Terrance Gainer (left) and Dr. James Crow (right) CHIEF JUSTICE ABRAHAMSON: Okay. With that we turn to Dr. Crow for a report on the Research and Development Working Group. Jim? COMMISSIONER CROW: Fine. Let me start by saying that when we talked about 50 years in the future, I decided to look 50 years in the past. And if we go back 50 years, which I do, and many of you do, at that time there were 48 chromosomes that had not been counted correctly. Seems crazy now. Actually, it was more like counting strands of spaghetti than counting discrete objects. But also we didn't know that the gene was DNA. The prevailing opinion was that the gene was protein. It was in 1953 that it finally, finally got straightened out to most people's satisfaction. And it wasn't until 1985 that this very deep scientific discovery was made practical from the standpoint of forensic use, had quite a long time delay but once it got started it increased exponentially. All of this is to say that any predictions for 50 years in the future are totally useless and predictions for ten years in the future are troublesome. The job of our subgroup has been prediction and not prescription. So, we have tried to make predictions as to what's happening within the next ten years but they're based mostly on the kind of things that are available now and are likely to be refined in the future. One thing I think is pretty clear, and that is that the 13 core loci are here to stay for a while. They will be superceded by better things but that doesn't mean that they will be replaced because the laboratories will have tooled up to use these 13 loci and they don't want to change even to something that's better. They don't want to change from something that works very well just to get something that's better.

So I think we can expect, at least we say so as a group, that we can expect the STR 13 core loci to be around for the next ten years, which is our period of time to look at this.

There are essentially two kinds of technical improvements worth talking about, there being two kinds. One is those that will simply make the 13 loci easier to use, more effective to use, more rapid, more automated, more robotics. Those things are almost certainly going to come within the next ten years but they'll use the same 13 core loci.

There are also going to be new systems that come and they're already here in many cases. The, I'm thinking about the same kinds of things that could be used, additional STRs, I think we can expect that pentanucleotides will come into the picture more than in the past.

They have two kinds of advantages. One is that there are a larger number of alleles per locus and the second is that the people who work in this field tell me that they're less prone to stuttering than the repeat of shorter lengths. So I think we can predict there will be more pentanucleotides.

COMMISSIONER SCHECK: Could you define what those are a little bit?

COMMISSIONER CROW: Yes. They're repeats of five units. Most of the ones that are now used are repetitions of four units and the fives just turn out to be a little bit better. And if they had been discovered first, they would probably be the ones that would be in use. I think there are, I think in two years those would be added to a lot of the kits that are available.

There is also, the one that Promega puts out right now has one pentanucleotide in it along with the 13 core loci and a sex-determining locus. What's coming up already here, we'll certainly have more emphasis on the Y chromosome and the Y chromosome will use SNPS, which is, which means single nucleotide probes or polymorphisms. That, the idea is not new, the DQ-alpha is already such a thing. But the widespread use of these is new and will undoubtedly increase in the future.

The last time I looked at this, which is a few weeks ago, there were 150 of these things on the Y chromosome. So that means that the Y chromosome is going to become just as useful as any other chromosome and it will have specific uses, especially with mixed samples from rapes. Having a Y chromosome arm will make it much easier to distinguish how many different males contributed to this.

Mitochondrial DNA is here already. We heard from Dennis Bauer. He had his first case based on mitochondrial DNA this past week. One, one disadvantage of both the Y chromosome and mitochondria is that these were essentially single unit things. There's no exchange of parts, no scrambling that goes on. And that means that the amount of information you have is simply restricted by the size of the database.

And so there will be pressure, and I think it will happen, to increase database sizes for mitochondria and the Y chromosome to the extent that it's used. Mitochondria has the advantage that it traces the female ancestry, so that means you can't always tell people apart but you can tell whether they are descended from the same female.

The Y chromosome has the same both important attribute and deficiency in that it traces male ancestry. We've seen, the newspapers have had quite a number of interesting examples of what has been transmitted by way of the Y chromosome.

