All right. Thank you, ladies and gentlemen. Please be seated. All right. Dr. Cotton, would you resume the witness stand, please. All right. The record should reflect that we have been rejoined by all the members of our jury panel. Mr. Clarke, you may proceed.
Dr. Cotton, if we can, I would like just to go back to the results from the Bronco, that is, the steering wheel stain, item no. 29. You have described the fact that Mr. Simpson is not excluded or could have been a person who provided that blood stain; is that right?
And let's just start with DQ-Alpha. How do you know that? What about those results let's you know that?
Item 29 has the following alleles: For DQ-Alpha it has a 1.1, a 1.2 and a 4. Mr. Simpson--
If I could, I'm sorry, I'm going to ask you if you would to point to the board as far as those markers that you just described for DQ-Alpha so that we can all see that.
Okay. Mr. Simpson is shown at the very top of the board and he has a 1.1 and a 1.2. The steering wheel, item no. 29 sample, has a 1.1 and a 1.2 and a 4, so two of the--two of the alleles or types that are found on the steering wheel are the same as Mr. Simpson's type, so he would be included based on the DQ-Alpha results as being a possible donor to the DNA taken from item 29.
Without getting into the individual types of the five polymarkers, was the same also true based on your typing results from that particular set of markers?
Now, what about Nicole Brown? You described the fact that she also could be a donor of that stain.
Nicole Brown has a 1.1 type. She has two 1.1 alleles. There is a 1.1 in the steering wheel; and therefore, based on the DQ-Alpha results, she would also be included as a possible donor to that mixture of people.
What statement, if any, can you make that about DQ-Alpha 4 that is also a result obtained from this steering wheel?
The DQ-Alpha 4 is shared with the allele 4 for Mr. Goldman, but his allele 1.3 isn't visible in the DQ-Alpha type from the steering wheel, so that 4 comes from someone, but we don't--we can't make any determination as to who.
As far as the results, and let's take the result you obtained in this case and let's take the steering wheel stain as an example--this is a picture, correct, of what you tested?
As far as when portions of a mixture may have been left on that steering wheel, can you scientifically state that they occurred at the same time?
KEY QUOTEAll right. If I could then, your Honor, I'm going to ask to be marked as People's exhibit 261 a further results board.
Which can be described, I believe, as the Rockingham residence and I believe that is right behind the current board.
Now, Dr. Cotton, referring you to this third results board labeled "Rockingham residence" and in particular to an item marked "no. 7, Rockingham driveway"--
--I believe you even discussed this item a little bit earlier this morning; is that right?
Did you in fact use that particular evidence stain from the Rockingham driveway and type it using the PCR technique?
The results that we got, umm--we got DQ-Alpha types. We got polymarker results. There was no C dot again on this DQ-Alpha type, so although we saw some dots and we recorded those, there was no C dot for the DQ-Alpha. The polymarker results had the appropriate control dot and they were reported.
As far as these DQ-Alpha results on the Rockingham driveway, stain no. 7, did you say that you don't--well, let me rephrase that. You didn't see a C dot; is that right?
Well, it is the same thing that I said earlier this morning. That means the types that you have are types that are in that sample, but you cannot be guaranteed that there are not types that you are not seeing. That is, there may be types in the sample that you are not seeing in your DQ-Alpha results.
Was there anything about the types that you did see that excluded, for instance, Mr. Simpson?
All right. Your Honor, with the Court's permission I'm going to reveal the first magnetic cover.
And as far as these results, Dr. Cotton, can you make any statement about the three parties in this case, whether any of them can be included or excluded from these results?
Based on the polymarker results, the--Mr. Simpson can be included and Nicole Brown and Ronald Goldman are excluded.
KEY QUOTEYour Honor, for the record, it should be noted same objection I made before be a standing objection for similar situations.
Let's turn to what's marked no. 12, the Rockingham foyer stain. Do you recall that?
