Thursday, December 15, 2011

The Cellmark report, and what it shows

Here is the Cellmark report, taken from the public files of the United States Supreme Court in Williams v. Illinois. A few points about it:

First, it seems to me that simply looking at the report demonstrates whatever degree of formality any justice is likely to require for a statement to be considered testimonial. It is not sworn, but we know that this fact in itself does not make a statement non-testimonial. It is on letterhead, dated, with a title, “Report of Laboratory Examination,” addressed to a recipient at the Forensic Science Center in Chicago, and signed by two laboratory directors. It bears two case numbers (I gather one for Cellmark and one for the submitting agency). It refers to the “exhibits received” and then to the disposition of “evidence.” Clearly it was made in contemplation of use in investigation and prosecution of crime.

Second, examination of the report should make clear that it was not simply the product of a machine (even assuming that mattered; the electropherogram, which was indeed the product of a machine, still needed human input at least to identify the sample tested). There is a one-page summary report, stating what tests the lab used and on what samples, summarizing the conclusions, and proposing a return of the evidence (which presumably would have allowed retesting!), together with a second page containing the critical allele charts. The male donor profile deduced by Cellmark is contained in the second of these charts. These first two pages are produced by humans. After these two pages, the report incorporates the electropherogram.

Third, the report makes clear that Lambatos could not have come to her own independent opinion as to what the male donor profile was simply by examining the electropherogram sent by Cellmark. As Lambatos testified repeatedly, the only electropherogram sent by Cellmark was of the mixed profile – meaning the sample contained DNA from both the victim and the male donor. There is no way to tell from that alone who produced what alleles. To deduce the male profile, it was also necessary to know what alleles were present in the victim’s DNA, and Cellmark did not send an electropherogram of her profile. (Of course, that could have been done, but it wasn’t.) What Cellmark did provide with respect to the victim was simply an allele chart (on the second page of the report), representing its statement of what the victim’s profile was. See p.5 of petitioner’s reply brief. From that, it deduced what the male profile was. Some of the deductions required relatively delicate judgments. (For example, at one of the loci Cellmark reported three alleles. It is possible, but rare, for an individual to have three alleles at one locus – but Williams does not; evidently, Lambatos interpreted the report to mean that the assailant’s profile included two of the three alleles, a factor that should diminish slightly the probative value of the evidence.)

I believe that Lambatos conveyed more than her opinion of what the deduced male profile was; I believe she really used the profile deduced by Cellmark, with which she said she basically agreed (notice, for example, that at JA 65 she testifies that she entered the profile from the Cellmark report to perform the data-base search), and the reliance on Cellmark gave the evidence extra heft. But in any event, Lambatos could not have reached an opinion as to what the male profile as based simply on the machine-generated information presented to her.

Fourth, and as a related matter, I think it is clear from the report that it was not merely used circumstantially. The circumstantial evidence argument, as I understand it, is that from the fact, as indicated by shipping documents, that the mixed sample was sent to Cellmark and a report purporting to be on that sample was received from Cellmark, one can infer circumstantially that the report received was the proper one. That seems to me to be a stretch, especially given that Cellmark and the state lab sent materials for numerous cases at once, but let’s assume it’s valid. If it were, and if the report was nothing more than a machine printout, then there might be some force to the argument. But it appears to me that the whole thing falls apart when one realizes that the critical part of the Cellmark report was not machine generated. It bears emphasis again: The electropherogram, the only part that was machine generated, showed only the mixed profile, not the male profile. The male profile that Cellmark deduced was presented as a human-produced statement in the allele chart on the bottom of page 2. It is essentially a statement that the donor of the sperm had 23 or 24 of 25 listed features. (I word it that way because of the three-allele locus; as I understand it, Cellmark was saying that the assailant had one or two of the three alleles listed at that locus.) In that sense, it is no different from a statement describing a couple of dozen visible features that the assailant had (hair color, eye color, height, and so forth). This is a detailed human statement that purported to describe the assailant, that was used to help identify him, and that could do so only if it was true. Frankly, the argument that the evidence was used circumstantially strikes me – notwithstanding the great respect I have for Michael Dreeben, who argued for the United States, and also for Paul Vinegrad, who has articulated the argument in commentary on this blog and elsewhere – as a lot of hooey once one understands the nature of the Cellmark report

I hope to post another entry soon discussing a way in which the Cellmark might possibly have been presented as circumstantial evidence – but that’s not what happened here.

Finally, similar reasoning should dispose of the argument that the problem was not one of confrontation but rather of adequacy of the proof. If Cellmark produced a thing, and that thing tended to prove guilt, but only on the assumption that it was what the prosecution contended it was, then there presumably would be no Confrontation Clause problem; there would only be a problem of authenticating the thing. But that simply isn’t what happened. Again, it seems to me that there is no getting around the fact that the Cellmark report was a detailed testimonial statement produced by humans that supported the prosecution case only if it was true: The report asserted not merely that a male profile had been found, and not merely that the testing had been done accurately. It also asserted what the profile was (allowing for some uncertainty at one of the loci). If it had not made that particular assertion, it would not have helped the prosecution. And the fact that it asserted a given profile – i.e., one that was later determined to match that of Williams – was clearly conveyed to the trier of fact. This is a Confrontation Clause problem, pure and simple.


Albert Locher said...

