|People v Belle|
|2015 NY Slip Op 50663(U) [47 Misc 3d 1218(A)]|
|Decided on April 29, 2015|
|Supreme Court, Bronx County|
|Published by New York State Law Reporting Bureau pursuant to Judiciary Law § 431.|
|This opinion is uncorrected and will not be published in the printed Official Reports.|
The People of the State of New York
Wendell Belle, Defendant.
"Do the math" is a well-known phrase; it usually means figure it out correctly. In this case, the phrase is at the crux of the arguments made by the parties, and it is also relevant to many aspects of this decision. Defendant adds his name to the list of those who seek to preclude the introduction of DNA-related probability analysis evidence proffered by the People based upon a challenge to the statistical methodology employed in a mathematics-based computer program called the Forensic Statistical Tool (FST) developed and used by the Forensic Biology Unit of the Office of the Chief Medical Examiner (OCME). However, for the first time, more than one defense expert has utilized the recently revealed FST equations and mathematically determined the probability that defendant's DNA is part of a mixture of DNA found on a significant piece of physical evidence. The motion to preclude FST statistical evidence, or for hearing pursuant to Frye v. United States, 293 F. 1013 (D.C. Cir 1923), is denied in all respects.
Defendant is charged with attempted murder in the second degree and related crimes. He is accused of firing a gun at another individual in the vicinity of 161st Street and the Grand Concourse. Defendant is alleged to have fled that location, still carrying the gun. A witness to the shooting directed police officers to defendant, who ran up Sheridan Avenue, a street just one block away from the Grand Concourse. Video surveillance purportedly shows defendant running into an apartment building on Sheridan Avenue. He was apprehended as he exited that building shortly after he entered it. No gun was found in his physical possession. However, police officers recovered a loaded semi-automatic handgun which had been placed behind a radiator in a public area of the Sherman Avenue building. Ballistics analysis proffered by the People shows that ammunition at the scene of the shooting came from this gun.
A police department employee "swabbed" various areas of the gun for the presence of DNA, and sent the swabs to the Forensic Biology Department of the Office of the Chief Medical Examiner. Human DNA was detected on swabs from three parts of the gun, and [*2]DNA profiles were prepared for two of those swabs. The swab from the "slide grip grooves/slide release" area revealed a mixture of DNA from at least three individuals. A second swab from the "front strap/back strap/slide grip" area also revealed a mixture of DNA from at least three individuals. The OCME criminalist determined that the DNA in these mixtures was suitable for comparison with a known donor. A third swab from the "trigger/trigger guard" also disclosed the presence of DNA. However, that sample was not suitable for comparison.
The People moved to compel defendant to provide a sample of his own DNA, via a buccal swab. Defendant strongly opposed that application. A judge of this Court found probable cause to grant the People's request, and ordered defendant to provide a DNA saliva sample. An OCME criminalist developed defendant's unique DNA profile from that sample, and then compared the results of that profile with the profiles developed from the two DNA mixtures from the gun deemed suitable for comparison. The criminalist first found that defendant could not be excluded as a contributor to the DNA mixtures found on the front strap as well as on the slide grip.[FN1] Then, using the FST computer program, the criminalist calculated, statistically, the probability that defendant was that contributor. This is commonly referred to as the "likelihood ratio." The criminalist concluded it was 14.7 million times more likely that the DNA mixture from the "front strap/back side/side grips" included defendant's DNA in addition to the DNA from two other unknown contributors, than the mixture contained DNA from three completely unknown individuals. In terms of the mixture of DNA from the "slide grip grooves/slide release" area of the gun, the criminalist concluded it was 864,000 more likely that defendant was one of three contributors to the DNA found in that mixture than if the mixture came from three completely unknown contributors.
In late 2014, attorneys from the Legal Aid Society's DNA Unit "cracked the code," so to speak, in that they ascertained the specific mathematical formulas used by the OCME in the FST computer program. This enables experts now in their employ, including Dr. Eli Shapiro, who worked at the OCME when the FST was being developed, to conduct their own statistical analysis. In this case, such a statistical analysis of the DNA profile on the "slide grip groove/slide release" DNA mixture was conducted twice by the DNA Unit. The original defense analysis, apparently not done by Dr. Shapiro, found not only that the OCME probability ratio "appears to be incorrect," but that, statistically, it was only 5.8 times more likely that defendant was a contributor to that DNA mixture, and therefore not a very likely contributor at all. (Affirmation in Support of Motion, dated March 9, 2015). Based on that conclusion, on March 10, 2015, defendant moved to preclude the introduction of FST evidence in this case as unreliable.
