BACKGROUND The Defendant, Mr. Anderson is charged with Murder in the - - PDF document

People v. Anderson: 17 CR 677 P a g e | 1

BACKGROUND The Defendant, Mr. Anderson is charged with Murder in the First Degree. Part of the evidence against the Defendant is DNA evidence. On February 21, 2019 the defense received a report of additional DNA testing conducted by the government. The DNA test results are provided using STRmix DNA analysis likelihood ratio and accompanying verbal scale. As it relates to the findings in this case, the DNA typing results from the front right thigh pocket (item 15.3) are at least “29 times more likely” if the DNA originated from Waller, the Defendant, and an unknown, unrelated individual than if they originated from Waller and two unknown, unrelated individuals. Assuming the presence of Waller, the analysis provides “limited” support for the proposition that the Defendant is a contributor to this mixture. The quoted likelihood ratio and verbal scale are the basis for the Defendant’s

• bjection to this evidence being in front of a jury.

District Court, Adams County, Colorado Adams County Justice Center 1100 Judicial Center Drive Brighton, CO 80601 ▲COURT USE ONLY▲ THE PEOPLE OF THE STATE OF COLORADO Plaintiff, vs. DANTE MARQUIS ANDERSON Defendant Case No: 17 CR 677 Division: F Courtroom: 401 ORDER CONCERNING PRESENTATION OF DNA EVIDENCE

DATE FILED: July 28, 2019 1:36 PM CASE NUMBER: 2017CR677

People v. Anderson: 17 CR 677 P a g e 2|

Both sides agree that the Prosecution, as the party sponsoring the evidence bears the burden of production and proof as to whether the evidence is ultimately admissible. The Court agrees with the Defense that this Court is the gatekeeper in determining the admissibility of expert testimony. The Defense filed an objection to the DNA evidence and the Court held a hearing pursuant to People v. Shreck, 22 P.3d 68 (Colo. 2001), on July 19, 2019. At the hearing the Prosecution called Dr. John Buckelton and Sarah Miller from the Colorado Burau of Investigation while the Defense called Dr. Daniel Krane. Each side also presented documentary evidence. At the conclusion of the hearing, each side had an opportunity to brief the issues in writing. The Court has now considered all written filings, the testimony, exhibits, and arguments. LEGAL ANALYSIS Courts throughout the country (Illinois, Michigan, the Virgin Islands, New York, Wyoming, Texas, as well as the US Army Courts-Martial) have found the reliability of probabilistic genotyping and STRmix sufficient enough to admit the results in criminal

• trials. United States v. Christensen, slip copy, 2019 WL 651500 (C.D. Ill. 2019); People
• v. Muhammad, 2018 WL 4927094, at *2 (Mich. App. 2018).

Shreck and C.R.E. Standards In Shreck, the Colorado Supreme Court “explained that the admissibility of expert testimony is governed by C.R.E. 702 and C.R.E. 403.” Kutzly v. People, 442 P.3d 838, 841 (Colo., 2019). “[U]nder these evidentiary rules, admissibility of expert testimony requires that the testimony be relevant and reliable, and that the probative value of the evidence not be substantially outweighed by any of the countervailing considerations contained in CRE 403.” Id. In addition to making findings “as to the reliability of the scientific principles upon which the expert testimony is based and the qualifications of the witness giving that testimony,” this Court is also required to determine “the usefulness of such testimony to the jury.” Ruibal v. People, 432 P.3d 590, 593 (Colo. 2018)

