Good afternoon, ladies and gentlemen.
THE JURY: Good afternoon.
CROSS-EXAMINATION (RESUMED) BY MR. NEUFELD
Now, Dr. Cotton, putting aside the--the absence of external blind proficiency tests at Cellmark, would it be fair to say that you have participated in a number of open proficiency tests since 1988?
And would it be fair to say that in--in these open--open proficiency tests that you have participated in you routinely would receive a few samples to examine or certainly no more than a handful of samples?
That you would receive no more than, let's say, five samples in any proficiency test?
Now, in this case, however, this case meaning People versus Mr. Simpson, you have received how many different samples from the Los Angeles Police Department to do DNA typing on?
All right. Now, unlike those various open tests that you just referred to a moment ago where ordinarily you would have three or four samples given to you to examine, were there two proficiency tests that you were participating in in 1988 and 1989 where you received 50 samples in each?
Okay. And those tests in 1988 and 1999 were sponsored by the California association of crime laboratory directors?
Now, let's take a look first at the proficiency tests run by the California association of crime laboratory directors that you participated in in the year 1988. First of all, in that proficiency test did you have any incorrect exclusions?
Now, in that particular test, Dr. Cotton, you were given these different samples and you were asked to type them and determine which, if any, matched any of the other samples; is that correct, in the same batch?
And at a subsequent point in time were you provided with a legend produced by the California association of crime lab directors telling you which samples in fact matched which other samples?
And when you were provided with that legend, Dr. Cotton, didn't they tell you that items 70, 72 and 76 should all match one another?
That may be right. I don't have my copies with me. Umm, I do actually have them in the building, but I don't have them down here with me, so I might need to look at yours.
Which is a six-page report from the California association of crime laboratory directors.
I show you this exhibit, this five-page--I'm sorry, six-page document from the California association of crime lab directors and ask you if it refreshes your recollection as to whether or not items 70, 72 and 76 all come from the same source?
Okay. And now that it refreshes your recollection, is it in fact true that the way this test was set up that samples 70, 72 and 76 all originated from the same person?
And when you did the DNA typing, "you" meaning Cellmark, not you personally necessarily, did the DNA typing on those three samples, didn't you conclude that although sample 70 and 72 matched one another, that sample 76 had a distinct pattern and did not match either 70 or 72?
That pretty much fits with my recollection, but of--if I'm remembering this group of three correctly, one of them is a mixture. You know, I don't have all my documentation of what we reported with me, so--
Well, Dr. Cotton, isn't it a fact--and by the way here, if you would like to refresh your recollection, here is Cellmark's report to the California association. And I would ask that this be marked next in line--next in order.
For a series of questions regarding this test, if I could have a copy of both the CACLD table and my report, it would make it a lot easier for me to answer the questions more easily.
Okay. Let me just ask you this question, and I will give you the report before you answer the question if you need it. All right?
Excuse me, Mr. Neufeld, you need a foundational question. Do you recognize what it is?
And is it your final report to the California association of crime laboratory directors with respect to your results on the 1988 proficiency test that they gave you?
And in that report, Dr. Cotton, didn't you state that the patterns obtained from stains numbered 70 and 72 were the same?
And didn't you state in that same report that you did obtain a pattern for sample 76 that was suitable for comparison purposes but was not found to match the other samples tested?
And would you agree, Dr. Cotton, that since 76 is supposed to come from the same person as 70 and 72, the fact that you got a pattern but you failed to match it with 70 and 72 would be a wrongful--I'm sorry--would be a false exclusion?
I wouldn't phrase it that way, first of all, because I have to--I would need to determine whether that was this mixture that I was remembering. And secondly, you have to go back to the films to look at the two patterns that you are talking about and see if there is enough data in each of those. That is, if you can't--not making a match is sort of like saying it is inconclusive. We didn't actually even make a list in this report of things which were specifically excluded, so it may mean, based on the data, that we were unable to come to a conclusion, not that we literally said something was excluded and then it was not.
