Thank you, ladies and gentlemen. Please be seated. All right. Dr. Weir, would you resume the witness stand, please. All right. Mr. Clarke.
Dr. Weir, while we're on item no. 78, if I could ask you just one or two more questions.
There were also estimates made by you of frequencies for this mixture based on the RFLP results; is that right?
And for two contributors, they range from 1 in 300 million to 1 in one trillion approximately; is that right?
And for three contributors, from approximately 1 in one million to 1 in 300 million; is that right?
The fact that you made changes or recalculations, as to the PCR estimated frequencies, do those have any effect whatsoever on the RFLP estimates?
Now, I'd like to refer you, if I could, to one of the Defense exhibits, and I believe it's exhibit 1208, the drawing that Mr. Neufeld did about mixture frequencies.
Dr. Weir, you expressed a desire to explain one of your answers to Mr. Neufeld. What did you want to explain?
Well, I wanted to explain how to interpret a mixture. There are some mixed stains, and we would like to assign a frequency to those mixtures on the assumption they came from some people unknown to us. So we need to calculate the frequency of two or three or four people, however many people we think are likely to have contributed to the mixture. And if we don't know, we'll just give the range of answers. And I've been trying to think of an appropriate analogy, and I was thinking of maybe playing a slot machine where in order to win, we had to have three different fruits showing, a lemon, a cherry and an orange. And we will win only if the whole three different fruits show. It just has three wheels on it. So what I've calculated is the frequency of getting that set of three fruits showing, a lemon, an orange and a cherry.
So in your example, there's only a payoff if a lemon, a cherry and an orange show up, one on each of the wheels?
That's right. The whole three have to be there. It doesn't matter on which wheel, but we've got to get the whole three. Now, before we even walk into the casino, we know it's going to be a rare event. That's the way things are stacked. Anyway, this is what we have to get to--we need to get the whole three. What Mr. Neufeld had me calculate was the frequency of getting some of them, either two of them or one of them. It depended on the circumstances. Two out of three doesn't pay. One out of three doesn't pay. I don't care how often an orange shows up. I've got to get an orange, lemon and cherry to pay off.
Now, how does that relate to--and there's this third column on these mixtures frequencies. First of all, the first two columns relate to frequencies that you estimated last week and then testified to today; is that right?
Those first two columns, one of them is much better than the other, they give us the frequency of the set of three; in this case, set of DNA bands, and my example, the set of three fruits. What Mr. Neufeld's--what the NRC's number is the chance that any of the three show up. Now, supposing there were only three fruit on the wheel. All three are going to--and the mixture is the three-fruit mixture, the chance of getting any one of the three is a hundred percent because those are the only three possible. So in some of his examples--in the polymarker examples, he said it wouldn't exclude anybody. Well, that's right. If you spin the wheels, you're going to get something showing up. So nothing is excluded. If there are more than three fruits on the wheel, there's also bananas which we don't see, then the frequencies won't be a hundred. But the essential thing is, this is a mixture. I think everyone has agreed it's a mixture because there's more than two bands that are probed that was contributed by more than one person. The question is, how unusual is this mixed profile to have come about if there were more than one contributor. It doesn't matter how many times the cherry shows up. It's the whole three needed for payoff.
Then what's wrong with this column labeled "NRC" where these numbers 1 in 4, 1 out of 10 or 1 out of 2 are written?
It doesn't give us the frequency of the evidence. The frequency of the evidence, if there were two people unknown to us or if there were three or if there were four, that's what we have to calculate. It's the frequency of the evidence.
And in your view, is that third column a scientifically appropriate way of describing the relevant frequencies of these mixtures?
Incidentally, is there any relation between proficiency tests and whether a mistake was made in this case?
Do either proficiency tests or whether a mistake was made in this case affect your calculations at all?
Now, you described, Dr. Weir, the fact that--and you expressed some reservations about being a witness for a criminal defendant. Do you recall that?
As far as your testifying, is that because your university has a contract with one of the laboratories in this case?
I'm very concerned that the statistical interpretation of DNA evidence is done correctly. I've been very angry over the last four years at some of the statements made in courts about the statistics attached to DNA profiles. I am very anxious that they be done correctly and I am therefore willing to testify when possible.
KEY QUOTEIf a criminal defendant wanted you to present evidence about statistics or frequencies attached to either a match or a mixture using the proper scientific manner of doing so, would you so testify for that Defendant?
As far as the university is concerned, what is your status while you're out here in Los Angeles testifying?
Now, these numbers that are presented on these boards--and we have before you and the jury exhibit 259, the Bundy results boards. But let's take all five of the results boards into account. What are these numbers that are written up there? Are they exact numbers?
Oh, no. No. They're estimates of the frequency with which these various profiles would arise from unknown people.
Well, there's two parts to that. They convey a very real sense how rare these profiles are. I think they're very good at doing that. But specifically, numbers in the billions are a little hard to understand. We don't have experience in having to deal with numbers in the billions. The only time we see them written down anywhere is to do with the national debt. I don't think anybody understands that. We just have these enormous numbers. They're good at saying this profile or that profile is rare, and I think when they are so rare, they're based on so many loci, I would be happy to just leave it at that.
I would like to direct your attention or I would like to have you direct your attention to a few of the stains on this board, in particular, item no. 47.
And would it help you first of all to use the smaller charts that you have in front of you?
I would like to direct your attention to items 47, 48, 49, 50, 52. First of all, amongst that group, that includes PCR results, and as to 52, RFLP results as well; is that right?
Your Honor, objection. All of this is beyond the scope. None of this goes to cross.
From a population genetics standpoint as well as a statistical standpoint, what do those results tell us in terms of the evidence presented on this particular board?
As far as those results are concerned, do those results--I'm sorry--results reveal a rare profile?
Yes. But--well, especially the RFLP frequencies. When they're in the millions, that certainly indicates a rare profile.
And as far as RFLP results, are you referring now to what would be one of the Bundy walkway stains, item no. 52, as well as the rear gate, item no. 117?
What can you tell us, Dr. Weir, about the meaning of those pieces of evidence referring to the RFLP matches from a statistical and population genetics standpoint?
Well, the statistics and population genetics is concerned only with determining how frequent that profile is in the general population and people other than--in some unknown person. So we--the numbers represent that frequency.
As far as Mr. Simpson is concerned, is he an individual that's simply not excluded or is this evidence from a frequency standpoint something more powerful than that?
Well, the evidence is that it's very unlikely we would see that evidentiary profile if it came from somebody else.
KEY QUOTEYour Honor, at this time, I would like to put on the easel the Rockingham residence results board.
I'm very concerned that the statistical interpretation of DNA evidence is done correctly. I've been very angry over the last four years at some of the statements made in courts about the statistics attached to DNA profiles.
Two out of three doesn't pay. One out of three doesn't pay. I don't care how often an orange shows up. I've got to get an orange, lemon and cherry to pay off.
I'm not being paid for any aspect of this case.
I'm on vacation.
The evidence is that it's very unlikely we would see that evidentiary profile if it came from somebody else.