Sunday, July 20, 2014

Retesting the knife for DNA: a dubious rationale, but an exculpatory result

Part 38 in the Knox/Sollecito case

Background on the DNA profiling of Sample 36-I
In overturning the Hellmann-Zanetti acquittal of Knox and Sollecito, the Supreme Court of Cassation (SCC) published a motivations document in 2013 (the previous blog entry critiqued section 12 of the motivations report on genetic investigations).  The present entry discusses section 11, another DNA-related portion but one that focused entirely on the retest of Raffaele’s kitchen knife, the subject of two previous entries here.  One of the court-appointed independent experts, Dr. Carla Vecchiotti, had found the sample in question (36-I) by swabbing the knife in 2011.  Because this amount was far below the recommended amount of DNA for the reagent kit, she decided that the amount was too small (its concentration was estimated to be 5 pg/┬ÁL).  Pro-guilt commenters sometimes imply that because the defense had the opportunity to object to certain aspects of DNA testing in 2007, that their subsequent requests for the raw forensic data should be turned aside (this argument is problematic on multiple grounds as discussed in the previous blog entry).  Yet if we accept that an objection to some aspect of testing should be lodged immediately, then either the prosecution's observers should have objected to not amplifying 36-I in the spring of 2011, or the prosecution should have admitted that it lost the opportunity to do so when it did not.

Why did the Supreme Court of Cassation demand an amplification of the new sample?
In section 11 of their motivations report the CSC wrote,
“This trace was not subjected to genetic analysis – through a decision made by one of the experts, Professor Vecchiotti, alone, without documented prior authorization to that effect by the Court, who had also given a mandate to attribute the DNA on the present findings on the knife and on the bra hook- because the amount was not sufficient to provide a reliable result, amounting to Law (sic) Copy Number. This choice, however, met the subsequent sharing of the group, on the assumption that such a small sample would not have allowed two amplifications necessary for a reliable result (p. 84 of sentence).

“So that when the Procurator General and the civil plaintiffs demanded to complete the examination, strongly as a result of the scientific contribution of Professor Novelli, geneticist of undisputable fame recognized by the same Court (p. 79 sent.) on the availability of equipment able of operate with safety also quantities of less than ten picograms, in the areas of diagnostic character (even on embryos) in which the claim to certainty is certainly no less pressing than that which animates the legal field, the Court rejected that proposal, assuming that the methods to which Prof. Novelli had referred were "in the experimental phase” (p. 84 sent.), thus freely interpreting and misrepresenting the assumption of bias, which in fact was to remember the use of these diagnostic methods in areas in which you can be certain of the result.

“…Especially since the renewal of genetic investigations were requested in 2011, after four years from the initial time and over which the evolution of instrumentation and methods of investigation had marked significant milestones, as was emphasised by the advisor to the Procurator General, Professor Novelli. Just on receipt of the information from the consultant mentioned who - under the constraint of the obligation to truth, spoke of cutting-edge techniques -, the Court fell into a new gross misrepresentation of argument concerning the reliability of the results of investigations carried out assuming no new findings of such remedies, even through developments emerging at a later time, concerning reasonableness of the grounds (Section I, 25.6 .2007, n. 24667).

“…[the survey must] bring to analysis also the newly sampled trace, according to the most accurate and modern “experimental” analytical techniques, under pain of violation of the law for not making a decisive test and the fallout in terms of manifest illogical reasoning (again for obvious incompleteness of the inferential platform, to have overlooked data that is not only important, but crucial), as was correctly pointed out by the public plaintiff.”

Leila Schneps and Coralie Colmez summarized Novelli’s testimony and offered their opinion in the chapter, “The test that wasn’t done,” in their 2013 book Math on Trial (pp. 83-86):  "In court on September 5 [2011] and supported by expert prosecution witness Giuseppe Novelli, Stefanoni explained that newer generations of DNA analysis kits existed in 2011 that had not been available in 2007, and these new kits could give results on as small as a couple of cells.  She wanted a new analysis performed to confirm that her previous work was correct.  The prosecution agreed and asked the judge to order the new tests...Judge Hellmann missed a major opportunity to get at the truth.”  According to the CSC, techniques that are appropriate for diagnosis in embryology are also appropriate for forensic investigations.  There have been advances in technology, and the new tests should take advantage of them.  The results will be "decisive."

Not every observer sounded convinced of the probability-based rationale Schneps and Colmez put forward.  Professor Daniel H Kaye wrote, “If these experts’ concern — that the original DNA test was simply detecting traces of Kercher's and Sollecito’s DNA that investigators inadvertently transferred to the knife and bra clasp, respectively — then repeating the tests could well continue to detect that DNA — and prove nothing more than the original tests did.” In other words one weakness in the rationale provided by the SCC is that a newer generation of kits cannot circumvent the problem that one is dealing with DNA in the low template range (see below). 