On March 18, 2015, defendant withdrew that motion. He indicated, in substance, that the likelihood ratio reported had been incorrectly calculated. On or about March 25, 2015, defendant filed a new motion seeking, in essence, the same relief. Defendant revealed Dr. Shapiro now determined the likelihood ratio for defendant's DNA being in the mixture found on the "slide grip groove/slide release" as 93,200. (Affirmation in Support of Motion, dated March 25, 2015). Defendant annexed twenty-nine "excel" spread-sheets as exhibits to the affirmation, to show how Dr. Shapiro made this calculation. Dr. Shapiro did not challenge the FST likelihood ratio that defendant was statistically 14.7 million times more likely to have been a [*3]contributor to the DNA mixture found on another part of the gun, and apparently no one from the DNA Unit calculated any likelihood ratio for that other location. Yet, defendant still seeks to preclude FST statistical analysis results on both DNA samples obtained from the gun as unreliable, and asked in the alternative for a Frye hearing. The People filed their affirmation in opposition on April 17, 2015. Defendant filed a reply affirmation on April 27, 2015.
At the outset, this Court denies the motion for a Frye hearing to determine the "general acceptance" of the FST in the relevant scientific community. In determining whether scientific evidence is admissible under the Frye standard, a court must assess the "elemental question of whether the [challenged] techniques, when properly performed, generate results accepted as reliable within the scientific community." People v. Wesley, 83 NY2d 417, 422 (1994). The "relevant scientific community" includes individuals who have knowledge of the particular scientific or technical area involved in developing, studying, and understanding the techniques. Id. Under the Frye standard, "the particular procedure need not be unanimously indorsed' by the scientific community but must be generally acceptable as reliable.'" Id. at 423 (citing People v. Middleton, 54 NY2d 42, 49 (1981)). Where a particular technique is found to be generally acceptable by relevant experts, a Frye hearing is usually not ordered. Moreover, where other judges have made rulings that the same technique has been found generally acceptable in deciding litigated motions, those decisions can "aid in determining the admissibility of the proffered testimony," in a particular case, and a Frye hearing is not necessary. See People v. LeGrand, 8 NY3d 449, 457-58 (2007); People v. Garcia, 39 Misc 3d 482, 486-87 (Sup. Ct. Bronx County 2013).
Here, the Court is able to determine that the FST satisfies the Frye standard without conducting a hearing, based on the documentation provided by both parties in their weighty written submissions in this case and numerous prior rulings by other judges faced with the same request. It is, of course, a bit curious that defendant would seek such a hearing after notifying the Court that his own expert has performed a mathematical analysis and determined a significant "likelihood ratio" on one DNA sample in this case using the FST algorithm. His argument, in part, is that the FST is not reliable based on his own expert's analysis, because his expert's conclusion differs from the one reached by the OCME criminalist. This, says defendant, supports his claim that FST methodology is not generally accepted as valid in the scientific community. Logically, the opposite is true. If two graduate students in a math program taking a test use the same formula to solve a problem, based on the same data, and both "properly performed" the calculations, then their answers should be the same. If one makes an error in calculation, that does not make the formula invalid. Here, the different results have nothing to do with the methodology, and the defense calculation has nothing to do with the Frye test for admissibility. See Wesley, 83 NY2d at 422. Thus, defendant's ability to utilize the OCME algorithm, albeit by hand and not by computer, is one reason to deny a Frye hearing challenging the FST methodology.
Even had the defense not been able to perform its own analysis, the application for a Frye hearing would still be denied. The FST algorithm is a series of complicated mathematical equations which compute the likelihood that the DNA of a known individual is included within a complex, mixed-donor DNA sample. The algorithm includes mathematical assumptions based on a well-documented phenomena in DNA analysis called "allele drop-out." See State v. Ott, 80 So.3d 1280, 1285-86 (La. Ct. Of Appeal 4th Cir. 2012). This part of the algorithm is usually the central challenge raised by defendants seeking a Frye hearing, and is in [*4]fact still part of the challenge in this case, despite the fact that defendant's DNA Unit experts have utilized these same FST assumptions in their calculations and come up with reportable likelihood ratios they believe are correct under that algorithm. Indeed, defendant has informed the Court that in "two other cases . . . the DNA Unit has recomputed the FST likelihood ratio [with results] very close to the OCME [FST] result." (Affirmation dated March 18, at page 7).
