People v. Lopez , 23 N.Y.S.3d 820 ( 2015 )

OPINION OF THE COURT

Ralph A. Fabrizio, J.

Defendant is indicted for criminal possession of a weapon in the second degree and related crimes. After his arrest, the gun in question was “swabbed” to determine whether any DNA suitable for comparison had been left by an individual who touched it. A mixture of DNA from multiple sources was discovered on the gun and DNA profiles were developed of contributors to the mixture. Defendant’s own unique DNA profile was obtained and compared with the DNA mixture profiles. A criminalist from the Office of the Chief Medical Examiner (OCME) determined defendant could not be ruled out as a contributor to that mixture. Then, using the Forensic Statistical Tool (FST), the criminalist formulated a likelihood ratio that defendant was a contributor. The finding: it is “approximately 828 million times more probable [that] the sample originated from [defendant] and one unknown person, than if it originated from two unknown, unrelated persons.” That gave “very strong support” for the criminalist to conclude that defendant’s DNA was in that mixture on the gun.

Defendant now moves, as other defendants have done on many occasions in many other courtrooms, to preclude the introduction of expert testimony by the People regarding the use of the FST to calculate the probability or likelihood ratio that his DNA is on the gun. In the alternative, he seeks another hearing on the admissibility of this evidence, pursuant to Frye v United States (293 F 1013 [DC Cir 1923]). Defendant argues, inter alia, that the relevant scientific community has not yet found the FST algorithm, including its use of statistically-predicted allele drop-in and drop-out rates, generally acceptable as a reliable method of statistical analysis in DNA mixture cases. These applications are denied.

Earlier this year, this court ruled on a motion seeking the same relief, based on some of, but not all of, the same grounds. (People v Belle, 47 Misc 3d 1281 [A], 2015 NY Slip Op 50663[U] [Sup Ct, Bronx County, Apr. 29, 2015].) At the time, the court was aware of a series of rulings made from the bench by a judge of coordinate jurisdiction in Kings County on the admissibility of FST evidence in two cases following a Frye hearing; since that time, a 32-page written decision has been published, amplifying the ruling made on the record. (People v Collins, 49 Misc 3d 595 [Sup Ct, Kings County 2015].) The Collins decision, finding FST evidence could not be admitted, was issued following a years-long Frye hearing concerning the admissibility of FST evidence, as well as about the admissibility of “low copy” DNA evidence.¹ The current motion embraces that decision in terms of the FST ruling and urges this court to follow it. Respectfully, this court declines to do so.

In Belle, this court found it unnecessary to conduct another Frye hearing, relying in part on the evidence presented in an already concluded Frye hearing specifically concerning admissibility of FST evidence. (People v William Rodriguez, Sup Ct, NY County, Oct. 24, 2013, Carruthers, J., indictment No. 5471/ 2009.) During that hearing, all witnesses called by the People and the defense rendered opinions about the general acceptance of FST methodology in the relevant scientific community. The validity of including allele drop-in and drop-out rates in the FST algorithm formed a major part of the opinion testimony at the hearing. In its lengthy decision, the court noted that only one expert claimed that using allele drop-in and drop-out rates to calculate probability ratios in DNA mixtures was not based on generally-accepted science. That witness was Dr. Eli Shapiro. Although the specific drop-in and drop-out rates used in the FST were considered proprietary to the developers of that program, and were unknown at the time, that fact was irrevelant to the court’s ruling that the logic behind the FST algorithm, including the use of specific allele drop-in and drop-out rates determined through the use of Bayesian statistical analysis, was generally accepted as reliable in the relevant scientific community.

An integral part of the defense argument in Belle involved the fact that following the testimony at the Collins hearing, the Legal Aid Society had been able to figure out the actual drop-in, drop-out variables used in FST calculations as well as the rest of the FST algorithm. The defense had calculated their own FST probability ratio to determine the likelihood that defendant Belle was a contributor to a DNA sample found on the gun he was alleged to have possessed. At first, the defense calculations produced results showing that Belle could not have been a likely contributor to the DNA mixture. Because of this, the defendant argued that the actual variables used in the FST algorithm were themselves flawed, and sought preclusion of the OCME’s FST results as unreliable. That motion was withdrawn after the defense recalculated the FST-based likelihood ratios. Based on the recalculation, performed by Dr. Eli Shapiro, who now works as a defense consultant on DNA cases, the defense actually agreed with the OCME’s conclusion that it was highly likely Belle’s DNA was part of the mixture, since the defendant’s own calculations showed it was 93,000 times more likely that Belle’s DNA was part of the mixture.