And you testified yesterday, as well as this morning, about the fact that that stain was tested using the RFLP method?
And then let's turn to the PCR results on that particular stain. Did you obtain some?
The results that we obtained on item 12 indicate that Mr. Simpson is a possible donor and that Nicole Brown and Ronald Goldman are excluded as possible donors.
Your Honor--your Honor, with the Court's permission again may I remove those labels?
Other than the fact that with regard to the Rockingham driveway stain no. 7 and the fact that no C dot was observed, are the types you actually detected the same in both no. 7 and no. 12, the Rockingham foyer stain, as far as the PCR results themselves?
As far as no. 7 is concerned, the driveway stain, did you even attempt RFLP typing?
All right. As far as the particular items in this case, were you provided an unstained control from either of these stains?
And did you test those--I'm sorry, are you able to determine if you obtained controls for each of those items or one item or what?
Umm, probably so, but are they in the set of controls that we obtained at the very end?
Now, taking you back to item no. 7, the Rockingham driveway stain, I believe you said the DNA was too degraded to use RFLP typing?
It means that the DNA had been broken up in a random manner by whatever affects it was exposed to and so it was--the pieces were basically too small to give an RFLP result.
All right. With respect to these substrate controls, again what does that negative result mean to you as an analyst?
It means that in--where--from wherever the substrate control was taken, presumably in a close--in an area close to the stain, that there was no DNA in that area.
All right. Your Honor, I would like to turn to one final results board. I believe that would be People's exhibit 262. I believe it can be described as the results board for the socks.
Dr. Cotton, referring you to this new board that has been marked or will be marked that refers to People's 262, the Rockingham socks, you have already described RFLP testing that was conducted on stain material provided to you; is that correct?
And you described that material or that blood staining as matching Nicole Brown; is that right?
Dr. Cotton, you described the fact that that five-probe match was with Nicole Brown; is that right?
Now, turning your attention, Dr. Cotton, if I can, to PCR testing, was that also done on this DNA this blood-stained material from that sock?
The results from the PCR testing, which is DQ-Alpha and polymarker, include Nicole Brown as being a donor and exclude Mr. Goldman and Mr. Simpson.
KEY QUOTEDoes that mean then, Dr. Cotton, that not only RFLP typing matched Nicole Brown, but so did six additional genetic markers using PCR?
KEY QUOTENow, Dr. Cotton, I would like to turn your attention to and return for a few moments to population frequency data. You use frequency data again in your description of the meaning of results; is that right?
For any set of genetic markers, a particular combination might be more common or less common depending on the alleles, because not--some alleles are common and found in lots of people and some alleles are not very common and not found in very many people, so it is a way to say this is a common type or a rare type that we have identified in a particular sample.
Now, you described a couple days ago about determining this estimate, this means of providing some description about how rare or common types are that may be shared as being able to be used when more than one genetic marker is concerned; is that right?
Essentially you are saying how often do I see the set of characteristics in the first genetic marker I look at and then how often do I see the set of characteristics in the second genetic marker I've looked at. So the overall combination is that how often would I see the combination genetic characteristics from the first marker and the second marker would be the product of those two numbers, that is, you would multiply them together and that would give you the overall estimation of how often you would see the group of characteristics from two markers. If you added an additional marker, then you would multiply that in as well and so on.
As long as the markers are inherited independently and are statistically independent of each other in their inher--that is, they need to be on separate chromosomes, but also people would ask are they statistically independent? As long as they are independent of one another, then the appropriate mathematical calculation is to multiply those individual frequencies of occurrence together to get the overall frequency of occurrence.
Now, with regard to this, and I believe you said different chromosomes; is that right?
Do you in fact use--and let's look at RFLP typing first. Do you use probes that have some different chromosomes or the same chromosomes?
Of the five probes that we used, they are all on different chromosomes, with one exception, that is, two are on the same, although they are located at a considerable distance apart, which means that they also would be inherited independently, and the remainder are on four additional and different chromosomes.