People v. Anderson: 17 CR 677 P a g e 3|

Specifically, to determine whether to admit expert testimony, this Court must find that: “(1) the scientific principles underlying the testimony are reasonably reliable; (2) the expert is qualified to opine on such matters; (3) the expert testimony will be helpful to the jury; and (4) the evidence satisfies CRE 403.” People v. Rector, 248 P.3d 1196, 1200 (Colo. 2011) With regards to a specific finding of reliability of expert testimony, the Supreme Court explained that a trial court is required to “apply a liberal standard that only requires proof that the underlying scientific principles are reasonably reliable.” Kutzly at 841. “Determining if expert testimony is reasonably reliable requires considering the totality of the circumstances surrounding the proposed expert testimony and is not contingent on any specific list of factors.” Id. Because this Court must consider the totality of the circumstances, the Colorado Supreme Court has provided factors to consider which include: whether the technique has been tested, whether it has been subjected to peer review and publication, whether it has been generally accepted, its known or potential rate of error, and the existence and maintenance of standards controlling its operation. Id. at 841-42. However, the Court also explained that these factors “will be crucial in some cases but inapposite in others,” because “making a CRE 702 reliability determination will vary depending on the particular subject matter at hand.” Id. at 842. Additionally, if this Court finds relevance and reliability, it must then conduct a C.R.E. 403 balancing test. Id. Application of Shreck and C.R.E. Standards and Findings The Court, below, makes specific findings which support the conclusion that in this case, the results are relevant, the scientific principles underlying the testimony are reasonably reliable; the experts are qualified to opine on such matters; the expert testimony will be helpful to the jury; and the evidence satisfies CRE 403. The Court starts with the relevance of the results. The Court finds that the result “29,” itself is not relevant as it does not explain, to a factfinder, what it means. However, when accompanied by the verbal scale, it becomes relevant because it puts the result into

People v. Anderson: 17 CR 677 P a g e 4|

• context. Additionally, when put into further context, and compared to the remaining

likelihood ratios and their verbal scales, it provides the factfinder the entire picture. As such, the expert testimony will be helpful to the jury in understanding where in the spectrum of proof this evidence stands against the Defendant. The Court finds the underlying scientific principles of STRmix and the reporting of the quantitative value likelihood ratios reliable. STRmix has successfully undergone developmental validation. It has been internally validated to be used in over 40 labs in the United States, including the FBI laboratory and the Colorado Bureau of Investigation (CBI)’s laboratory. As it relates to its reliability, STRmix uses Markov Chain Monte Carlo to examine DNA profiles. The evidence proved that this is well-established statistical method. STRmix uses the available genetic information it has to differentiate true DNA donors from non-DNA donors within whichever DNA profile is in question. The Court finds the testimony of Dr. John Buckleton, credible and relevant for the determination of this motion. Dr. Buckleton is currently employed by the New Zealand Government Forensic Science Service and is currently assigned to the to the United States to assist with the transition to probabilistic genotyping. He has a number of degrees from the University of Auckland in New Zealand. The top two are a PhD in chemistry and a DOC in forensic science from the British Commonwealth. Dr. Buckleton has been responsible for publications as it relates to STRmix, and those publications have been in peer review journals. Additionally, he has given presentations surrounding DNA forensics and STRmix generally. Dr. Buckleton has been qualified as an expert in STRmix probabilistic genotyping many times before. As one of the co-creators of STRmix, and with

• ver 30 years of experience as a forensic scientist, the Court finds his experience sufficient

as an expert in STRmix generally and the statistics, probability, likelihood ratios, and verbal scales. Similarly, the Court finds the testimony of Sarah Miller, credible and relevant for the determination of this motion. She is responsible for the quality of the laboratory as it pertains to the DNA section. She’s also responsible for providing new technologies to

People v. Anderson: 17 CR 677 P a g e 5|

CBI, insuring the analysts are properly trained and competency tested as well as being responsible to ensure that CBI abides by all of standardization requirements. Finally, Ms. Miller is a qualified DNA analyst at CBI. She is a member of the American Academy of Forensic Sciences and a member of the American Society of Crime Laboratory Directors. She has a Bachelor of Science degree in biochemistry from the University of Missouri as well as Master of Science degree in biology from the University of Denver. She too has testified as an expert in the DNA analysis. Ms. Miller, the Program Manager and Technical Leader at the CBI, oversaw the successful internal validation for the use of STRmix across CBI’s five laboratories. The CBI has adopted SWGDAM’s recommendations in the reporting of likelihood ratios and the accompanying verbal scale. With regards to the testing that was performed in this case, it did not go beyond the limits established during the internal validation. As it relates to the analyst in this case, April Hrenya-Wood, Ms. Miller testified that