Let me come back to this. No, item 76 was not a mixture, but item 72 was a mixture.
Right. So item 76, according to the way this test was set up, should have matched item 70; is that correct?
Okay. And even though you got a pattern for item 76, which was not a mixture, you did not state that it matched item 70; isn't that correct?
It would if the two patterns were both of equal quality and we just missed it. Remember that at the--remember, you don't--I guess you don't really know this, but each of these samples, these 49 samples, was extracted as if it were a sexual assault sample because we had no information, and in that type of extraction, which is not something that was done in this particular case, you try to separate DNA from--from a sperm fraction from any DNA that is in something else. So of the 49 samples you end up with really two samples from each of the 49, giving you a total of 98, and so you have 98 patterns which were manually visibly compared to one another, and if the patterns are both good, and I can't recall if they are, the fact that we missed seeing it in all this comparison of 98 patterns with, you know, every other pattern of the 98 is a--is an error in finding that match, but in my mind does not--I do not think of it as a false exclusion, a wrongful exclusion.
Well, I appreciate that you don't consider it a wrongful exclusion, but--and you said that one of your concerns was that you don't know how clear the pattern was; is that correct?
Well, would it--the fact that your report states that the pattern for 76 was suitable for comparison purposes, would that refresh your recollection as to whether the pattern for 76 was in fact suitable for comparison purposes?
I understand that it says that, but that doesn't give me any information. The mixture pattern was a problem and it doesn't provide any information about the pattern for 70, so--
Well, what I'm saying is I don't have enough information in front of me to give you a clearance as to whether or not that was not mentioned, because it was basically an inconclusive or whether or not we simply missed the--seeing the match and--and we could have made a match and did not.
Dr. Cotton, is there any qualification in this report, when you are describing the pattern comparisons of sample 70, which is not a mixture, with sample 72?
And of course in your laboratory you wouldn't have called a match if that pattern on item no. 70 wasn't sufficiently clear and robust to compare with 72; isn't that right?
That would be correct in general, but I remember that--I mean, I've looked at these films on several different occasions. They are not as good as the films that we produce now, and so there are things about those films that might lead to problems in interpretation.
And Dr. Cotton, you wouldn't have said in your report, quote, referring to item no. 76 "These patterns were suitable for comparison purposes but were not found in the other samples tested," unquote, unless that was true, would you?
Dr. Cotton, I show you this photograph and ask you is that the photograph reflecting your comparison of samples 70, 72 and 76 that we were just discussing?
Yes, it seems to be, except these weren't all on the same film, so this appears to be various--one, two, three, four, five patterns from different films lined up to be able to photograph them.
Okay. But the actual lanes themselves represent the samples that were tested by your laboratory in this proficiency test, correct?
Yes. Can I ask a question about the picture? Do you know whether or not the labeling across--the labeling across the top must have been put on in order to make the picture, so I--
The label across the top, very much the same way the label was put on the autorads in this particular case.
So could you identify which lanes are attributable to which samples? The same thing was done in this photograph. Does that help?
Your Honor, what is the easiest way for me to point to different lanes for the witness?
Dr. Cotton, would you like to step down? I think it would be easier for you to see this.
Now, in this particular proficiency test series of autorads, it is basically divided up into three different samples. As you can see at the top of the photograph there, there is sample 70, sample 72 and finally sample 76. Do you see that?
Okay. And the lane that represents the non-semen fraction of sample 76 is the second lane from the right where the pink arrow is pointing at. Do you see that?
Okay. And now, what I would like you to do is not look at the mixture, but I would like you to compare--can we have a pink dot just so we will have a permanent point of reference there, so the witness can go back and forth. All right. Now, what I would like you to do is compare the pattern that you see in that lane, the second lane from the right, to go all the way across to sample 70 to the third lane which is also the non-semen fraction of sample 70. Do you see that?
Now, Dr. Cotton, according to the providers of this test, the profile on item 70 that the pink arrow is looking at on the left should match the pattern in sample 76 where there is an arrow there on the far right; isn't that correct?
