of mtDNA testing at the time -- how viable it would have been to test HVR2 on degraded remains in 2002, and here's her response to that:
"I think it would have been possible if they'd had a sufficient sample. HVR2 is more difficult in general, though, because of the length variations with insertions and deletions around 309 and 315, and databases for comparisons are very heavily weighted toward HVR1 results, so the first step would have been HVR1 (which obviously consumed some of the sample).
In hindsight, they could have used techniques developed a few years later for identification of WTC victims, where DNA was degraded into very short fragments. If the Titanic project had tested HVR2 of the living descendants, they could have zeroed in on a short segment containing the critical difference(s). That would have a higher probability of success (but not guaranteed, of course)."
So yeah, they could have done more sophisticated testing -- but it would have been iffy based on the sufficiency of the available sample and more effort than time permitted -- and it still might have been inconclusive.
So thanks to Ann for the clarification. My bad for not delving deeper before venting, but I confess I'm happy to learn that this is an error of dental forensics, rather than genetics. Now, if only the article would correct its errors and give genetic genealogy credit for correcting a dental misjudgment!