That was going to be my next question. How can you use, for instance, these two probes or two genetic markers that you said were on the same chromosome?
It is--the long explanation, which I won't try to give you, has to do with how DNA is distributed when eggs and sperm are created, that is, because they each have half. When two markers are relatively far apart on a chromosome, the chances of them being passed on to an offspring are equal. That is, it would be--they wouldn't tend to be passed along together and if something--I'm not doing this part very well. If two markers are close together on a chromosome, when that DNA goes into an egg or a sperm, they tend to go together. There is exchange of DNA between the chromosomes while eggs and sperm are being created, so that if two markers are far apart, they are not any more likely to go into the same egg or sperm as two markers that are on different chromosomes. That is the best I can do with--without giving a long biology lesson.
All right. With regard to these two markers that are on the same chromosome, have you done or have you established or shown to your satisfaction or other satisfaction that in fact they are not inherited together?
The--the work was not done at Cellmark, the work was originally done by Alec Jeffries, and he established that they are in fact sufficiently far apart to be considered to be inherited independently.
Now, you have described this multiplication process, that is, the frequency at one marker, the frequency at the next marker and so forth. Is that any different than what has been used in these conventional serology cases about which we spoke for a number of years?
Now, as far as calculating frequencies, did you calculate frequencies in this case based on your laboratory's results?
And was that for the various samples that you have described over the last two days?
It is basically a sample of a population of people from which you will derive your estimation of how often you would see a particular genetic characteristic.
To ultimately assign a frequency to the genetic characteristics in a particular test and then assign an overall frequency for some combination of characteristics.
As far as these RFLP results in this case, and I believe you have described samples that match, for instance, Mr. Simpson; is that right?
Do you compare those set of characteristics or have you compared those set of characteristics to the number of individuals in your databases?
Now, as far as your calculation--actually, let me ask another question. As far as your calculation of frequencies in this case, are those simply estimations of how rare these matching characteristics are?
What role, and I think you touched on this the day before yesterday, what role do major racial groups have in this frequency calculation process?
For some genetic characteristics they can be more common--let me give a concrete example, maybe it will be a little bit easier. Let's just take DQ-Alpha, for example. And I'm making up these figures, since I don't remember the real ones. Let's say the type 1.1, 1.1. It could be very common in Caucasians and very rare in Hispanics or it could be very common in Caucasians and very rare in blacks, or the reverse, and so in order to give an appropriate range of how common or rare a set of characteristics is, it is usual to say this is the--this is the figure for how often you would see this in Caucasians, this is the figure for how often you would see this in African Americans and this is the figure how often you would see this in Hispanics. And if you had other racial databases at your disposal, you could go and do that for other groups as well.
Why don't you simply calculate an estimate of the frequency based on one group as opposed to another?
Well, you could. It would sort of give you an overall kind of averaged figure, but it wouldn't give you the range of figures that using individual racial groups would, and therefore using individual racial groups is generally considered to be a more informative way of giving that number.
In the way of calculating these estimates, do you take any steps along the way--well, do you have to make decisions about kind of which way to do something when you calculate these frequencies? Does that make any sense?
Okay. Let me try that again then. Do you take any steps along the way to ensure that you are not making something sound, as far as a match, rarer than it really is?
The particular steps that are taken--let me just refer to RFLP, because this is really where that becomes--that question becomes applicable. If you remember a couple days ago, I guess, I talked about--we sort of drew out a small database of five people and talked about looking at a window of sizes around a particular DNA fragment size, and it is that window that you are using to reflect the accuracy of your gel system and to not overestimate how rare a particular characteristic is.
As far as this calculation process itself, do you go through it, whether it is an RFLP type test or a PCR type test that is used?
Are there different--I think you said that there were a little bit different considerations in each of the two; is that right?