The OCME has developed a formula which utilizes documented, verified allele drop-in and drop-out rates compiled after multiple, painstaking validation studies. One such validation study occurred in connection with the OCME developing a methodology for low copy/high sensitivity DNA analysis. Validation studies involving allele drop-in, drop-out rates were cited as a reason that low copy DNA evidence was found "sufficiently reliable" for admissibility under the standard set out in Daubert v. Merrell Dow Pharms, Inc., 509 U.S. 579 (1993). See, United States v. Morgan, 2014 U.S. Dist LEXIS 145687, October 2, 2014 (SDNY Marrero, J) The prediction of such allele drop-out rates is recognized to be part of overall validated DNA science, and "occurs when alleles from the principal DNA donors fail to appear in the DNA profile." Id.; see also Guy v. State, 2014 Tex App. LEXIS 11577, October 22, 2014 (Ct of Appeals, 3rd District) (Testimony of Dr. Chakrobarty at trial).
DNA likelihood ratios ascertained by computer programs utilizing demonstrable allele drop-in and drop-out rates as part of an algorithm have been found admissible by courts in other states. See Commonwealth v. Foley, 38 A.3d 882, 888 -90 (Pa. Sup Ct. 2012) (TrueAllele program, applying Frye test); State v. Bander, 208 P.3d 1242, 1253 - 55 (Court of Appeals Washington, Division One, 2009); see also People v. Wakefield, 2015 NY Misc LEXIS 306, February 9, 2015 (Sup Ct Schenectady County) (also TrueAllele). Defendant does not dispute any of this; he complains that the allele drop-in and drop-out rates used in the FST program do not satisfy the Frye standard. The FST itself has been the subject of "exhaustive validation procedures" and "numerous professional, scientific groups that have peer reviewed and accepted the validity of OCME's use of the FST." Garcia, 39 Misc 3d at 488. "OCME's empirical estimation of drop-in and drop-out rates have been presented at numerous national and scientific conferences." Id. And, the FST's overall methodology has been certified and accredited by the New York State Commission on Forensic Science and the National Forensic Science Technology Center. Id at 487-88.To this Court, this more than satisfies the FST's general acceptance in the scientific community as a reliable method of determining likelihood ratios.
In denying the application for an FST Frye hearing, this Court joins the vast majority of judges who have rejected the same request. These Courts have also ruled FST evidence admissible. The consistent reasons stated in each of these decisions is, as this Court also finds, that based on published articles and proffers made in voluminous motions and responses, FST calculations utilize nothing more than statistical formulas based on firmly-rooted mathematical principles and that the FST methodology utilizing the OCME's allele drop-in and drop-out rates has been repeatedly validated and approved by experts in the field. See People v. Williams, Ind. No 3445/2008 (Sup. Ct. Bronx County March 5, 2014, Moore,J); People v. Styles, 40 Misc 3d 1205 (A) (Sup. Ct. Kings County 2013, Donnelly, J.); People v. Garcia, 39 Misc 3d 482 (Sup. Ct. Bronx County 2013, Iacovetta, J.); People v. Caballero, Ind. No 10278/2011 (Sup. Ct. Queens County, April 15, 2013, Knopf, J); People v. Whetstone, Ind. No. 5728/12 (Sup. Ct. NY County, June 17, 2013, White, J.); People v. Wortham, Ind. No 3148/11 (Sup. Ct. NY County, November 21, 2012, Ward, J.); People v. Foster-Bey, Ind. No 9239/10 [*5](Sup. Ct. Kings County, May 22, 2013, Goldberg, J.); People v. Bowens, Ind. No. 336/11 (Sup. Ct. Richmond County, April 29, 2013, Rooney, J.); People v. Brissett, Ind. No. 847/10 (Sup. Ct. Bronx County, February 25, 2013, Webber, J.).
During the course of two or three years of challenges to the FST, two judges ordered Frye hearings. In New York County, a Court issued a forty page, unpublished, written decision on October 24, 2013 in the case of People v. William Rodriguez, Ind. No. 5472/2009 (Sup. Ct. NY County, Carruthers, J.). In that decision, the Court exhaustively summarized evidence provided by witnesses called by the People as well as the defendant. The Court noted that three of the four defense witnesses even agreed that "the concept of using a likelihood ratio incorporating the probability of drop-out is generally accepted within the scientific community." The fourth defense expert, Dr. Eli Shapiro, conceded he "was not as knowledgeable" in the "field of forensic statistics" as another witness, Dr. Adele Mitchell, who, while at OCME helped to develop the FST software and tested it. Dr. Shapiro was the lone witness at that hearing who stated that "no likelihood ratio program could ever assign a true statistical value to a complex DNA mixture." He endorsed another form of analysis called CPI, which no other expert called by either side believed could produce valid results where any type of allele drop-out has occurred. Following the hearing, the Court found the FST algorithm satisfied the Frye standard, and denied the motion to preclude admission of FST evidence.