Well, both sample 70 and sample 76 are not--are not mixtures; isn't that correct?
That's right. I think you are misreading their key, because it is sample 72 that is a mixture of 70 and 76 and 76 clearly does not match 70, but 70 should match one pattern in 72 and 76 should match the other pattern in 72.
Dr. Cotton, again referring you to the code that was given to you by the California association of crime laboratory directors after you did these tests, wouldn't you agree that sample 70 is simply semen, not a mixture that was collected on 7/85? Would you agree with that?
And would you agree, Dr. Cotton, that for item no. 76 it is also the same semen that was drawn on July of 1985, according to this log?
That is not--yes, it is semen, but it is not the same as 70. I don't think that is the intention of that sample description.
On the face of this sample description, Dr. Cotton, would you agree that it says that item 70 and item 76 come from the same person?
From looking at this report, is it your understanding that the California association of crime laboratories grouped all samples that came from the same person in the same box?
What they've done is grouped all samples that could be related in the same box and these are grouped together, (Indicating), because both 70 and 76 should match 72, because 72 is a mixture, so that is why they are in the same box. But it doesn't mean, and I'm sure that they could confirm this, that 70 and 76 were the same. I mean, there wouldn't be any point of making a mixtures of 70 and 76 if they were from the same person, and so I feel confident that that is what that means.
If we work off of your premises, Dr. Cotton, then nonetheless, according to this log, sample 76 should be consistent with the mixed stain in sample 72; isn't that correct?
And would you agree, Dr. Cotton, that in your report you failed to state or include 76 as being consistent with the profile in 72?
It is not actually--it is going to be one page of the--of the six-page report of the CACLD.
Dr. Cotton, did you say you recognized this six-page report as a report that you received from the California association of crime laboratory directors assessing your results in this proficiency test?
Umm, what I did was look at the two keys and I recognized them. If you would like me to look at the six pages, then I can--
(Witness complies.) The report was, I believe, not just to us; the other two labs that participated and other people in--within the CACLD.
But this is an accurate copy of the report that you received from the California association of crime laboratory directors regarding the results of the 1988 proficiency tests?
And by the way, when you ordinarily participate in proficiency tests, is it the ordinary course of business that the provider of the proficiency test furnishes the participant with a report evaluating your performance?
And when these proficiency tests are given to laboratories such as your own, is the provider of the proficiency test under an obligation to report the results accurately?
Yeah, I think they certainly would try to do that. I have seen mistakes in reports, but you know, everybody, when they do this, is trying to report the results accurately.
I would offer the page again, your Honor. I think it is a business record exception, your Honor.
And is a copy of this report actually retained by you in your record as part of the or in the course of Cellmark's business?
Counsel, why don't you take a look at the evidence code. You are still missing a few items.
Doctor, are the results that are contained in this report recorded at or near the time of the events that it reflects?
Since I didn't participate in writing the report, I wouldn't know. I mean, I know that it was done after the proficiency tests were completed, but that is all I know in terms of its timing.
And are they relied upon as being accurate by those persons, members of that organization?
I'm sorry, I still have an objection as to at least one of the elements, your Honor.
Now, Dr. Cotton, in this particular test you received 49 samples from the California association of crime laboratory directors; is that correct?
And in fact the data that applies to Cellmark would be in the column on the far right of that particular exhibit?
And of the 49 samples that you received, you were able to generate results on 44 of those 49; is that correct?
And so as to the remaining five samples, that would be the third number in that column, you were unable to generate DNA results?
And now the fourth item on that list is a no. 1; is that correct, with a series of little asterisks next to it?
And that no. 1 indicates that out of the 44 samples that you were able to type that you made one incorrect match; isn't that correct?
And when you say one incorrect match, that would be the equivalent of one false positive; isn't that correct?
Now, the way you got this false positive, Dr. Cotton, on this particular proficiency test, that is there were two samples, namely 57 and 58, that came from the same person; isn't that right?
Well, again, I don't remember the sample numbers, but there were two samples--you can go on and I'm pretty sure I can follow you without the specific sample number.