In the PCR test, when you have, say, a 1.1 allele, you don't have any particular window you need to put around that. That is clearly defined as a 1.1; it is not close to a 1.1. It doesn't have the same features of creating a DNA fragment size where you know that that fragment size is also an estimation. So you have a 1.1 and you don't need to build a window around that, so you can go directly to your database and use the frequencies in the databases that you have available to you.
Now, as far as these various samples, do you calculate then this estimate or did you calculate this estimate based on different major racial groups?
We did the RFLP and the PCR estimates for African Americans, Caucasians and Hispanics.
Those--those are the groups that we have data on at Cellmark, and in addition, the PCR data, some of that population data was compiled from other laboratories as well. The only other major racial group that we might use in the United States might be oriental, which would include Japanese and Chinese and Korean and so on. We do not at Cellmark have a population database that is currently in use for that racial group, so we aren't providing any statistical estimates for that group.
As far as these racial groups, and you have described that you have reported, for instance, in this case, results from three major groups?
What about the existence of other population groups around the world? There is more than three, obviously, and you have named a fourth.
I think that is a question, but I better make sure it is a question. You just named a fourth group, for instance, that you don't report results for?
And I believe you described the fact that that is because you don't have database material; is that right?
How do you know, when you report frequencies from these three groups, that you did perform that estimate process in this case, that somehow you are not overstating the rarity dramatically or significantly when other population groups may have in fact a more common set of these characteristics?
Besides the population data that is available--available in our laboratory, there is an enormous amount of population data that has been produced by forensic laboratories all over the United States and all over Europe and a few laboratories elsewhere, including Australia and Japan. The system that we use that I discussed for this testing is not--I can't take our population data at Cellmark and compare it to population data that has been produced by the Federal Bureau of Investigation or the California Department of Justice, because we used a different restriction enzyme, and if you will just accept for the moment that that makes the DNA sizes a little bit different. However, the European community uses the same enzyme that we do, so there has been a lot of work mostly spearheaded through the Federal Bureau of Investigation to get all of the population data that has been obtained all together in one major volume. And when you look at all of this data together, what you see is that there are some differences in the populations, but they are not huge for RFLP loci, for those loci. This is not applicable--what I'm saying is not correct about the PCR loci. So I'm just talking about RFLP, so that you could go on and do more racial groups, but they would continue to produce numbers in the same range that you are getting with three or four, so it would be a more refined estimate, but it wouldn't really provide you with a huge amount of information that you aren't already getting with three.
Now, I believe you said that it does not apply to the PCR groups. What do you mean?
The PCR locations that we've tested here do differ significantly from one racial group to the next, so if you did more than three groups, you might still see some differences, that is, you might get something that is more common than the three examples that we've provided or less common than the three examples we've provided. You could do that. We don't have that data at Cellmark. The data is available in the literature, but for convenience purposes we still in our reports for PCR provide the three groups that I've already talked about.
Now, as far as your reporting in this case, did you report estimated frequencies, in other words, estimates of how rare match characteristics are in the three major racial groups you have described?
Now, as far as those estimates, can you or will you be able to tell us basically the range, that is, among those three groups as far as a set of characteristics to a sample, let's say, which is the most common group, that is the group with the most common estimate, versus the group with the least common or rarest estimate?
Would that then by doing so represent the range of rarity of matching characteristics?
Now, your Honor, at this time I would like to return to the Bundy crime scene results board, exhibit 259.
Based on the polymarker results, the--Mr. Simpson can be included and Nicole Brown and Ronald Goldman are excluded.
The results from the PCR testing, which is DQ-Alpha and polymarker, include Nicole Brown as being a donor and exclude Mr. Goldman and Mr. Simpson.
Does that mean then, Dr. Cotton, that not only RFLP typing matched Nicole Brown, but so did six additional genetic markers using PCR? It certainly does.
As far as when portions of a mixture may have been left on that steering wheel, can you scientifically state that they occurred at the same time?
The stain was too degraded--the DNA was too degraded.