In Kings County, a combined Frye hearing about the FST took place in the unrelated cases of People v. Andrew Peaks, Ind. No. 7689/2010, and People v. Jaquan Collins, Ind No. 8077/2010 (Sup. Ct. Kings County, Dwyer, J.). Here, the Court ruled that the FST did not satisfy the Frye standard. On November 7, 2014, the Court issued a brief oral ruling in anticipation of a more lengthy written decision. The Court adopted the "defense position [which] is that there is difficulty approving of the methods through which drop in and drop out rates and other stochastic effects are considered in the FST." (Minutes annexed as Exhibit E to defendant's motion).
Since the Peaks/Collins decision, defendants, including this defendant, have asked various judges to adopt that Court's ruling and preclude FST-based likelihood ratio evidence, or order additional Frye hearings. It appears no judge has yet agreed to grant either request. See People v. Graham, Ind. No, 2784/2012 (Sup Ct Bronx County February 20, 2015, Benitez, J); People v. Colon, Ind. No. 5073/2010 (Sup Ct New York County November 19, 2014, Carro, J).[FN2] There are motions to reargue pending in some cases where judges denied applications for Frye hearings prior to the Peaks/Collins ruling.
At bottom, this Court agrees with the reasoning of the many judges who have found that the FST program is neither new, novel, nor experimental. FST is based on long-established principles used to calculate statistical probabilities, and utilizes those principles in calculating where possible an easy to understand mathematical number measuring the likelihood of whether a particular defendant, already determined by DNA analysis to be a contributor to a DNA mixture in question, was statistically-speaking, that contributor. See People v. Rodriguez, Ind. No. 5471/09 (Sup. Ct. NY County, October 24, 2013, Carruthers, J., after Frye hearing). This Court is aware that there will always be some scientific dissenters, as there were when Galileo was imprisoned, and others who are merely skeptical and want to see more validation [*6]studies, and that these individuals do not endorse the methodology of the FST. Defendant has presented the opinions of these individuals in his motion papers, This Court is persuaded that the FST uses mathematical formulas that are generally accepted within the relevant scientific community. The assumptions for allele drop-in and drop-out rates have been sufficiently validated. And, the precise FST algorithms can be and have been utilized by defense experts to determine whether the actual results achieved by OCME criminalists are accurate. To the extent that Peaks/Collins holds otherwise, this Court respectfully declines to follow that opinion.
The second, and additional case-specific reason for denying the application to preclude all FST expert testimony in this case as per se unreliable is based on the defendant's analysis itself. Dr. Shapiro and at least one other unnamed expert working with the Legal Aid Society's DNA Unit have accomplished what no other defense expert, including, apparently, the experts who testified in the Peaks/Collins case, had been able to do: they applied the FST mathematical formulas, including allele drop-in, drop-out assumptions, and determined likelihood ratios utilizing those very principles. Moreover, even though the most recent FST-based analysis performed by Dr. Shapiro on the DNA from the "slide grip grooves/slide release" produced results that were significantly different than those used by the OCME, his results were nonetheless independently significant. His conclusion is that defendant is 93,200 more likely to be a contributor to the that mixture of DNA found on a gun ballistically-linked to an attempted murder. While his likelihood ratio is far less than the OCME criminalist's calculation, Dr. Shapiro's result, which is proffered by defendant to be correctly calculated, would be relevant and admissible to place that gun in defendant's hand.
Although defendant argues that this disparity alone shows that the FST formula produces unreliable results, that position is difficult to accept. The People contend Dr. Shapiro must have made an error in calculation. They point to the fact that the first FST-based likelihood ratio, made by an unnamed expert in the DNA Unit and which found defendant not to be a likely contributor to that DNA sample, is conceded to have been incorrectly calculated. This Court does not find either Dr. Shapiro's calculation or the OCME's calculation to be inaccurate; it only finds that one of them must be. They used the same exact data and used the same algorithm. That has nothing to do with the reliability of the method; it is merely proof that one of these experts made a calculator error.[FN3]
What has now been set up by the defendant in this case is a potential "battle of the experts" over the results obtained from using the same scientific and mathematical procedures. Defendant seeks to preclude all of these results from consideration by a jury. The attack is no longer on the method; it is on whether it was "properly performed." Wesley, 83 [*7]NY2d at 422. That is a question of weight, not admissibility. As Justice Benitez found, "the prevailing view of courts that have considered this issue is that FST is generally accepted in the relevant scientific community and that disputes concerning . . . the conclusions reached, and any potential weaknesses or shortcomings perceived by the defense, as with much scientific evidence, bear on the weight as opposed to the admissibility of the evidence. These are subjects to be explored and challenged through cross-examination and/or defense expert testimony and ultimately evaluated by the jury." People v. Graham, supra. That is precisely what should happen in this case, especially since the defendant has the ability to call his own experts to challenge the actual likelihood ratios calculated by the OCME. See Guy v. State, supra ("jury was free to assign weight to these various expert opinions" about probability that the defendant's DNA was on evidence, where allele drop out occurred).