Let me show you a photograph. Next in order--let me show you, Dr. Cotton, what is Defense 1157, next in order.
Dr. Cotton, does that photograph appear to illustrate the banding patterns for three different samples that you tested in that proficiency test; namely, 57, 58 and 59?
When you did the DNA testing on these three different samples, 57, 58 and 59, did you utilize what you earlier testified to in this trial as a single locus cocktail?
And that is a situation where instead of simply probing the DNA with one genetic marker, you do more than one genetic marker at the same time?
Dr. Cotton, is that profile the DNA profile that the pointer is pointing at, the DNA profile from sample 57?
And Dr.--whoops--and Dr. Cotton, is that second arrow pointing at the DNA profile representing sample 58?
Now, in this particular proficiency test Dr. Cotton, sample 57 and 58 came from one person, correct?
Right. Okay. Now, when you analyzed these three samples at Cellmark, were they all processed as part of the same batch?
I think that they were, but since I was not--I wasn't really supervising any work in the laboratory at the time that this test was done, the records, if I remember correctly, indicate that they were processed in two separate batches, so it is possible that 59, 58 and 57 were processed at the same time.
Thank you. And would it be fair to say, Dr. Cotton, that somehow during the processing of these samples that either--that DNA from either sample 57 or sample 58 was accidentally put into the tube with sample 59?
And that mistake led to the false match of 59 with sample 57 and 58; isn't that correct?
And wouldn't you agree, Dr. Cotton, that the bands in this false match are as strong and as clear as the bands in the match between Mr. Simpson's sample and item 5 to in this case?
Dr. Cotton, with respect to band density and relative band densities, would you agree that the bands in this proficiency test where you had a false match between 59 and 57 and 58, are as dense as the bands in item 52 in our case?
Did you say earlier during your testimony, Dr. Cotton, that at the time you did the proficiency testing in this--in 1988 and 1989, that you were not producing as high quality autorads as you are currently producing in your laboratory?
Given that statement, Dr. Cotton, wouldn't you agree that the banding profile in item 59 and its comparison with item 57 is as high quality a banding pattern as the banding pattern you have comparing 52 with Mr. Simpson's reference sample in this case?
Dr. Cotton, would you agree that just by looking at the photograph of the autorads in this case, that there is no way just by looking at them that you can tell that you had in fact gotten a false match?
And would you agree that the reason that you have this false match is because the original sample that was in the tube for 59 was so degraded as to not be observable?
I don't think we know whether it was degraded or it was just too small of a quantity.
But both of those explanations could produce a result where you don't see a DNA pattern; isn't that right?
And in our case, Dr. Cotton, had the DNA, the original DNA from the Bundy drops been so degraded due to handling or packaging or processing, such that you wouldn't see a DNA pattern in the original sample and then it was cross-contaminated with Mr. Simpson's reference sample, would you also have this same phenomena?
Well, would you agree, Dr. Cotton, that this type of phenomena that you observed here for sample 59 can occur whenever the original sample is so degraded that it will not leave a visual pattern and is then contaminated with another sample of DNA?
And Dr. Cotton, had the--had you then been asked to type a frequency for this particular match, that frequency would indicate a very rare event, would it not?
Yes, you would type--you would make a frequency and it would almost certainly indicate that banding pattern was a rare event.
In fact, Dr. Cotton, for this particular banding pattern, didn't you generate in your report to the CACLD that it had a frequency of approximately one in 1.8 billion?
If you are reading to our report, then in--yes. I know that we generated frequencies. Obviously I can't remember what the frequency was for a particular pattern.
And so what you are saying in this particular report is the frequency of that match between item 57 and item 59 occurs in the population with a frequency of approximately one--I'm sorry--occurs with a frequency of approximately one in 1.8 billion people; is that right?
Right. We are seeing the frequency of the pattern. Not the frequency of the match, but the frequency of the pattern would be--well, for that report it says 1.8 billion.