This court denied the Belle motion, in part, because the calculation performed by Dr. Shapiro showed the efficacy of using the specific allele drop-in and drop-out rates that had been discovered to be in the FST algorithm. Dr. Shapiro disputed the math, and found a different likelihood ratio. However, the ratio he calculated was significant in and of itself. The defense argued that the different mathematical results showed FST was not reliable. It seemed just a matter of common sense for this court to conclude the opposite was true, and that this could be explained away as a calculator error. It was a matter of weight, not admissibility. The defense had a witness to offer a challenge to a mathematical result, and there would now be a garden-variety jury issue to determine which calculation was done correctly.

Citing the written decision in Collins, defendant now takes issue with the Belle decision,² opining, as does the Collins court, that it “jumps over those aspects [involving allele drop-in and drop-in rates] without discussion.” (Collins, 49 Misc 3d at 628.) This is quite a baffling assertion, since there was not only a discussion in Belle of the acceptability of allele drop-in and drop-out rates in likelihood ratio calculations in the Rodriguez case, but also citations to cases from other courts in which the use of these specific stochastic effects had been recognized as acceptable in the relevant scientific community in likelihood ratio analysis. (See State v Ott, 80 So 3d 1280, 1285-1286 [La Ct of Appeal, 4th Cir 2012]; Guy v State, 2014 WL 5423760, 2014 Tex App LEXIS 11577 [3d Dist, Oct. 22, 2014, No. 03-12-00466-CR] [both cited in Belle].) To the extent that defendant suggests that a court must explain the science in a decision discussing whether a certain scientific procedure is generally accepted in the relevant scientific community, such a suggestion misstates the court’s role in a Frye application.

In any event, the science behind DNA analysis, including the specific scientific terms, is not only the subject of scientific journals, but was published as a lengthy, technical “appendix” to the majority opinion recognizing the admissibility of DNA evidence to prove a defendant’s guilt in People v Wesley (83 NY2d 417, 430-435 [1994]). That appendix explains the meaning of the terms relevant to DNA analysis itself: a “genome” is the “individual’s complement of DNA”; the “genome” is contained in an “individual’s chromosome”; a “gene” is a “DNA sequence”; each “gene” is situated at a “specific locus on a specific chromosome”; and there are “alternative genes” that are also found “at a particular locus [and] are called alleles.” (Id. at 430-431.) One of the foundations of DNA analysis concerns “polymorphism” (id. at 431-433), and this relates to the alleles. (Promega Corp. v Life Tech. Corp., 773 F3d 1338, 1341-1342 [Fed Cir 2014].)

“Although an individual has, at most, two alleles at a given locus, many different alleles can exist for the same locus within a given population. When multiple alleles exist at a particular locus, the genetic variant is referred to as polymorphism. Polymorphisms are simply the genetic differences among members of [the genetic] population, and are caused by the variations in base sequences in the DNA in the population.” (Wesley, 83 NY2d at 431-432.)

This science has not changed. In fact, DNA analysis in forensic cases is universally recognized as accepted in the scientific community, and courts even take judicial notice of this fact. (See United States v Beasley, 102 F3d 1440, 1448 [8th Cir 1996].)

Notably, all of the witnesses who testified at the Collins and Rodriguez hearings agree about basic DNA science, and polymorphism. In this case, as in Belle, the DNA profiles of potential contributors to the mixtures found on the guns was detected by methods of DNA analysis unchallenged by these defendants. The challenge is to the population-study based probability calculation of the likelihood that DNA in that sample belongs to the defendant. In Wesley, the Court also found that the statistical conclusions reached by the DNA examiners about the likelihood that the murder victim’s DNA was on the defendant’s clothing were properly admitted at trial. Significantly, the majority found that “challenges to the population studies ... to estimate the probability of a coincidental match go not to [the] admissibility, but to the weight of the evidence, which should be left to the trier of fact.” (83 NY2d at 427 [citation omitted].) The appendix notes that “population genetic studies” had been performed over a “two-year period” using DNA “from approximately 900 unrelated individuals.” (Id. at 434.) The Court noted that the population study evidence was actually “reduced by a factor of 10,” to eliminate a possible challenge to the validity of the expert’s conclusion. (Id. at 435.)

