Now, sir, just to remind you of what's up on the screen at this point, we're looking at the genotypes for the Bronco console stains, 303, 304 and 305.
Okay? And the DQ-Alpha profile that's seen in that mixture is a 1.1, a 1.3 and a 4 and there may or may not be a 1.2 there; is that correct?
Okay. Now, do the frequencies and the sums of the frequencies for the Caucasian, African American, Hispanic databases that I present on that table as you're looking at now correctly give you the aggregate, the sum for the various matching genotypes that are consistent with that mixture for an individual?
Well, it's--correct me if I'm wrong. It looks like you've listed all the possible genotypes because every genotype could be in that mixture.
So I'm puzzled as to why the frequencies don't add up to a hundred percent. I must be missing something.
Okay. Because there are some alleles in the system, if you may recall from your review of how the system functions, like allele 2, allele 3. Do you remember that, sir?
Do you recall in your discussion of the systems with Mr. Sims that there are more alleles in the system than alleles that are present in this mixture?
Correct. All right. So when you--when you then look at the frequencies reflected in the Cellmark tables for the genotypes that are present in that mixture, are the numbers reflected on Defendant's--
And in 1210, we were told that there are three possible alleles present in the D1S80 system; is that correct?
Okay. And the various genotypes that could comprise those three alleles, 18, 24 and 25, are reflected in the first column; is that right?
And if we were to sum up all the possible genotype frequencies that could be part of that mixture, you would arrive at the numbers that we have on that last line entitled "Sum" for the three different ratio databases; is that correct?
Okay. Now, if you wanted to then look at what is the frequency of the population who could not be excluded as a potential contributor to this mixture both in the D1S80 level and the DQ-Alpha level, you would then have to combine the frequencies from those two different loci; is that correct?
And does Defendant's 1211 now reflect the combined frequencies of the DQ-Alpha system, the D1S80 system for those three databases as I've just described?
Okay. And would you agree, sir, that if we simply look at the most common frequency of people who could be included as contributors to that mixture of those three databases, what is the most common?
Okay. So what we're saying, sir, is that given the mixtures that we see--just looking for the black.
--that if we're just considering 1 percentage of the population cannot be excluded as being a contributor to the mixtures reflected on the bloodstains on the console using that NRC method as you see it up there, would be approximately 1 in 4; is that correct?
And what that says, when you say 1 in 4, that means approximately 25 percent of the population, if you simply type them at random, they would come out with genotypes for these two systems that would be included or consistent with the profile seen in that mixture on the console; is that correct?
Well, included, but not consistent. The profiles are a mixture of more than one person.
I understand. But you wouldn't be able to exclude 25 percent of the population approximately?
Thank you. And now, finally, for that last item, sir, G10, which is the stain on the glove, again, the DQ-Alpha profile for that one, for the DQ-Alpha mixture, would be reflected on Defendant's 1209, would it not, because you have again the 1.1, the 1.3, the 4 and you may or may not have the 1.2?
Fine. And now, for the D1S80 frequencies, however, instead of seeing three alleles present, you only see two; is that correct?
And what you see in the first column of 1212, sir, are the various genotype combinations that could go into that mixture; is that correct?
Now, once again, now for this stain, if you wanted to take the same approach that we did for 303, 304, 305, you would then combine the frequency for the D1S80 locus with the frequency from the DQ-Alpha locus; is that correct?
Okay. And on 1213, sir, have we done just that? Did we combine the frequencies for each of these genotype sums, both for Caucasians, African Americans and Hispanics?
Now, of these three databases, what is the most common frequency of individuals who cannot be excluded as having contributed to a mixture such as this profile?
All right. So if the question was, Dr. Weir, what percentage of the population would be included as potential contributors to this mixture on the glove, that was the question being posed, the answer would be 1 in 10; is that correct?
I think I have some new ones in the red cup down here. Mr. Neufeld, try that red cup there.
And so with regard to that last fraction, 1 in 10, what you're saying is, sir, that if we look at the entire population, that approximately 10 percent of that population would have genotype frequencies which would be includable or represented in that mixture on stain G10; is that correct?
Now, Dr. Weir--Dr. Weir when you failed to take into account in your calculations last week, was it that the--when you see the 1.3 and the 4 alleles in a mixture, that you don't know whether the 1.2 allele is there or not; is that correct?