Well, if had been a real case, item 57, 58 and 59, and not simply a proficiency test, you would have declared a match between item 57 and 59; is that correct?
And the frequency that you would have generated would have been one in 1.8 billion people, correct?
But in fact in this particular example there was a 1 in 50 chance that it was a false positive; isn't that correct?
Well, was it a one in 50 chance that it was a false positive in this particular proficiency test?
Well, then let me ask you this simple question, Dr. Cotton. I hope it is simple. In this particular instance you said the frequency of the pattern was one in 1.8 billion people; is that right?
So the one in 1.8 billion would incorrectly create some value to that match which would be a distortion, wouldn't it?
Well, would you agree, Dr. Cotton, that since 59 and 57 don't match it would be wrong to suggest, in that particular instance, that that match is valued at a frequency of one in 1.8 billion people?
You keep referring to a value of a match. The frequency doesn't tell you anything except how often you would see this banding pattern, so the frequency is only that. It doesn't say anything more than that. And yes, you are correct, this is a false match. This is an incorrect match. It was caused by laboratory error.
Because it is a false positive match that number of one in 1.8 billion doesn't apply to the frequency of the pattern in lane 59; isn't that correct?
Well, it actually still applies to the frequency of that pattern. The frequency of that pattern is whatever it is. In fact, that pattern matches because it is the same DNA as I actually believe it is 58 that that--we demonstrated that sample came from. It is like you are talking about--I'm hearing what you are asking is talking about apples and oranges.
Well, Dr. Cotton, would you agree that if this particular proficiency test was a real case that the statistic of one in 1.8 billion would be irrelevant if in fact you knew it was a false match between 59 and 57?
Now, after you got this false positive in this series of proficiency tests in 1988, did you make changes in your procedures in an attempt to prevent or at least to minimize the likelihood that DNA from a sample could cross-contaminate another sample?
We made changes in our procedures to specifically address the way that this particular contamination occurred.
And did you believe that when you made those changes that that took care or solved the problem?
Well, I think it solved this problem. That--that clearly doesn't imply that it could solve every possible problem.
But in 1989, the very next year, you participated in a second round of California association of crime laboratory directors proficiency testing, didn't you?
Now, this again was caused by DNA from one sample cross-contaminating a second sample, was it not?
And did you conduct a full investigation to try and discover what caused this cross-contamination?
And in that investigation did you conclude that the amount of the contaminant which caused the false positive was somewhere between 20 and 50 nanograms of DNA?
And in this particular case, Dr. Cotton, this case being the Simpson case, wasn't it your estimate that the quantity of DNA in lane item sample 52 was approximately 25 nanograms of DNA?
It was my estimate that the amount of DNA that is participating in giving that pattern is 25 to 50. The total amount of DNA in that sample was more like 200, and I remember giving an explanation of where those two numbers came from.
But the amount of DNA that is actually in the lane is somewhere around 25 nanograms? Was that your initial estimate?
The amount of DNA that is giving the pattern is between 25 and 50. The amount of DNA handled with that sample in total is about 200.
So the amount of DNA that is 25 to 50 nanograms in that particular item, at least as it appears in the RFLP test, would be within the 25 to 50 nanogram range that you attributed to the cross-contamination in the proficiency test; is that correct?
KEY QUOTENow, how did you go about attempting to figure out how you got this false match in 1989?
We went into all the handwritten notes. Two different analysts had done the samples. We spoke with the analysts to find out what samples they had done adjacent to one another. And the reason was--when I say "adjacent," that the test-tube rack holds, I don't know, say, 15 samples across the top row and another 15 samples across the bottom row, but anyway, about 15 could be side-by-side. We knew that--we knew a whole lot of places where we could figure out a whole lot of places where the contamination didn't occur, and so--we were finally able to analyze that the contamination had to have occurred in the extraction of the sperm fraction. Again, everything was extracted as if it could have a non-sperm and a sperm fraction, it could have occurred in the extractions of the sperm fractions, and the samples that were involved in producing the contamination were worked side-by-side. So we were able to eliminate other places where it could have occurred, for example, as you load the gel. It couldn't have occurred as you load the gel because the two samples weren't loaded side-by-side. So anyway, we narrowed it down to this one section and beyond that we weren't able to definitively say exactly how the contamination occurred. We could just say it occurred in these series of steps.