We can certainly, I think these new markers will mainly be SNPs and they can be and can be and will be popularized a great deal in the next few years. It's already happening. Most people who are involved in the research in this area as opposed to forensic practice are using SNPs all the time and they'll clearly come out in medical research. The genome project, which is due -- well, the year in which it's due changes every year. I think it's clear that the reason that the target date changes is not because the science changes, it's because the definition of completeness changes, that is, relaxed a little bit from time to time.

So we're going to know, it's a very large substantial fraction of human DNA very soon and this will certainly pick up genes. The latest estimate I saw of the number of genes was 140,000. When I was starting to study genetics, we estimated it at 10,000 and the number keeps going up and up. I say that partly to emphasize the point that we really don't know and it's a guess, guess right now.

There will be much more integration of computers into the analysis, even from mixed samples, which are troublesome things to do statistically by hand, at least, and computer programs can do this more effectively.

Robotics will come into the act sooner or later and I think probably sooner. I think you can also expect to find much more studies of nonhuman DNA plants and animals, will enter into this, and also within the human DNA, we have DNA that not only comes from our genes but we have the DNA that comes from viruses that we carry so it would be possible, could be right now for some cases to identify quite a bit about a person in terms of what particular disease it carries, and of course, their normal viruses as well as the abnormal ones.

Mostly this sounds like a gee-whiz science fiction stuff right now but my conviction and a very strong one is that we have always been surprised at the past. When new developments start developing, we say it's going to happen in ten years, it turns out to happen in five years.

And there are a lot of things right now that could be done very soon to the satisfaction of a hospital laboratory but, of course, we demand much more higher, much higher standards for forensic use.

So, it's a question of not predicting when something is theoretically possible but predicting when it's developed to the point that it would be acceptable by the forensic community. I don't want to talk too long but I want to say a few more things.

CHIEF JUSTICE ABRAHAMSON: Go ahead. No problem.

COMMISSIONER CROW: So far no problem but I'm not really wound up yet, for sure.

CHIEF JUSTICE ABRAHAMSON: Tell it to me on the plane going home.

COMMISSIONER CROW: One thing that's happening and will continue to happen is the gathering of population data from more and more different populations. American Indians, for example, have been a problem. Their tribal structure means that there are many different subsets of the population compared with the larger, with the larger groups.

One point that our committee is interested in and I personally am strongly interested in is this. Whenever we make a calculation for the probability of a match or a likely ratio, that's based on some assumptions. And we think they're pretty good assumptions but they're far from really established to the point that you would like to have it. It's clearly not true that the United States population is one panmictic, random unit.

So, using just strictly Harvey Weinberg ratios, clearly can't be exactly correct although it could be a remarkably close approximation, likewise the achievement of what we call the linkage equilibrium, equilibrium between genes and independent chromosomes, that's, supposing somebody asked about the topic, one would expect it not always to be correct even for genes that are on independent chromosomes.

This is all a prelude to saying that I think that our methods that are being used right now are really quite good but they're not as good as you would like. What can we do to make it better? Here's one idea that I think may catch on.

If we were worried about the fact that the population of people that we're looking at contain relatives unknown to us who is looking at them or that they come from a small subset of a population, which means they're partly related. Essentially I'm saying the same thing. If we don't know that, is there something we can do about it?

Well, one way is this. We can increase our standards slightly. It only takes a few more loci to do this, and make the process sufficient to distinguish among, between sibs, between brothers.

One thing that I think is obvious but I'll try to explain it without waving my hands too much but if you have two parents and two brothers that are children of those, those, those brothers are going to share one-fourth of their genes irrespective of anything else that's happening, just by the laws of Mendel.

So it means that there's a factor of one-fourth that comes into the resemblance of brothers; that's not true for any other degree of relationship.