• Ms. Hrenya-Wood has successfully completed training for use of STRmix with her case
• work. However, if Ms. Hrenya-Wood is to testify, she will need to lay the foundation for

her own qualifications. For purposes of background in STRmix, the following is taken directly from People

• v. Superior Court (Dominguez), 239 Cal.Rptr.3d 71, 74, 28 Cal.App.5th 223, 228 (Cal.App.

4 Dist., 2018), which relied on (Chin et al., Forensic DNA Evidence: Science and the Law (The Rutter Group 2018) ¶ 11:7.), (See People v. Bullard-Daniel (N.Y.Sup.Ct. 2016) 54 Misc.3d 177, 42 N.Y.S.3d 714, 717, fn. 4.) and (Chin et al., Forensic DNA Evidence: Science and the Law, ¶ 11:7.). It provides an overview into the STRmix probabilistic genotyping program and provides the setting for this order. STRmix is a probabilistic genotyping program—i.e., one that comprises software, or software and hardware, with analytical and statistical functions that entail complex formulae and

• algorithms. It produces a likelihood ratio, which is generally

expressed as follows: a match between the suspect and the evidence is (x number) of times more probable than a coincidental match. Probabilistic genotyping has been described as particularly useful for low-level DNA samples

People v. Anderson: 17 CR 677 P a g e 6|

and complex mixtures since it can reduce subjectivity in the analysis of DNA typing results. (all internal marks removed from the original.) Associated with STRmix DNA analysis likelihood ratio, is the accompanying verbal

• scale. This verbal scale is based on the recommendations set by the Scientific Working

Group on DNA Analysis Methods, (SWGDAM). In 2018, SWGDAM formed an ad hoc group on Genotyping Results Reported as Likelihood Ratios. The Court acknowledges the Defense’s position that prior to 2018, there was a different value associated with certain likelihood ratios. Specifically, Dr. Buckleton admitted that his opinion had been, at one time, that likelihood ratios below 1,000 were uninformative. Nonetheless, his views are now aligned with the recommendations SWGDAM. Specifically, Dr. Buckleton agrees with SWGDAM that a likelihood ratio conveys the weight that should be given to the evidence and a likelihood ratio approaching one should not be labeled inconclusive based

• n its magnitude. First, however, the Court is analyzing the likelihood ratio and the

accompanying verbal scale which will be in front of a jury, not the one from several years

• ago. Second, this might be more relevant for the amount of weight a jury might ultimately

put on the evidence. The ad hoc working group was comprised of experts in the application of statistical principles to forensic evidence and forensic practitioners with expertise in the interpretation

• f mixed DNA specimens and probabilistic genotyping. The group also provided

recommendations on the reporting of genotyping results. With the exception of results which are deemed exclusionary, SWGDAM recommends that the numerical value of the likelihood ratio is to be reported as a quantitative estimate of statistical weight. As to the verbal scale, SWGDAM recommends a qualitative statement that conveys the degree of support to be given the numerical value

• f the likelihood ratio. The purpose behind this is to assist the reader/analyzer/fact finder

to understand the level of strength or lack thereof with the likelihood ratio. As indicated above, this is why the Court finds that the result “29,” would not relevant by itself, nor would it be helpful to the jury because it does not explain what it means. However, when

