Fourth Annual DNA Grantees' Workshop
Tuesday, June 24, 2003
MORNING SESSION
Novel STR Multiplexes with Reduced Size for the Analysis of Degraded and Contaminated Forensic DNA Samples
Bruce R. McCord
Biography
MR. TAMBASCO: Our next speaker, Bruce McCord, has been in Ohio for about 4 years, and we're so glad to have him. Bruce earned his bachelor's degree in chemistry at the University of William and Mary, and I believe that he has a Ph.D. in analytical chemistry from the University of Wisconsin. Bruce is currently on staff as the director of the forensic chemistry program at Ohio University, and he is turning out some wonderful students. He is going to speak to us today about novel short tandem repeat (STR) multiplexes with reduced size for the analysis of degraded and contaminated forensic DNA samples.
DR. McCORD: This is a project that actually started in the summer of 1999. I had an exceptional student, whose name was Kerry Opel (she had a master's degree in anthropology at the time), who came to me and said, "I want to look at some anthropological samples." She had a joint interest in forensics and anthropology, and we had a forensic anthropologist, Nancy Tatarek, on staff. McCord: Slide 1
So we collaborated and started to think about how we were going to look at some of the samples. I had one 310 at the time and didn't want to switch it over to do mitochondrial work, so we started to think about what we could do about that. I had recalled that an associate, John Butler, had been working on multi-time of flight (TOF) and we didn't have a multi-TOF at the time, so we decided to just try some of those primers on the 310.
That's how this project got started. It's on the development of CODIS (Combined DNA Index System) multiplexes with reduced size for the analysis of degraded DNA. The primary justification for doing this is the fact that we're all very familiar with large multiplex DNA kits. They're rapid, they're efficient, and they're great, especially for the analysis of convicted offender samples. McCord: Slide 2
Unfortunately, not every sample run in a lab is going to be spotted on a stain card. While these large multiplexes work very well for stain cards, skeletal material is quite a different matter. These types of samples represent a special challenge, as do other types of degraded and disturbed DNA.
With degraded DNA (this is from a presentation by Lynne Helton at a workshop at the MAFS STR Symposium in Indianapolis), you get this degradation curve, where the large amplicons don't amplify very well and they disappear, and you worry about allelic dropout when you have this problem. You worry about intensity thresholds and a lot of things. McCord: Slide 3
In these situations, where you have no amplification or poor amplification with STRs (short tandem repeats), you instantly start thinking about mitochondrial DNA. But if you think about it—and of course I am going to be provocative here—the real question is, why don't you have amplification? Is there no DNA there? McCord: Slide 4
Well, in the case of hairs maybe that's true. But in the case of a bone sample, because the person grew and lived, there ought to be some DNA in this sample other than mitochondria. It may not be that there is no DNA. Instead, it may be that the DNA itself is degraded, or alternatively, you have PCR (polymerase chain reaction) inhibitors. You know the DNA is there, but you can't access it. Maybe it's too badly degraded and for all intents and purposes it's not accessible. But maybe it's just the fact that your amplicon is too large.
The other issue, of course, is that mitochondrial DNA statistics are bad (the heteroplasmy causes some problems), but you want access to the CODIS data for your degraded samples. McCord: Slide 5
So we took this approach whereas, instead of going for massive multiplexes, let's take little multiplexes. They're called miniplexes. Take the primers from each STR amplicon and make them as short and close as possible to the primer dimer sites. Obviously, there's a minimum size that you can work with because you've got the size of the two primers. You've also got the repeat region in the center, so the minimum size is somewhere around 60 with 20 base primers. McCord: Slide 6
You can then take this situation. Obviously you can't put more than one STR in each lane because otherwise they would overlap with each other as you push them all down to that 60 base sort of cutoff. Then you look at this and say, "Well, how is this going to work? How well will we be able to amplify samples and see if we haven't developed a new kind of way to access degraded DNA?" McCord: Slide 7
There are some wonderful things that occur once you start reducing the sizes. To do this, you take the current primers—and you've got your repeat region, so you have to be concerned about the whole situation—and move them in. The problem, of course, is that you know you've got an unstable region around your repeats, so you have to deal with that. You've got to try to get as much information as you can and take advantage of the fact that you can do this. McCord: Slides 8 and 9
So, we got together and designed a whole series of different miniplexes. We tried to encompass the entire set of CODIS loci as well as some of the other loci in Powerplex. Each one is labeled with a different colored dye. You can also, if you've got the five-color system, which, thank God, we do in our lab now, look at that orange lane as well. McCord: Slides 10–12
When looking at the data, one of the things that the postdoc in my lab said was that we could put more than one loci in each site because some of sites were big enough to squeeze two of them into the site if we designed the loci just right. As a joke between John, her [postdoc], and I, we called this the "big mini." So, we have little minis and big minis. McCord: Slides 13 and 14
The next big question was how does this work with degraded DNA. Our initial approach, was, well, we were going to physically degrade the DNA. We got several students to donate large amounts of blood, and we digested all this blood and looked at different size ranges. This was done by Denise Chung in my lab. Denise did the standard thing where you digest the DNA, but to put a little twist on this, we cut out sections, so we knew exactly what the sizes and size ranges were (not just some vague ranges but absolute ranges) and where the digestion occurred. McCord: Slides 15 and 16
We first looked at the CTTv multiplex, which is composed of four STR loci: CSF1PO, TPOX, TH01 and vWA. (This is an aside to you: To get tiny ladders from the regular ladders, we just amplify the regular ladders. The primers are inside because they're smaller, so it's a perfectly simple thing—well, not completely simple—to take a an existing commercial allelic ladder and make it smaller.) If you just take a standard setup of CTTv and look at what the effect of size is on these samples, you can see that we're doing just fine once we get above 1,000 base pairs in template size or average template size. But once we start getting smaller (template size range of 350), we lose CSF1PO. When we drop it a little smaller, we then lose both TPOX and THO1. McCord: Slide 17
The strong effect that size has on the ability to amplify the size of your template is very clear, and looking at comparisons with Powerplex, again you see the effect of the small size of these markers and the efficiency of amplification. The bigger the difference in size between the amplified product, the more intense the effect. For very small products, you get very efficient amplification; that is, when you drop the size of the STR that's being amplified, the efficiency of amplification should improve. McCord: Slides 18–20
Of course, it's not all that easy. There are bugaboos or problems. One of them is adenylation. You have to add special sequences to the 5' end and make sure that you don't get plus A or minus A effects. A big problem we have is interference from free dye—the dyes that fall off. Because we're purchasing the primers, you don't go through quite the quality control that you would normally do. Most of the dyes that we see are there as we purchased them; they're not an artifact of thermal cycling. McCord: Slide 21
We also have issues of balance, sensitivity, and mutations. What about mutations? Well, when you move these primers in, it's possible that you can have mutations, and you all should be aware of this from the continuing wars between the different types of kits.