So there you have it. That is the science behind a DNA likelihood analysis, all based upon methods found generally accepted within the scientific community 21 years ago in Wesley. It includes calculations about whether certain alleles are likely to appear at a particular locus. As the Rodriguez decision explained, “Drop-out occurs when an allele that is actually known to exist at a particular locus in a DNA sample is not found in the analysis . . . Drop-in occurs when an allele is detected during analysis that is not known to the person or persons contributing to the sample” (Sup Ct, NY County, Oct. 24, 2013, Carruthers, J., indictment No. 5471/2009, slip op). To the extent that any current defense experts challenge basic DNA science as it relates to polymorphism based on analysis of population-predicted allele location, that science was ruled acceptable despite the testimony of experts who opined otherwise in Wesley. Moreover, the use of population studies using known samples—precisely what the OCME did here, with adding observable stochastic effects, such as allele drop-in and dropout rates for particular loci—was a methodology approved of in Wesley as well. The Court of Appeals majority was unconcerned with any of this. In fact, the majority opinion “noted that novel scientific evidence may be admitted without any hearing at all by the trial court.” (83 NY2d at 426.) This court continues to believe that FST-based opinion evidence is neither new nor novel, and completely admissible.

Moreover, this court also continues to rely on the testimony and findings in Rodriguez as settling the questions posed by the defendant. DNA science is thoroughly addressed in the 42-page decision in Rodriguez. The Frye hearing in that case concerned the same defense challenge raised in Belle and now raised in this motion. Eight witnesses testified about the general acceptance of an algorithm employing calculated allele drop-in and drop-out rates, as the FST does, in the relevant scientific community. In that case, the defense challenged the general acceptance of the FST because of “unfounded assumptions concerning allele drop-in and drop-out.” All but one expert, including those called by the defense, agreed that using predicted allele drop-in, drop-out rates was generally accepted as valid in the relevant scientific community. Thus, the court reported that at the conclusion of the Frye hearing, “the prosecution and the defense now concur that the concept of using a likelihood ratio incorporating the probability of drop out is generally accepted in the scientific community” (Sup Ct, NY County, Oct. 24, 2013, Carruthers, J., indictment No. 5471/ 2009, slip op). This court believes Rodriguez was correctly decided.

The written decision in Collins does not convince this court that the conclusion it reached in Belle was incorrect. In Collins, the court relied on the oft-quoted phrase from Chief Judge Kaye’s concurrence in Wesley (83 NY2d at 439), that determining whether a particular scientific method meets the Frye standard involves merely “counting scientists’ votes.” While that opinion concurred in the result, it did not concur in the majority decision that either the DNA evidence or the scientific study evidence was properly admitted at trial. Thus, to the extent that the concurring opinion stated that DNA evidence did not pass the Frye test, it may more properly be called a dissent for the purposes of this type of motion. The majority even noted that “new and more stringent requirements that would be added under the [Frye] test proposed by the concurring opinion.” (Id. at 426.) The concurring opinion rejected the position that a Frye determination related to admissibility could be based upon experimentation and published articles by scientists who actually developed and implemented the challenged methods. (Id. at 439 [concurring opinion].) It also equated the appearance of experts who testified that the methods were unreliable as a scenario where “controversy rages.” (Id.) Those were things with which the majority in Wesley was unconcerned. These are the same arguments raised by the defendant in this motion. This court is equally unconcerned.