People v. Anderson: 17 CR 677 P a g e 7|

accompanied by the verbal scale, it becomes relevant because it puts the result into context. Additionally, when put into further context and compared to the remaining likelihood ratios and their verbal scales, it provides the factfinder the entire picture ensuring reliability. Furthermore, an average juror would understand that as the numbers in the likelihood ratio increase, so do the verbal scales. This is to say that an average juror would understand that a likelihood ratio within the “limited support” quadrant is not as convincing as “moderate support,” “strong support,” or “very strong support,” especially when presented with the associated likelihood ratio. According to SWGDAM, and which the Court accepts, the use of a verbal scale is supported across various disciplines of forensic science and has been adopted by the Association of Forensic Service Providers as well as the European Network of Forensic Science Institutes. SWGDAM recommends labs use the following verbal scale in reporting the qualitative value of likelihood ratios: (a) uninformative, (b) limited, (c) moderate, (d) strong, and (e) very strong. The numerical values associated with each are as follows:

Likelihood Ratio Verbal Scale 1 Uninformative 2 – 99 Limited Support 100 – 9,999 Moderate Support 10,000 – 999,999 Strong Support More than 1,000,000 Very Strong Support

The likelihood ratios above are a mathematical relationship between two explanations for the DNA profile. With probabilistic genotyping, there is an inclusionary hypothesis (associated with one party, for example, the prosecution) and an exclusionary hypothesis (associated with the other party, for example, the defense). The first hypothesis proposes that the suspect is the contributor to the DNA and the second hypothesis proposes that an unknown, unrelated individual is the contributor to the DNA. The likelihood ratio is the relationship between these two competing hypothesis. By way of further example, if the likelihood ratio is one, it is because both hypothesis are equally probable. If this is

People v. Anderson: 17 CR 677 P a g e 8|

the case, it is considered “uninformative,” as indicated in the chart above. The greater the number, the stronger the verbal scale support will be, which will be helpful to the factfinder. The Court finds that the Defense’s reliance on People v. Marks, 374 P.3d 518 (Colo.App. 2015) is misplaced as it dealt with DNA results which were either “inconclusive” or “no conclusion.” That would be more akin to the “uninformative” category from the chart above. The Court finds that the statistical or numerical conclusions are reliable and

• relevant. Similarly, the Court finds the assignment of the verbal scale is not arbitrary. The

Court finds that SWGDAM is comprised of the relevant scientific community, which includes experts in the application of statistical principals and forensic practitioners. The recommendations at issue here underwent public comment and were adopted by SWGDAM’s board. Here, consistent with this Court’s conclusion and order, SWGDAM recommends that the likelihood ratios in the “limited” range should be reported. Further, the Court accepts Dr. Buckleton’s testimony that there are three layers of conservatism built into the STRmix algorithm. This tends to cause STRmix to underreport the inclusionary/prosecutor hypotheses as a smaller likelihood ratios. As such, this has a tendency to the benefit of the defense hypothesis. As explained above, the Court finds that STRmix has been repeatedly tested and widely accepted by the scientific community and the results in this case are relevant. Nonetheless, pursuant to C.R.E. 403, although relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence. Here, the numbers in this case are low when compared in the chart above. This is at least an argument to exclude the results in this particular case. However, the Court finds that the number here, when attached with the associated verbal scale, does not affect STRmix’s reliability. The Court does not find that the probative value of the DNA STRmix evidence is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative

People v. Anderson: 17 CR 677 P a g e 9|

• evidence. It will be up the jury, after the sponsoring witnesses have been put through the

crucible of cross examination, to determine how much weight to put on the number “29” when compared to the rest of the chart. Additionally, any prejudice, confusion, or weight issues will be addressed by the appropriate instruction as to how to treat testimony from a court-recognized expert. CONCLUSIONS As articulated above in the specific findings, in this case, the results are relevant, the scientific principles underlying the testimony are reasonably reliable; the experts are qualified to opine on such matters; the expert testimony will be helpful to the jury; and the evidence satisfies CRE 403. Therefore, the Court will allow the testimony. However, as indicated above, if Ms. Hrenya-Wood is to testify, she will need to lay her own qualifications on the record. Ordered on July 28, 2019. BY THE COURT: ROBERTO RAMÍREZ DISTRICT COURT JUDGE