Now, this one-fourth is augmented by smaller numbers that depend on the frequency of the genes in the population but this means that for the most part this factor of one-fourth slightly augmented is quite robust against most of the things that we would worry about in the assumptions that I was mentioning earlier. If we could separate brothers, we could certainly separate anything that's less related than brothers.

So, how much, how much more would we need? Oh, one other point about that. The 1992 committee, though, was strongly criticized for its ceiling (phonetic) principle, and I think it was properly criticized, but the aim of the ceiling principle, and it is a very laudable one, and that was to have a procedure that made it unnecessary to distinguish among racial or individual ethnic groups. By the use of sibs we come pretty close to accomplishing that.

The value of this system depends mainly on the loci that you're using, having a large number of alleles. So that's part of the explanation for my personal enthusiasm for pentanucleotides to be added to the list. Right now with 13 core loci we can separate sibs from each other. The probability of two sibs matching is about one in 300,000.

What I'm trying to say here is that probably many courts would be more convinced by a robust, certain kind of an explanation, free of assumptions, of one in 300,000 than they would be one of 300 million when you're not quite so sure of your assumptions. So, to me this can be an important step.

If we want to go up to the kind of levels that are traditional probability levels, the VNTRs (phonetic) and then STRs, if we go up to 21 loci, we're down to an average match probably of one in 300 million, which ought to satisfy almost anybody.

So I think, I think our committee believes and I certainly do personally that I'm willing to trade a small probability for a higher probability that's made with much more certainty. So, I really hope this will catch on. And it's essentially a part of our report.

One other point that comes up, and the FBI has gotten into this, and that's the question of defining what individualization or uniqueness means. And what the FBI did was say they could take, we asked for a probability that has the following properties. We take the population of the United States or rather its reciprocal and they would say if two people have a probability less than the reciprocal of the population of the United States, they're probably the same person. Then they introduced a number of factors of safety.

They put in an arbitrary factor of ten, which came from our 1996 committee report and they introduced confidence limits, a fairly elaborate calculation but simple in principle. Whether that will catch on I don't know; we don't know. It has been used a few times. It's also been strongly criticized because it's based on the more naive assumptions of strictly Harvey Weinberg and linkage equilibrium proportions and composition of a genous structure. And whether these conservative modifications that were built into that take care of all of the doubts that a person can have, that's an open question and I think that it will remain.

My guess -- this can only be a guess -- is that, of course, this is not really a scientific question, and maybe the courts or some other influential group will decide that if the match probability is much less than the reciprocal of the United States population, it will simply declare that this is unique and make it a legal or a formal definition that hasn't much scientific basis but can act, but can be a basis of procedure. It's not the first time in the history of society that things like this have been done.

A few other points. Right now it's already come up around the table. We could take a drop of blood or a sample of blood and oftentimes determine the geographical origin of the person.

We can even do it from the 13 core loci with certain probabilities, not too impressive in most cases but with the addition of the kind of things that could be looked at in the breadth (phonetic) sample and one could. There are, there are genes that have a very high frequency in persons of African descent and very low otherwise and vice versa.

So, the recognition of the geographical ancestry of a person is almost certain to come with increasing knowledge. To my taste much more interesting, and much more acceptable, really, would be of the finding of individual traits.

I don't know what the state of the art is now but it must surely be changing rapidly and the genome project is going to produce this willy-nilly, a gene for eye color, for hair color, for baldness, skin pigment, the kind of things we obviously think of are not very far in the future as far as being discernible from a blood sample. But I would emphasize -- and I think everybody in the room knows but let me emphasize anyhow -- that, that in order to do that, you're looking for traits that are not part of the 13 core loci.

So if you try to do something like this with a database, you would have to do it with probes other than the kinds that are now, that are now being used.

I think we will also begin to learn things that are based not just on the presence or absence of a DNA type but on the amount of expression of the genes produced by that DNA. That's a field that's advancing very rapidly right now, just the study of the gene expression. And we're going to find that it's possible to distinguish people on the basis not just what genes they have but how strongly this gene exerts its process.