To the extent that “counting heads” is an appropriate method of determining admissibility under Frye, even the Wesley concurrence does not call for blind acceptance of all experts’ opinions. The concurring opinion viewed several witnesses’ testimony as irrevelant because “[n]one . . . was expert in Forensic DNA analysis.” (Id. at 438.) Even Frye (293 F at 1013-1014) discussed the need for testimony from specific “physiological and psychological authorities” in order to find expert testimony concerning the results of a “deception” test to be admissible. Thus, at a Frye hearing, a court should not just count the scientist’s head, but should consider what expert hat can be worn on that head. This is the assessment that is supposed to be made at every trial when an expert testifies. The jury is instructed to consider, inter alia, “the qualifications and the believability of the witness” in determining “whether or not to accept [their] testimony.” (CJI2d[NY] Expert in General— Expert Witness.) Notably, there are even different approved jury instructions related to different areas of proffered expertise, such as Crime Victim Syndrome and Drug Trafficking. (See CJI2d[NY] Expert.)

This court has read the excerpts of the testimony of each of the witnesses proffered by the People and the defense in the Collins hearing. It has also read the qualifications of each witness relevant to their rendering a decision about the acceptance of FST in the relevant scientific community. This court is not persuaded that all “votes” should have been given the same weight, based on the different “qualifications and believability” of these witnesses.

As the People correctly note, and even defendant agrees, in Collins, the court relied extensively on the testimony of Dr. Eli Shapiro to support its conclusion that the FST methodology and algorithm, including the use of its assumed allele drop-in and drop-out rates, was not generally accepted within the relevant scientific community. However, Dr. Shapiro has no training or experience in statistical analysis or the use of probability software. He testified that he lacked any basis to understand the statistical algorithm developed in part by Dr. Adele Mitchell, who testified at the same hearing. That statistical analysis is at the heart of the FST program and its development. Dr. Shapiro, who never performed any statistical analysis or even had the ability to check whether Dr. Mitchell’s allele drop-in or drop-out probability analysis produced viable results, nonetheless was credited as an expert in precisely these areas.

As noted, this court agrees with the finding in Rodriguez not to give weight to Dr. Shapiro’s opinion. Indeed, he stood alone in that hearing in reaching the conclusion that no likelihood ratio program could ever be developed that would assign a “true statistical value to a complex DNA mixture.” Even now, this court is uncertain what Dr. Shapiro’s current opinion might be about the FST, and his basis of knowledge to render such an opinion. In his reply papers, defendant discusses the types of factors he believes are relevant to developing a statistical model to analyze the probability that an individual’s DNA is part of a mixture. (Reply mem dated Oct. 9, 2015 at 10-11.) At no time does he argue, as his own witness opines, that no such program could be developed.

A court is permitted to rely on conclusions reached in other Frye hearings by judges who addressed the same issue presented for its resolution. (People v LeGrand, 8 NY3d 449, 457-458 [2007].) This court agrees with the decision in Rodriguez that Dr. Shapiro’s vote on the acceptance of FST statistical analysis lacked sufficient credibility to be weighed in making a finding precluding the admission of such testimony.³ Since it appears that his “vote” was essential to the conclusion reached in Collins, this court does not find that ruling persuasive. Moreover, since the expert witnesses who testified for the People, and whose expertise in this specialized area has never been the subject of any dispute, actually performed the probability analysis, came up with conservative allele drop-in and drop-out rates based on their own experimentation, and were cognizant of each and every one of the stochastic effects that are well-documented in DNA science when drafting their algorithm, this court also believes the Rodriguez court correctly concluded that the overall FST methodology is based on principles that are generally accepted as reliable in the relevant scientific community.

In the six months since the Belle decision, and the four months since the Collins written ruling, motions challenging the admissibility of FST evidence continue to be made. Certainly, that is a path that defense attorneys have every right to pursue. However, this court is unaware that any court, anywhere, other than Collins, has ruled FST evidence, or similar Bayesian-based probability programs with allele drop-in or drop-out rates included, to be inadmissible. Significantly, there has been one very persuasive New York court ruling made since the Collins decision setting out detailed, sound reasons for trial judges to not follow that ruling and follow Rodriguez.