Can one equate pre-implantation genetic diagnosis with DNA forensics?
It may be that Professor Novelli left the court with the impression that single-cell PCR was new technology, but single-cell PCR in diagnostics is at least twenty years old.  Yet the problem with using diagnostic practices to justify further low template testing on the knife is that it ignores that one is still dealing with low template quantities and all the problems that such analysis has in a forensic setting.

In the article “LCN DNA Analysis: Limitations Prevent 'General Acceptance'” Dr. Angela von Daal wrote, "The use of PCR for analysis of very low levels of DNA in the field of pre-implantation genetic diagnosis (PGD) has been used to justify general acceptance of LCN analysis in the wider scientific community(12). This argument is flawed. PGD analysis is not analogous to LCN analysis for several reasons. PGD uses pristine cellular DNA from a single source, whereas forensic LCN samples are mostly mixtures and are likely in a state of apoptosis. The complexities of profile interpretation issues seen with STR analysis (e.g., stutter) are not an issue for PGD testing(13). However the most significant difference is that the samples derived from the mother, father and embryo are single-source and the parental samples are of known genotype."  In other words the connection between pre-implantation genetic diagnosis and low template DNA forensics is tenuous and trying to extrapolate from the former to the latter is dubious at best.  Why Professor Novelli chose to ignore the problems in his analogy is something only he can explain.

Did the SCC understand the unique problems of low template DNA profiling?
And the problems of innocent transfer of DNA get more severe as the sample size gets smaller.  Sara Gino testified for the defense in the trial of the first instance, and some of what she had to say is pertinent to this issue. From the Massei report (p. 258, English translation): “She reaffirmed that [the risk of] contamination exists, and emphasised that in minimal quantities of DNA there is not necessarily a greater risk of contamination but it was easier to notice the effects of the contamination and be misled (‘...It's not that the risk of contamination is greater; but it is easier to see the contamination...’ page 92).” In response to a question on this subject, Professor Dan Krane expounded on Sarah Gino’s testimony, “There is absolutely no question but that contamination is a much greater problem in LCN cases than conventional DNA testing. The reasons that it is a greater problem are both because it is easier to detect contaminants ([Sarah] Gino's point) and because it is easier to transfer (and to transfer without knowing) smaller amounts of DNA than larger amounts of DNA.”

Proper low template profiling demands that one rework the entire process of DNA forensics, from collection of the evidence all the way through the final step of analysis of the electropherogram.  True low copy number LCN profiling requires ultra-clean, dedicated facilities in order to perform it properly, as clearly discussed in nontechnical language in an article in the New Zealand Herald. Without those extra precautions, the odds of contamination are necessarily greater (or else the precautions are a waste of time and money).  In a review article in Nanomedicine Professor Novelli and coauthors wrote, “Very few laboratories perform low template DNA typing properly, because it requires dedicated facilities and great experience, although there are several published methods for the interpretation of such profiles [80-82].”  This blog has previously discussed the need for proper air handling facilities.

Even so, within the forensic DNA community there are a number of critics of LCN profiling.  Dr. Budowle and colleagues, for example, argued that low template analysis should be restricted to identification of bodies and the generation of leads.  In other words using the practices in embryology to justify further testing on the knife was predicated on a very weak analogy and one that has been seriously questioned by molecular biologists.  Professor Novelli should have known this and should have acknowledged this limitation to the court.

What other forensic issues did the Supreme Court of Cassation ignore?
In forensics it is impossible to interrogate the DNA as to the time or manner of its deposition.  With respect to the knife, the DNA may have arrived from secondary or tertiary transfer prior to its collection or during its handling.  Regrettably, the knife was removed from its package at the police station, a breach of good practice.  The person who repackaged it had previously been to the women’s apartment earlier that day, a second error.  In addition, the negative controls have never been produced in the form of raw data or probably any other form.

What did the proponents of the retest expect to learn?
As this blog has repeatedly argued, the DNA from Sample 36B in 2007 was probably from laboratory contamination.  Pro-guilt commenters have resisted this interpretation.  In a story from 27 April 2013 the BBC’s Ruth Alexander asked Coralie Colmez this question.  “’So what this means in the case of the knife in the murder is that if it were tested again, and once again the DNA was Meredith's profile we could be a lot more certain that the DNA on the knife is indeed Meredith's,’ Colmez says.”  On the other hand suppose that the test came out differently.  “And if the knife were tested again and the DNA did not match Meredith Kercher's profile? That would be good news for Knox and Sollecito, she says.
‘This would mean that this major piece of evidence against them would be discredited.’”  In other words either way the results turned out, there would be useful, arguably decisive, new information.