Now, so what you did, in other words, even after this thorough investigation on your part, was to narrow it down to about what point in time the cross-contamination occurred; is that right?
But even to this day, Dr. Cotton, would it be fair to say that you don't know how the accidental contamination occurred during the extraction process?
And would it also be fair to say that this cross-contamination, this accidental cross-contamination happened in spite of the fact that all of the sample transfers were witnessed by a second person to make sure that no errors occurred?
And again, Dr. Cotton, after you got this second false positive, did you make certain changes in your protocol again to try and avoid this type of problem from recurring in the future?
We mandated that the DNA extractions for the known samples could not be done out on the bench at the same time as the DNA extractions from the evidence samples.
Would the DNA extractions from the known sample be put into a different extraction rack than the DNA samples from the evidence?
I'm sorry, I thought you mentioned before there was this rack and it might have 15 samples in it or some such number. Did you say something to that effect?
Well, do you also have a similar kind of rack process when you are analyzing case work?
So what I'm asking you is under your current protocol will you ever have the extraction from the reference sample at the time of extraction sitting in the same rack as the evidentiary specimens?
And based on your experience, Dr. Cotton, and your expertise, would you agree that a laboratory which places the reference sample at the time of extraction in that same rack with the evidentiary specimens runs the risk of falsely incriminating innocent people?
Dr. Cotton, would you agree that a laboratory that extracts the reference sample and puts it in the same rack, same extraction rack as the evidentiary specimens, runs the risk of getting false matches between the reference sample and the evidence?
Well, based on our experience and that that happened in our lab, then obviously it is possible.
KEY QUOTEAnd since 1989 I believe you said you haven't participated in any proficiency tests where you have been asked to handle more than a dozen samples; is that correct?
Well, I think I said we haven't--the proficiency tests normally have three or four samples.
In this case in total we analyzed two samples. That was done over a relatively long period of time, so they weren't all analyzed--they weren't all received at the same time and of course they weren't all analyzed at the same time.
And Dr. Cotton, these false matches that you made on these open proficiency tests, they were made during the first two years that your laboratory was doing DNA typing?
Now, would it be fair to say that you have been improving your technique over the last six years?
But I think you also said that you have no control over the possibility of cross-contamination that occurs on samples before you receive them?
And in this case every sample that you received was initially collected, packaged and processed by the Los Angeles Police Department?
It is a foundation question for the next one and there is no other question on that, your Honor. With the Court's permission I would just like to ask that one question.
Now, since you became involved with the Los Angeles Police Department, have you ever been asked to review the Los Angeles Police Department's procedures that were used in this case?
Do you have any personal knowledge, as a profession, as an expert, as to the extent that the Los Angeles Police Department has been subjected to external blind proficiency testing?
Would you agree, Dr. Cotton, that you are more reliable now as a laboratory than you were in the first two years of your operation when you made these two false positives on the CACLD tests?
I don't know that because we made these errors in the tests that we were unreliable in our other work. I--I don't know how to answer that question.
Well, let me ask you this, Dr. Cotton: Do you think that you are still getting false positives with the same degree of frequency now that you got them in those two proficiency tests in 1988 and 1989?
And do you know, Dr. Cotton, as to how long the LAPD have been doing DNA case work before they did the work in this case?
Your Honor, with the Court's permission I would like to substitute 1157-A, which is the printout, for 1157.
You are exactly right.
Yes, it would be irrelevant.
Yes, that would be fair to say that.
So the amount of DNA that is 25 to 50 nanograms in that particular item, at least as it appears in the RFLP test, would be within the 25 to 50 nanogram range that you attributed to the cross-contamination in the proficiency test; is that correct?
Well, based on our experience and that that happened in our lab, then obviously it is possible.