The view of our, of this subgroup on sample retention, it certainly was not, certainly we did not have the kind of detailed discussion that's been going around the table here. And perhaps a discussion was not needed but the viewpoint of the group is that samples should be retained, there are enough things that you can do with this that are worth preserving and that the best way to avoid abuse is by making abuses illegal rather than trying to make them impossible by withdrawal of the samples. Whether the -- let me state this. Whether our Working Group would have reached this same conclusion had we all participated in the discussion today, I'm not at all sure of.

One other thing, two other things, then I'll end. The, we are impressed by the tremendous amount of attention that's been given to technical and statistical aspects of DNA forensics, that exceeds virtually anything that's happened in the crime scene up until now, the identification scene.

Even fingerprints have not been subject to the kind of statistical rigor although I don't think that we seriously doubt them right now. But I'm wondering if it isn't possible that the kind of really meticulous care the DNA has had will not have a salutary influence on other sources, ballistics, handwriting, all the other things that are considerably more dubious but which could be made more precise by the same kind of attention to statistical minutia that the DNA has had.

The group also has talked some -- we're getting out of our area pretty quickly -- that, about witnesses. There's a feeling within the group that there's a real place for expert witnesses not brought in by either the prosecution or the defense so that a person would be free as a witness to give evidence on either side of the issue or both sides of the issue, I probably should say, rather than being more or less constrained, as the adversary system calls for, for presenting just one point of view. We commend this to the Commission as a possible study but you're doing it already. So that's just one more piece of advice to people. It's already happening.

What about universal databases? There's no strong advocacy of it anywhere but I will point out to all of you here that the country of Iceland is now in the process of getting everybody DNA typed and whatever the results of that are, it's a place for us to look at and see what that's led to, what have been the social problems, what have been the gains.

There are certainly going to be some goods coming from it; almost certainly it will create problems and I, for one, will watch that with great curiosity.

As far as our report itself, we have it, after this read, written, it's undergoing quite a number of modifications, most of which are stylistic, though, not content-wise, and I hope in the next month or two that we will have approved by e-mail, mainly, a final working that can be turned over to the editors and then finally subject to all your approval.

COMMISSIONER SCHECK: Would it be possible to send e-mails if it's substantively in place but not stylistically? I would love to see an e-mail of it. Is that possible?

COMMISSIONER CROW: You know, what I can do for anybody I think is simply send the latest version of the report. That's probably better.

COMMISSIONER GAHN: Yeah, that would be great.

COMMISSIONER BASHINSKI: That would be great.

COMMISSIONER REILLY: Could it just be uniformly distributed?

DIRECTOR ASPLEN: Sure.

COMMISSIONER GAHN: Realizing it's a draft but --

COMMISSIONER CROW: Yeah. Well, since I'm trying right now to do some revisions, the next week or two is tied up for me but within a month or so I think I could circulate the next edition of the draft and I think we should do that.

DIRECTOR ASPLEN: Sure.

COMMISSIONER CROW: We clearly welcome the input of all of you or anyone else, for that matter.

CHIEF JUSTICE ABRAHAMSON: Any comments, questions? One, circulating the draft will be good, familiarize everybody. Phil?

COMMISSIONER REILLY: Just a footnote. I happen to be fairly involved with what's going on in Iceland. So I have the statute. I have been in touch with the people and talked with them and I've continued to paragraph some for them.

COMMISSIONER CROW: Well, who did you write it out for? I heard a lecture by the man who --

COMMISSIONER REILLY: Kerry Sevrenson (phonetic)?

COMMISSIONER CROW: Yes.

COMMISSIONER REILLY: Yeah, I was with him just a few days ago, so.

CHIEF JUSTICE ABRAHAMSON: The next meeting will be in Iceland.

COMMISSIONER REILLY: You could do worse.