What did the new test actually tell us about the knife?
Andrea Vogt wrote, "The RIS Wednesday deposited their forensic report on trace 36i, a spot of DNA identified (but not earlier tested) on the kitchen knife alleged to be the murder weapon. “Cento Percento” (100 percent) said Major Berti, discussing compatibility. The RIS found that the DNA was compatible with Amanda Knox, and excluded that it was that of Sollecito, Guede or Kercher. The RIS expert was asked only a few questions from attorneys and the judge. The judge asked why the RIS had done two amplications of the DNA and not 3 or 4. Major Berti described that two is considered the minimum number of amplifications necessary, according to today’s forensic standards, doing less (or more) might have diminished the reliability of the results. The judge also asked about the age of the equipment used. Berti responded that the forensic kit used this time has been commercialized since 2010 and available for use since 2011. At one point the judge stopped a line of questioning by Knox’s Rome attorney Carlo Dalla Vedova, who was asking why the RIS described Knox’s DNA as “fluids” when a prior expert had said the trace did not come from blood. Nencini said: That question was not put to the RIS by this court, it was not their job to determine that. The other experts’ reports are in the case files for everyone to read, he noted, adding: “We cannot put words in the mouth of this expert that were said by another expert.”

The presence of Ms. Knox’s DNA on the handle knife has never been in dispute, and the finding of additional DNA on the blade is not inculpatory.  A recent review article on trace DNA noted that DNA transfers can occur after an item of evidence is packaged; therefore, the DNA on the blade may have arisen from transfer after the knife was packaged, or from direct transfer: chef’s knives are often gripped both at the handle and the base of the blade.  If anything, the result is modest additional exculpatory evidence.  If the knife were the murder weapon, it would have Meredith’s blood.  Somehow all of Meredith’s blood would have to be removed, yet DNA from Meredith and Amanda would remain, along with starch.  A letter to the court from Professor Bruce Budowle indicates that it would be difficult to clean a knife of blood.  There was additional DNA in sample 36-I belonging to an unknown person, a finding which further calls into question the hypothesis that it was the murder weapon.

Are the new results exculpatory?
In early 2014 Leila Schneps wrote, “By the third trial, when a new attempt was made to collect DNA from the knife (which had been swabbed again during the appeal trial, though no tests were then conducted) there was no match to Meredith – a result welcomed by Knox's defence team, though it did not in fact impact on the findings of the first trial.”  This passage presents quite a contrast with the earlier quote from Coralie Colmez (above).

There are two important issues.  The first point is that Meredith Kercher’s DNA was not found, and this was the whole reason for performing the test.  The people who argued for the test did so on the basis that the retest might improve the reliability of the 2007 result.  They should at the very minimum concede that the original result on the knife has now been “discredited,” as Ms. Colmez wrote.  The second point is that the 2013 result did indeed impact on the finding of the first trial.  The Carabinieri noted that at least two amplifications of the DNA must be done, and Stefanoni apparently did only one.  Therefore, the lack of retesting with respect to the original result is indeed makes it unreliable by the expert testimony of the Carabinieri, which support the testimony of Conti and Vecchiotti.  In other words the result and the testimony are a one-two punch that should have knocked the kitchen knife right out of the trial.

Is the Nencini court being impartial and objective with respect to the forensics of this case?
An early report from Andrea Vogt on Judge Nencini’s motivations document from the Florence court indicates that Nencini harshly criticized the rationale for Conti and Vecchiotti’s not completing the test in 2011.  For the reasons given above, this argument has no merit.  A cynical observer might conclude that it is a canard to distract future courts from the real issue, which is that Meredith Kercher’s DNA was not found, and this was supposed to be a decisive test.  Pro-guilt commenters might now argue that the test would have been decisive if Meredith’s DNA had been observed in 36-I but any other result is inconclusive.  However, such an argument prompts a question:  Why should the Nencini court order a test that might only be beneficial to the prosecution and not also grant requests that are favorable to the defense?  This blog has long argued that the failure of the prosecution to release the raw DNA data constitutes a serious breach of discovery, rendering the trial unfair.

Another serious logical problem with the SCC and the Nencini court is the selective indignation about the lack of amplification of 36-I in 2011.  Mrs. Stefanoni has failed to produce many electropherograms which showed positive quantification of DNA, some of which may contain highly probative evidence.  The defense has every right to see these data under any reasonable standard of discovery.  One profoundly wishes that the Nencini court had demanded the production of the electropherograms that originated from a putative semen fraction, to take just one example.  The fact that it did not do so is only one reason to call its objectivity into question.

Update (13 August 2014)
In the Lindy Chamberlain case a forensic worker, Joy Kuhl, invented a new test for fetal hemoglobin.  The defense witnesses were sharply critical of her methods.  “On the matter of specificity and testing solution, [Ian] Barker wanted to know whether [Professor Richard] Nairn believed that over two hundred tests, referring to Kuhl's tests, which showed a negative reaction to adult haemoglobin would be sufficient to conclude that the anti-sera was not specific to adult haemoglobin. Nairn answered that it would depend on the test, a bad test is a bad test no matter how many times it is repeated.”  Another author quoted Richard Nairn: "Two hundred bad tests are poorer than one good test."