Fourth Annual DNA Grantees' Workshop

Monday, June 23, 2003

MORNING SESSION

California Department of Justice Experience with Linear Array System for Forensic mtDNA Analysis
Mark Timken
Biography

(Note: The PowerPoint presentation that supplements Mr. Timken's transcript is not available.)

MR. FRANK: Our next speaker is Dr. Mark Timken. Mark works as a criminalist with the California Department of Justice. He's a member of the Methods and Development Group. He received his Ph.D. from the University of Illinois and has worked in the areas of automated extraction and real-time PCR. Today, Mark is going to continue our first presentation and present on the California Department of Justice's experiences with linear array systems for forensic mitochondrial DNA analysis.

DR. TIMKEN: Thank you. I just want to spend a few minutes describing some of the results that we have obtained at the California Department of Justice (CA DOJ). The actual work that I'm going to present was done by Mavis Chong and Rom Kishore in the Methods Development Group. So if you have questions that I can't answer, I can certainly refer you to them, or to Sandy, and they'd be able to answer any questions you might have.

All of this work was done for the Missing Persons DNA Program that the CA DOJ is establishing. We've recently completed a validation of the mitochondrial DNA sequencing, and we're working on validating the Roche mitochondrial linear array kit. We are actually going to use it to generate amplicons, not only for the linear array, but also for DNA sequencing, so we don't have to waste any of our template.

We plan to establish a DNA typing and haplotyping reference database that we can use to screen and prioritize large numbers of samples for sequencing. What we've found is that it is a very sensitive measure of heteroplasmy, so we plan to use it for confirming heteroplasmy.

First I'll just show some data from our validation work. The first thing here is a sensitivity. I should mention that the actual arrays look better in person. It's nice that you get to see blue dots. They actually look better; the scan really doesn't do justice to them.

This is showing the sensitivity on the array strip as a function of input DNA, and this is input nuclear DNA, so it's 51.5 picograms of nuclear DNA. With 34 cycles, we can get good allele calls down to about 0.5 picograms, about 10 times less than the recommended amount. If we go up to 38 cycles, we can actually get good allele calls down to 0.05 picograms. So it turns out to be quite sensitive.

This shows the results of a study that was done mixing two different templates and two different individuals in 10-percent ratios. What you can see is pretty much what you'd expect: 100 percent of A down to 100 percent of B. The balance is good, and we can make calls or see the mixture down to at least 10 percent in the mixture, so it's fairly sensitive to mixtures.

Here are a couple of examples of possible case type situations. These data are actually from a case that was solved with STRs (short tandem repeats), because this isn't really validated for casework yet at our lab. But in this case, we (Rom Kishore) looked at two bones from a missing individual, bone number 1, bone number 2, as well as a reference sample from the putative mother.

What you can see, I hope, is that these individuals are related by STRs, as indicated by the good match with the linear array strip.

From the sequence data in the first locus (16093 locus), there's actually a faint band that is a heteroplasmy at this site. In the sequence work that's done, there's a little bit of T/C heteroplasmy that actually is reproduced and inherited in this case. This method is very sensitive to heteroplasmy.

So I think I'll just conclude by restating what I did to begin with. It seems to me that the linear array kit is very sensitive and a good tool for limited samples. There is some hands-on stuff, but it's simple to use, especially compared with sequencing. We use the same amplicons for sequencing and the arrays.

I guess there are two recommendations: One, make sure that you have a good, very stable hot shaker, because the hybridization is very sensitive to that; and two, calling alleles can be tedious. We worked at automating the interpretation, but we were very happy to hear that Roche is involved in trying to get an automated interpretation system that makes that part of the assay simpler.

So with that, I'll end. If you have questions, you can ask me, or feel free to ask and I'll forward them to my colleagues. Thank you very much.

MR. FRANK: Thank you very much, Mark.