In People v Debraux (50 Misc 3d 247 [2015, Kahn, J.]), the court found that the use of allele drop-in and drop-out rates as well as other stochastic effects in calculating probability ratios was an entirely valid methodology that is generally accepted in the relevant, credible scientific community. That court declined to follow Collins for a number of reasons. One of them also involved what appears to be the “believability” of defense witness opinions in that case. In Collins (49 Misc 3d at 612), the court indicated that it was swayed by the testimony of the “impressive defense witnesses.” In Debraux, the court noted that three of those witnesses all worked for one laboratory in Texas, and one other witness ran a defense consulting firm. The Debraux ruling was also critical of Collins in its interpretation of and application of the Frye standard itself. The court recognized that the Collins court reached its “conclusion principally on the basis of [its] agreement on the merits with the arguments of the defense experts criticizing the manner in which stochastic effects at each locus are assessed by the FST, and therefore rejected the FST technique overall (see People v Collins, 49 Misc 3d at 613-620).” (People v Debraux, 50 Misc 3d at 255.) As the Debraux decision correctly pointed out, “in a Frye analysis, it is not the role of the court to evaluate and appraise the merits of the scientific technique under consideration . . . [and] Verify[ ] the soundness of a scientific conclusion.’ (Parker v Mobil Oil Corp., 7 NY3d 434, 446-447).” {Id. at 259 [citations omitted].) That court adopted the Rodriguez ruling and also declined to order another Frye hearing. This court agrees with the analysis and the result reached in Debraux.

There is also a decision from a New Jersey court addressing admissibility of both low copy DNA and FST evidence obtained pursuant to a request by New Jersey authorities to the OCME to perform the analysis. In State v Rochat (Sup Ct, NJ, Bergen County, Aug. 6, 2015, Steele, J., index No. 13-07-1002-1,⁴ Iv denied Sup Ct, NJ, App Div, Sept. 24, 2015), the Bergen County prosecutor’s office sought to introduce testimony of OCME criminalists in a New Jersey homicide case. Blood evidence swabbed from the kitchen sink of the crime scene was sent to New York City for low copy DNA analysis. The criminalist involved determined that the defendant could be the source of the DNA in the blood. It appears the FST was used to calculate a likelihood ratio. The defense sought to preclude admission of the DNA evidence or have the court order a Frye hearing. The motion submitted by the defendant took the position that the FST was not generally accepted in the scientific community. The court in that case, using the correct Frye standard, ruled the evidence admissible without a Frye hearing. The Rochat court referred to the myriad of New York State court decisions which found the FST methodology not at all novel. The court was not persuaded by affidavits from three proffered defense experts, including Dr. Eli Shapiro, that there were “specific problems with . . . [the] FST.” The New Jersey court considered the Collins ruling and was also not persuaded by its analysis. Notably, Dr. Shapiro told the court in his own submission that he continues to test the FST program and that in one case he found that false-positive results were generated. No mention is made in the current motion about Dr. Shapiro’s apparent continuing work with the FST.

In sum, based on a thorough consideration of the arguments raised in the motion papers, the testimonial exhibits submitted from the Collins case, its consideration of the testimony of the various witnesses who have opined about the general acceptability of the FST methodology in the relevant scientific community, including the specific allele drop-in and drop-out rates used in the algorithm, the Rodriguez, Debraux and Rochat rulings, and majority decision of the Court of Appeals in Wesley, this court once again declines to order a Frye hearing. For these reasons, as well as reasons stated in the Belle decision, the motion to preclude the introduction of FST testimony is denied.

Defendant also argues that should this court find the FST calculations based on a methodology generally accepted in the scientific community, the evidence should still be ruled inadmissible based on a pure evidentiary weight analysis. Citing the decision in People v Graham (128 AD3d 572 [1st Dept 2015]), defendant argues that his motion to preclude FST testimony should be granted because the probative value of such evidence is outweighed by the potential for prejudice in this case. His reliance on Graham is curious for a number of reasons. First, that decision did not involve the FST. In fact, it appears that no quantitative statistical analysis was performed by the OCME, probably because the FST did not exist at the time of the original testing. The First Department found that the jury’s verdict that the defendant possessed a gun was against the weight of the evidence. (Id. at 572-573.) The court further found that the DNA evidence in that case “did not connect defendant [to] the pistol.” (Id. at 573.) Multiple people were found at the scene of a shooting. Two guns were found at the scene. The criminalist testified that a “codefendant’s DNA conclusively matched DNA found on the pistol’s trigger.” (Id.) Defendant Graham’s DNA was part of a “mixture of DNA from at least three persons.” (Id.) The criminalist testified “only that defendant ‘could’ have been a contributor to that mixture.” (Id.)