The defense calculations themselves provide another compelling reason to deny the motion to preclude FST-based likelihood ratio calculations. Assuming, arguendo, defendant proffered that his original calculation finding him not to be a likely contributor to a mixture of DNA found on the gun was correct, that is certainly something the defense would seek to introduce in this case. Perhaps in a future case, FST analysis done by a defense expert will challenge the OCME's likelihood ratio calculation that a defendant is a even contributor to a DNA mixture. A defendant should be able to present their own FST expert at a trial.
In fact, the People point out that in at least two cases since the Peak/Collins ruling, attorneys from the Legal Aid Society's DNA Unit have called criminalists from the OCME specifically to testify for the defense about likelihood ratios calculated by the FST which produced results favorable to other defendants. In one case, the OCME witness was called by an attorney from the Legal Aid Society DNA Unit who had argued against the admission of FST evidence and presented witnesses in Peaks/Collins who testified that FST was not generally accepted in the scientific community as reliable. Just weeks after the Peaks/Collins ruling, the same attorney asked that an OCME criminalist be qualified as "an expert in the field of DNA analysis and statistical analysis of DNA profiles." That criminalist then testified about the FST likelihood ratio calculated in that case. People v. Barnett, Ind No. 3533/2012 (Sup. Ct. Kings County, DelGuidice, J; Minutes annexed as Exhibit One to People's Response).
The Kings County District Attorney's Office subsequently moved to reopen the Peaks/Collins hearing, based, in part, on the blatant inconsistency in position. That Court declined the application to reopen the hearing, but clarified its ruling. In a sweeping statement based on its ruling following the Frye hearing, that Court said that henceforth in all cases before that judge, "neither side will be introducing FST evidence." (Minutes dated January 5, 2015, People v. Collins, Ind. No 8077/2010 at page 4). That ruling is consistent with the Court's initial Peaks/Collins finding that it was specifically adopting the defense position that the FST is not sufficiently recognized within the scientific community to be reliable.
This Court is somewhat troubled by the same attorneys arguing in one case that FST is generally accepted in the scientific community and therefore admissible to exculpate their clients, but is not admissible to incriminate their clients because it is unreliable (Reply Affirmation dated April 27, 2015 at pages 14- 18). In terms of a Frye analysis, when the defense is based upon scientific testimony, the test for admissibility has to be the same as it is for the prosecution — which is apparently the holding in Peaks/Collins. This Court finds FST reliable, and as far as this case is concerned, that means both sides may introduce their FST results and explain to the jury that defendant is either 93,000 or 834,000 more likely than not to be a [*8]contributor to the contested DNA sample, unless one side finds a calculation error.
Finally, defendant also asks this Court to rule that the probative value of FST testimony in this case is outweighed by a "prejudicial effect" on the defense. That application is denied as well. This is not a DNA-based prosecution. The People have an eyewitness to the shooting. They have the gun, which was ballistically matched to the shooting. The gun was found inside a building the defendant entered, and left, shortly after the shooting. The DNA found in substantial form on at least two areas of that gun and deemed suitable for comparison and which has been interpreted to mean defendant cannot be excluded as a contributor to those sample is additional, relevant, admissible evidence. Comparison by experts from the OCME as well as the Legal Aid Society's DNA Unit conclude that the likelihood that defendant's DNA is on that gun is very high. Given the other evidence, DNA likelihood ratio testimony is extremely probative of establishing defendant's guilt of possessing that gun and corroborating eyewitness testimony that he not only possessed it, but fired it at another person. The Court finds no prejudice in the introduction of such evidence based on the FST in this case.
WHEREFORE, defendant's motion for a Frye hearing and to preclude the admission of FST scientific evidence is denied in all respects.
This opinion constitutes the Decision and Order of the Court.