And I believe you also said that before you would undertake all these statistical calculations, you wanted to be absolutely sure of what these mixtures meant, and that's why you contacted Mr. Sims; is that correct?
Well, sir, do you realize as you sit here today that just as the presence of the 1.3 and 1--I'm sorry--withdrawn. Do you realize as you sit here today, Dr. Weir, that just as when you have the 1.3 and 4 allele, you can't tell whether the 1.2 is there or not, and that same phenomena occurs if you have a mixture in the 1.1 and the 4 allele are present? Are you aware of that, sir?
Dr. Weir, there's already been testimony in this case that Cellmark's no. 19 is their internal number for LAPD item no. 29, the steering wheel. I would like to show you what's been marked as Defendant's 1213--1214 and show you their notes and show you their comments and their characterization of the typing for DQ-Alpha on their no. 19, which is the steering wheel. Here, here, okay, and here (Indicating).
All right. Now, in the Cellmark laboratory notes, sir, for their DQ-Alpha typing of the steering wheel, do not they say that it is a 1.1, a 4 and a possible 1.2?
All right. Before you commenced to do your statistical work on Cellmark's data in this case, did you take it upon yourself to learn how the data should be interpreted?
For the interpretation of the mixtures on this item specifically, I used the chart in evidence. I did not--I have not seen these pages before.
Are you aware of the fact, Dr. Weir, that any time there is a 1.1 and a 4 allele present, the 1.2 might also be massed? Are you aware of that?
Dr. Weir, when you calculated the frequencies for the mixture on the steering wheel, did you assume that the 1.2 allele was not massed?
On both DOJ and Cellmark. I saw the same pattern from both laboratories. I assumed that they were correct and made that calculation.
So are you saying, sir, that when you did your calculations on the steering wheel, you assumed in that calculation that the 1.2 allele was not massed or was actually present?
Okay. And, sir, would you agree if in fact a 1.2 allele could equally be massed whenever the 1.1 and 4 are present, as you've already acknowledged it is when the 1.3 and 4 is present, would that also change your frequency estimates?
If you were to learn, sir, that the--that when you have a 1.1 and a 4 allele present as you do on the steering wheel--
Excuse me, counsel. Why don't you more clearly frame that as a hypothetical question.
Well--all right. Assume for the moment the hypothetical, sir, that whenever you see a 1.1 and a 4 allele in a mixture, you can't be sure whether the 1.2 is there. I just want you to assume that.
Sir, I'm just asking you to assume that. Would you assume that for the next question I'm about to ask you?
Okay. Well, if you assume for the moment that whenever you have the 1.1 allele and the 4 allele present in this hypothetical, that you don't know whether the 1.2 allele is present, in that hypothetical, if you then calculated the frequencies for the mixtures on the steering wheel, would you arrive at frequencies that are more common than those frequencies that you arrived at either in your Friday calculations or in your Monday morning calculations?
The way you've described this hypothetical sounds to me like the situation we've already discussed, the 1.1--excuse me--1.3 and 4.
If that's the facts, and I have no knowledge one way or the other, if that was the case, and that was ignored, then that would affect the results, certainly.
And, sir, would that also affect the results in such a way that your calculations on the steering wheel would again be statistically biased against Mr. Simpson?
Well, would you agree, sir, that if that condition in the hypothetical was correct, then the statistics that you've articulated to this jury both last week and this morning would incorrectly state the frequencies in such a way as to make them seem more rare than they really are?
I want you to take a look at this page. It begins here and ends here, okay. This and then this table (Indicating).
Okay. Now, having read that document, okay, sir, would you agree that apparently, given the construction of this DQ-Alpha test--
Sir, have you ever--before you did any calculations in this case, you said that you spoke to Gary Sims and you attempted to learn the limitations of this system; is that correct?
Yes. I've asked them to confirm that the results on the charts are accurate, and he confirmed that they were. I've not seen this document before.
Sir, plain and simple, wouldn't you agree that you've also made another mistake on your calculations for item 29?
Sir, as you sit here today at this moment on the witness stand, did you make a mistake in your calculations for item 29?