This is concordance. If the primers have no problems, then your original primers and your redesigned primers are going to produce the same result. If there is a small point mutation that's not too close to the 3' end, then you get peak and balance effects. If it's very close to the 3' end or it's more than what you would expect, then you can have allele dropout. When you're looking at the effects of these things, try to avoid these situations. McCord: Slides 22 and 23
To find out where these things occurred and to have a better feel for what sorts of things we were going to do, we looked at 541 samples. Using 4 miniplexes, we compared our results from the 310 with results from Identifiler in John's lab at the National Institute of Standards and Technology (NIST). We found 23 discrepant alleles at 7 loci. That seems like a lot, but actually 11 of those, or half or so, were at vWA—most of them associated with allele 14 but that may not have anything to do with it. There was just a binding site issue at this allele dropout but it could be fixed easily by looking at redundant primers. McCord: Slides 24 and 25
We also had some deletions at D13 and some problems with Identifiler. We identified four instances where the problem wasn't in the multiplex but it was actually a binding site mutation in Identifiler.
Sensitivity and peak balance and cycle number are obviously important things. We wanted to establish how sensitive these things were. McCord: Slide 26
Down to 31 picograms, we're still getting nearly a 1,000 RFUs (relative fluorescent units) with mini 2. McCord: Slide 27
With the mini 4, we're still getting up to 500 RFUs at 31 picograms, which is not quite as good. Nonetheless, we get great sensitivity. McCord: Slide 28
Big mini still needs some work. We think we need to increase primer concentration. In fact, we have yet to do an optimization study for primer concentration on these things. McCord: Slide 29
We know that we're going to get better results as we up the concentration of primer. We're already getting great results, so this is kind of exciting. McCord: Slide 30
Another thing we know is that increasing cycle number tends to affect the smaller amplicons more, so we're also wondering about that. McCord: Slide 31
Peak balance is the big issue with this type of analysis. Here is the precautionary tale: If you want to keep the peak balance and avoid these stochastic effects, even though you've got great sensitivity because of your amplification efficiency for the smaller amplicons, you've got to be careful because your effect of template concentration affects peak balance. McCord: Slides 32 and 33
You can be fooled. Although you get an intense result, it doesn't necessarily mean that you aren't seeing some potential for imbalance effects. Around 100 picograms is probably where you start to lose peak balance. If you're going to have some kind of sensitivity threshold for reliability, that might be what you want to do. If you're less concerned about that and just wanted to get a type, then maybe go down, you're still going to get great sensitivity.
Again, we still have some balance problems with D21 and CSF in the big mini. McCord: Slide 34
From there, our next goals are to optimize these things and to start looking at bone samples. We've developed an arrangement with the University of Tennessee Forensic Anthropological Center where we've sent a couple of students to look at the effects of different types of burials and different types of age of decayed samples. McCord: Slide 35
We are seeing some interesting results in terms of the effects of degradation as well as inhibition. We're still trying to deal with inhibition. McCord: Slides 36–41
So what we've got is an alternative to mitochondrial DNA. I'm not saying that it's going to work for every sample for which you would normally use mitochondrial DNA, but I am saying that you ought to look at the assumptions that you have. If the DNA is there and it's too small, well, just make your STR small. McCord: Slide 42
There are some advantages that come along with that. For example, your sensitivity goes up. You should also check for allele dropouts and check for concordance with your other results. It brings up a whole set of different ideas.
In the future we need to continue to optimize our primer sets. We want to look at low copy number. We want to look at redundant primers. We want to increase the primer concentration to optimize our sensitivity. McCord: Slide 43
I have this idea. I don't know whether it's going to work or not, but it seems to me that one of the effects of inhibition may be that the enzyme falls off early. If that's the case, then the effect might be less for smaller amplicons.
Again, in collaboration with the Anthropology Department at Ohio University and the University of Tennessee, we are going to look at degraded samples under some controlled conditions and continue validating. Something I didn't discuss is that we're also working with real-time PCR in terms of getting better ways to quantitate these samples.
Lastly, again I want to acknowledge Kerry Opel and John Butler at NIST; Yin Shen and Jiri Drabek, postdocs at Ohio University; Denise Chung (Ohio University), who did a lot of the degradation studies; Nancy Tatarek (Ohio University); Lee Meadows from the University of Tennessee; Janice Nicklas and Eric Buel for helping us with real-time PCR; and of course, the National Institute of Justice for bringing me here. McCord: Slide 44
Here are some additional slides. McCord: Slides 45–52
Thank you very much.