Addressing Prof. Friedman’s comments on the Cellmark report (Part 1):

1. Data Base Search: The fact Lambatos used the Cellmark allele chart for the initial database search for a match is irrelevant. This process identifies a suspect. After that, the analyst does further work, examining the original data from both the known suspect sample and the rape kit. The analysis and opinion which comes from this new examination (prompted by the database hit) is the evidence presented in court; any mention of the database hit is superfluous. (JA 55-56) The same thing happens with automated database fingerprint matches, and other investigative procedures that point to a suspect.

2. Mixed Sample Interpretation: The Cellmark report shows that the rape sample was a “mixed sample”; the victim’s alleles, found in her reference sample, were subtracted from the overall to deduce the male portion; and the victim’s profile was only represented in the report materials by a table, not by the victim’s electropherogram. One should note there will be little if any interpretation to deduce the victim’s DNA from analysis of her blood. If Cellmark had included the victim blood sample eletropherogram so Lambatos could have reviewed it, would that satisfy you? If that is all it would take, we are only talking about transferring specific data results from the electropherogram to a table like page two of the Cellmark report. That hardly seems a tipping point of Constitutional dimension; if it is, including the victim’s electropherogram solves the issue.

3. Identity of the Sample: In the blog you say the electropherogram is not simply machine produced data, because the machine still needed human input, “at least to identify the sample tested.” True enough. But that is true of not only the processing of done at the phase which produces the electropherogram. The opening of the rape package at Cellmark occurred several steps before the electropherogram step. The swab was screened, and a cutting was taken; the cutting was subject to DNA extraction; the extracted DNA went through PCR amplification; the amplified DNA was processed to attach fluorescent markers, all before it was fed into the genetic analyzer. In an assembly line process, for each step a different analyst was involved, documenting the work as being done on this case, The appendix to the OCME brief (from another lab that uses assembly-line processing) gives examples of how some of that paperwork appears in their lab, at pp. A1 through A11. Why is only the last step when “identity of the sample” is entered into the last machine testimonial, requiring a live witness? Why is not each person who materially processed the sample along way required to testify? Each successive technician relies on the identifying information produced by the predecessor(s), to ensure the proper ID for the next person who deals with the sample or data. If the “identity of the sample” is a point requiring live testimony of a witness who can give direct evidence identifying the sample, no one has offered a principled explanation that will avoid the “all analysts/technicians must testify” result.

Albert Locher said...

Addressing Prof. Friedman’s comments on the Cellmark report (Part 2):

4. The Fact Illinois sent Multiple Kits to Cellmark: Your blog entry says this makes it "a stretch" that the data Lambatos received from Cellmark referred to the correct sample (for L.J.). Why? The testimony of Hapack and Lambatos make it clear that each separate rape kit is packaged and labeled separately. These smaller boxes may all be put into one larger box and then shipped together, but why does that implicated the intergrity of the evidence any more than if they were shipped separately? In either case, they wind up in the lab at Cellmark, together with boxes from other jurisdictions, where they are all processed separately, to avoid cross-contamination, as accreditation standards demand (at least, a trier of fact is entitled to draw that inference; see below).

5. Who must testify: The Cellmark report was signed by Dr. Robin Cotton, Dir. of the Cellmark Forensic Lab, and Dr. Jennifer Reynolds, Dir. of the Cellmark Identity Lab. Neither of them did any of the lab work, nor handled the samples at any step. Each relied on the work of others. Must we call Dr. Cotton and/or Dr. Reynolds? Must we also call the analysts A1 and A2 who put the forensic sample and the reference sample into the genetic analyzer machine? That is 3 Cellmark witnesses. But then don't we need others to establish that the samples A1 and A2 processed are the samples from this particular ISP case (the "identify of the sample" issue, see above)? At a minimum, that brings us to 3 or 4 added witnesses each for the various steps for the forensic and reference samples, a total of at least 9 (or more) Cellmark witnesses. Certainly, everyone involved in the process (with the possible exception of the FedEx employees who were the shipping intermediaries) knew their work was for a criminal case, and that if it led to identification of a suspect, it would wind up in court.

6. Circumstantial Evidence: Here, the circumstantial evidence includes: (a) the testimony of Hapak as to how he packaged and labeled the rape kit items; (b) the testimony of Lambatos as to ISP rape kits are packaged, labeled, and shipped to Cellmark, with a record on the shipping manifest; (c) the shipping manifest showing the return shipping and return Cellmark of those items, with the report; and (d) the fact that Cellmark is an accredited lab. Dreeban (from the Solicitor General) discussed these in oral argument. I would add another piece of circumstantial evidence -- the fact that L.J. identified in a lineup Williams, whose DNA matched the DNA from the rape kit. From these, the trier of fact can reasonably draw the inference for the ultimate facts (the definintion of circumstantial evidence) that the L.J. rape kit was sent to Cellmark, analyzed in accord with accreditation standards, and it was those results Lambatos relied on in her analysis. This may not be the strongest, but if such circumstantial evidence is not allowed, you are saying only direct evidence (a witness who personally observed and/or conducted each step) satisfies the Confrontation Clause; circumstantial evidence does not.

TASA said...

thanks! another top rate post!

Veronica Carly said...

Your blog is very nice.It is very helpful for the public files of the United States Supreme Court.

Waste Disposal