No, I didn't. This is the first time I've seen this document, and it says here apparent heterozygous, type 1.1 and 4. If we see that "Apparent," meaning that's what we see, that apparent type may have a 1.2. The evidence presented to me quite clearly states a 1.1, 1.2, 4 mixture. It does not show a 1.1, 4 mixture.
Dr. Weir, before you did your calculations in this case, had you ever even looked at a DQ-Alpha strip?
I have not looked at a DQ-Alpha strip at all ever. As I've said repeatedly, my analysis starts with what the forensic scientists tell me in general and in particular what's written on these charts.
KEY QUOTEBut, sir, don't you think it's important as a statistician that's relying on someone else's data, that you first learn what the limitations of that system is in terms of ambiguities?
I don't have enough time to learn all of the molecular biology. I must analyze data presented to me. If a reputable forensic scientist says this is the data, that's what I analyze. I can't do everything. I've been very careful I hope to explain I'm not a molecular biologist. I must rely on what's told to me.
KEY QUOTEWell, sir, wouldn't you agree that at least in the hypothetical that I gave, if in fact when you have a 1.1 allele and a 4 allele present in a mixture, if in fact you can't tell whether the 1.2 is there or not, if that was true in the hypothetical, wouldn't that mean that your statistical frequencies for item 29 understate the true frequency of that profile?
The hypothetical as you've described is not consistent with what I have in front of me.
But is--the hypothetical as I've described it to you, sir, would mean that your frequencies once again would be in error, correct?
Well, the hypothetical as I've explained it to you would mean once again that your frequencies understate--
Okay. Well, accepting the hypothetical as I've given to you, wouldn't that mean--I'm sorry--would that mean that the frequencies of occurrence would be more common than those that you testified to previously?
My results, which I believe are correct, are based on the assumptions of the data being as it is. I have been very careful. I've given you the answer for the profile in front of me. That is not the answer for a profile not in front of me.
If in fact the condition I put into the hypothetical is correct, would you agree that the frequencies that you've testified to here for item 29 are in fact less common than what they actually would be?
If I calculated frequencies under one assumption when the other one was true, then I would be misstating them, yes.
KEY QUOTEOkay. Thank you. Now, the methods that you employed in this case in calculating the frequencies of these mixtures, have you ever testified to those methods in any other criminal case involving DNA evidence before this one?
To your knowledge, Dr. Weir, has any other expert in the United States ever used your approach that you used with this jury to calculate the frequencies in a mixture in a forensic DNA case?
I believe Dr. Stoney in Chicago uses these calculations. I don't--I'm not sure whether he's testified. He's certainly written about them. He is the scientist in this country and there are scientists outside of this country who use that method.
Once again, sir, to your knowledge, has anybody, any expert anywhere in the United States of America ever testified in a court of law in a criminal case applying the methods to calculating mixture frequencies that you've testified to in this case?
Your Honor, I'm sorry. I would ask the witness be instructed to answer the question.
And I believe you said that the California Department of Justice relies on the FBI database for its RFLP numbers; is that right?
And you've also been consulting to Cellmark since about the same time on their statistical methods and their databases; is that correct?
And I believe you testified that you reviewed--at the conclusion of your direct testimony, that you reviewed the statistical work done by both DOJ and Cellmark in this case; isn't that right?
And you said after your review of that data, it meets with your approval; is that correct?
I'm not sure if I said those words, but I think the estimates they gave are good estimates. They follow naturally from the databases they use, yes.
Dr. Weir, when you reviewed Cellmark's work in this case, were you as careful in your evaluation and assessment of their work as you were of your own work in this case?
Well, Dr. Weir, were you more careful with your evaluation of the FBI's statistical work and DOJ's statistical work in this case than you were with your own statistical work in this case?
I have not looked at a DQ-Alpha strip at all ever. As I've said repeatedly, my analysis starts with what the forensic scientists tell me in general and in particular what's written on these charts.
I'm not aware of any, sir.
If I calculated frequencies under one assumption when the other one was true, then I would be misstating them, yes.
Oh, I can assume anything. I don't think it's going to impact my analysis.
I don't have enough time to learn all of the molecular biology. I must analyze data presented to me. If a reputable forensic scientist says this is the data, that's what I analyze. I can't do everything.