Here, of course, the FST allows the criminalist to testify to more than a vague possibility. The likelihood ratio found—that defendant is 828 million times more likely to be a contributor to the mixture—is relevant and has substantial weight. It is also the type of statistical evidence that was recognized as admissible by Wesley even without a Frye hearing. Moreover, it is only part of the other strong evidence in this case. According to the sworn statements in the felony complaint, the gun in this case was recovered during the execution of a search warrant in a Bronx apartment. Defendant made a statement that he lived in the apartment. When the police arrived to execute the warrant, an officer on the team states she saw defendant toss the gun from a window in that apartment. The People served notice of a statement made by the defendant to the police admitting that he threw the gun out that window. The FST likelihood ratio which provides “very strong support” that defendant held that very gun is not only probative as additional evidence that defendant possessed the gun, but because it is based on a methodology which has been generally accepted by the scientific community, there is nothing prejudicial in its introduction at this trial.

Finally, defendant seeks an order compelling the People to provide discovery of the “raw data” compiled by the OCME in conducting the DNA analysis, as well as the “forensic statistical tool program” itself. No discovery of any computer program is necessary. The Legal Aid Society already has the algorithm used in the program. To the extent they claim it would be easier to perform the calculations with the actual program software, the computer program itself is proprietary and the court is not ordering its disclosure. In fact, according to court filings in New Jersey, Dr. Shapiro is using the algorithm and possibly writing an article about his findings. As far as the “raw data” is concerned, the People contend that no raw data exists in this case. There’s no explanation about why it does not exist here when in other cases the People have stated that the OCME has such data. Thus, the court orders production of any “raw data” already in the possession of the OCME. This is consistent with this court’s earlier ruling in People v Javon Grant (Sup Ct, Bronx County, indictment No. 604/2013, decided June 12, 2014), and annexed to the defense motion in this matter as exhibit J.

1 . Recently, an appellate court in Maryland ruled that low copy DNA results were admissible in that state under a Frye analysis. (Phillips v State, 2015 WL 6472286, 2015 Md App LEXIS 154 [Ct Special Appeals, Oct. 27, 2015, No. 456].) In Maryland, the Frye standard is met by a showing that the scientific opinion evidence is “generally accepted as reliable within the expert’s scientific field.” (Reed v State, 283 Md 374, 381, 391 A2d 364, 368 [1978].) The Maryland appellate court referred to the Collins decision as holding that low copy DNA evidence is not “sufficiently reliable . . . under New York’s version of the Frye standard,” and therefore did not follow it. (Phillips v State, 2015 WL 6472286, *10 n 10, 2015 Md App LEXIS 154, *30 n 10.)

2 . After that ruling, defendant Belle accepted a sentence offer on a plea to the indictment made by a different judge. When defendant Belle pleaded guilty, it is believed he waived all appellate rights. Thus, the FST decision will not be the subject of appellate review.

3 . Dr. Shapiro testified as a defense trial witness before this court in People v Garcia (Sup Ct, Bronx County, indictment No. 2650/09). In that case, in a published opinion, another judge in this courthouse ruled that evidence of low copy DNA testing as well as the FST likelihood ratio calculated by the OCME was admissible without a Frye hearing. (People v Garcia, 39 Misc 3d 482, 487-490 [Sup Ct, Bronx County 2013].) Dr. Shapiro gave opinion testimony for the defense at the trial. In terms of the FST results, he testified he had no training in statistical analysis and, although he had worked at the OCME, had never used the FST. Although the bulk of his testimony challenged the low copy DNA profile results testified to by the People’s witness, he also testified, in substance, that he believed it important to calculate likelihood ratios, and showed some familiarity with how different allele drop-in and drop-out rates can be found at different DNA sites. The case is currently on appeal.

4 . The decision is annexed as exhibit 6 to the People’s response to the current motion.