CHIEF JUSTICE ABRAHAMSON: I know. It will be fun.

COMMISSIONER SCHECK: They have very good records, too, right? Excellent records.

COMMISSIONER CROW: No, the Iceland people have been much more zealous in maintaining their, or assiduous in maintaining their genealogical records than most of the rest the world has. So you can combine the modern DNA technology with unusually good record keeping.

COMMISSIONER SCHECK: Aren't they being paid for it, though?

COMMISSIONER REILLY: Well, we could get in a long discussion about it but basically there's a, it's a major source of income to the government of Iceland, is over a hundred million dollars.

CHIEF JUSTICE ABRAHAMSON: Who provided the money?

COMMISSIONER REILLY: Roche.

CHIEF JUSTICE ABRAHAMSON: Roche, okay. Well, that's not bad. It beats taxation, said facetiously.

COMMISSIONER REILLY: Interestingly, the law does not compel the taking of DNA. The law basically licenses out the health records and then each individual must consent to have his DNA drawn. But I just asked Kerry this. About thus far 99 percent of the citizens approached are giving their DNA on the spot.

CHIEF JUSTICE ABRAHAMSON: Norm?

COMMISSIONER GAHN: One question, Doctor, just from what you said. Do you believe that there's a possibility that down the road there may be a turn from the, to these traditional noncoding genetic locations you look at and a turn to using these medical diagnostic or viral DNA or phenotypic expressions in the forensic setting, be going that way?

COMMISSIONER CROW: Oh, I think we're going that way, yes, willy-nilly. Of course, I still think the main identification will be these 13 core loci and modifications therefrom.

COMMISSIONER GAHN: But hasn't one of the selling points all along been we're looking at noncoding areas --

COMMISSIONER CROW: Yes.

COMMISSIONER GAHN: Then why go and get worked up about it?

COMMISSIONER CROW: Well, I emphasize it over and over again, that these loci are

deliberately chosen not to be associated with those traits.

COMMISSIONER SCHECK: Is that true, Norm, the STR?

COMMISSIONER CROW: It's not absolutely true.

COMMISSIONER SCHECK: It's not true, is it?

COMMISSIONER CROW: There's one disease locust, probably you know about it but the VWA is really, it's part of a gene. It's Von Hildebrand's (phonetic) blood syndrome, coagulation syndrome. This is not the business part of the gene but the theoretical possibility of finding a believer of this that has the correlation with the disease state exists. Not A very useful correlation, I suspect.

COMMISSIONER SCHECK: So, DQ-alpha is definitely?

COMMISSIONER CROW: DQ-alpha, very much, oh, yeah. So although this is an ideal, that's not absolutely achieved. That may be one reason why maybe you might want to replace one locus of the 13. There are some others, though, that are not so satisfactory for other reasons.

COMMISSIONER SCHECK: One of the, following up on Norm's comment, just to make it clear, would it be fair to say that if you were to look at the science profiling, which is, for lack of a better phrase, au courant, you, I mean, you know, is a booming industry in law enforcement, I'm talking here about looking at the modus operandi or the characteristics of a crime or crime scene and trying to make a prediction of a profile of who the person is.

I mean, what you're talking about here is that, that we can use these genetic techniques to build, to create a form of profiling in an investigative sense from crime scene evidence.

COMMISSIONER CROW: I think so. Of course we're talking about the future and what's here now.

COMMISSIONER SCHECK: Yeah. Part of it is here in terms of race and disease states and, you know, obviously where we're going, I mean, there's physical disease states, private medical facts as well as psychological profiling that could arguably be done from this.

COMMISSIONER CROW: Well, it's troublesome but it's a real possibility. Presumably it would be used to narrow the search for suspects.

COMMISSIONER SMITH: Well, actually you'll hunt wiser.

COMMISSIONER CROW: Yeah.

Previous	Contents	Next
Back to National Commission Main Page