People v. Coy

• Decision Date: 07 August 2003
• Docket Number: No. 238112.,238112.
• Parties: PEOPLE OF THE STATE OF MICHIGAN, Plaintiff-Appellee, v. LAURENCE DEANE COY, Defendant-Appellant.
• Court: Court of Appeal of Michigan — District of US

Page 1

PEOPLE OF THE STATE OF MICHIGAN, Plaintiff-Appellee, v. LAURENCE DEANE COY, Defendant-Appellant.

No. 238112.

Court of Appeals of Michigan.

August 7, 2003.

Appeal from Calhoun Circuit Court, LC No. 98-001925-FC.

Before: Markey, P.J., and Saad and Wilder, JJ.

PER CURIAM.

Defendant was originally convicted in 1998 of second-degree murder, MCL 750.317, and sentenced to forty to sixty years' imprisonment. This Court reversed defendant's conviction and remanded for a new trial. People v Coy, 243 Mich App 283; 620 NW2d 888 (2000). On retrial, a jury convicted defendant of voluntary manslaughter, MCL 750.321; he was sentenced to twenty to thirty years' imprisonment as an habitual offender, third offense, MCL 769.11. He now appeals by right asserting that the trial court erred by admitting evidence and by not granting an adjournment to permit additional DNA testing. Defendant also claims that his sentence is disproportionate. We affirm.

Defendant first argues that the trial court erroneously admitted evidence of statistical analysis of DNA profiles developed from mixed blood samples found at the crime scene. Specifically, he claims that the methods employed interpreting the mixed DNA result samples are not sufficiently appropriate and scientifically acceptable to justify the admission of the test results. We disagree. A trial court's decision to admit evidence is reviewed for an abuse of discretion. People v Herndon, 246 Mich App 371, 406; 633 NW2d 376 (2001).

In our prior opinion, we found that the admission of evidence that defendant's DNA profile was consistent with DNA profiles from mixed blood samples was plain error warranting reversal because no testimony illuminated the statistical significance of a potential match. Coy, supra at 301. Neither defendant nor the victim could "be excluded as a possible contributor" to the mixed samples recovered from a broken knife blade found in the victim's bedroom and on the victim's bedroom doorknob. Id. at 293. Anita Matthews, a forensic serologist and the associate director of forensic identity testing at Laboratory Corporation of America in North Carolina (Lab Corp) testified that "`once we determine that two samples could have come from the same source then we could calculate a statistical estimate to give a likelihood of how common or how rare it is to find that set of characteristics in another individual.'" Id. at 293-294. However, Matthews did not offer such testimony because Lab Corp's "`policy is we do not calculate statistical estimates for mixed samples.'" Id. at 294. The evidence also showed that the police investigated several suspects other than defendant, including the victim's roommate, Kristina McKee. Id. at 285, 308. We concluded that "absent some analytic or interpretive evidence concerning the likelihood or significance of a DNA profile match, Matthews' testimony concerning the potential match between defendant's DNA and the DNA contained in the mixed blood samples found on the knife blade and the doorknob was insufficient to assist the jury in determining whether defendant contributed DNA to the mixed sample." Id. at 301. We did not, however, declare or delineate the appropriate articulations for expressing the extent or meaning of a potential match, but merely held that "some qualitative or quantitative interpretation must accompany evidence of the potential match." Id. at 302.

On retrial, Megan Clement, a Technical Director at Lab Corp testified about the statistical significance of the potential match of the DNA profiles from the mixed blood samples and also presented evidence of additional DNA testing: McKee and two other suspects were excluded as possible contributors on all of the evidentiary items. The parties also stipulated that a fourth person was excluded as a possible DNA contributor to the mixed blood samples on the knife blade and bedroom doorknob.

Before the retrial, the trial court held an extensive evidentiary hearing to decide the admissibility of the statistical analysis offered by the prosecutor concerning the mixed DNA evidence. The trial court recognized Clement as an expert, and she testified concerning Lab Corp's use of proficiency testing to ensure reliable DNA results. It was determined that the sample had DNA from more than one contributor because multiple loci showed three characteristics.¹ Neither defendant nor the victim could be excluded as contributors because characteristics of the mixed sample on the knife blade were contributed by either defendant or the victim. Thus, no evidence existed that anyone other than defendant and the victim contributed to the mixed sample.

Testing of the mixed sample found on the victim's bedroom doorknob produced reportable results at five loci. The doorknob sample, like the knife blade sample, clearly contained a mixture of DNA from more than one person. The DNA profiles of the victim and defendant were compared against the profile from the doorknob. Neither the victim nor defendant could be excluded as contributors. Again, the characteristics in the doorknob sample were shared by the victim or defendant. Therefore, like the sample from the knife blade, the evidence pointed to only two contributors.

Before the end of 2000, Lab Corp did not calculate statistical ratios for mixed sample DNA. In July 2000, the DNA Advisory Board endorsed two methods for calculating statistical ratios for mixed samples: the likelihood ratio and the probability of exclusion or probability of inclusion calculation. The Federal Bureau of Investigations (FBI) had also developed a computer program using accepted statistical methods to replace handwritten probability calculations used with samples containing mixed DNA contributions.

Clement testified that Lab Corp followed the recommendation of the DNA Advisory Board and used the FBI computer program to calculate the probability of inclusion or exclusion as to the mixed sample found on the knife blade. Clement testified that the combined probability of selecting an unrelated individual who could be included as a contributor to the mixture was 1 in 1,210 for the African-American population, 1 in 952 for the Caucasian population, 1 in 1,115 for the Southeastern Hispanic population, and 1 in 916 for the Southwestern Hispanic population. Clement further testified that the combined probability of exclusion was that 99.17 percent of the African-American population would be excluded as contributors to the mixture of DNA found on the knife blade. Hand calculations confirmed the accuracy of the computer calculations.

A likelihood ratio was also calculated for the profile obtained from the knife blade. The evidentiary sample from the knife blade was 164,000 times more likely to be a mixture of the victim's and defendant's DNA than to be a mixture of the victim's and an unknown African-American person's DNA, 868,000 times more likely to be a mixture of the victim's and defendant's DNA than to be a mixture of the victim's and an unknown Caucasian person's DNA, 1,030,000 times more likely to be a mixture of the victim's and defendant's DNA than to be a mixture of the victim's and an unknown Southeastern Hispanic person's DNA, and 944,000 times more likely to be a mixture of the victim's and defendant's DNA than to be a mixture of the victim's and an unknown Southwestern Hispanic person's DNA.

Clement also testified concerning the probability of inclusion/exclusion of the doorknob sample. The probably of randomly selecting an unrelated individual who could be included as a contributor was 1 in 319 for the African-American population, 1 in 260 for the Caucasian population, 1 in 444 for the Southeastern Hispanic population, and 1 in 350 for the Southwestern Hispanic population. The combined probability of exclusion supported the conclusion that 99.68% of the African-American population be excluded as potential contributors to the mixed DNA.

The likelihood ratio method was also used for statistical calculations for the mixed sample on the doorknob. Two alternative hypotheses were used. One was that the mixture derived from the victim and defendant, and the other is that it came from the victim and an unknown contributor. The sample from the doorknob was 3,100 times more likely to be a mixture of the victim's and defendant's DNA than from the victim's and an unknown African-American's DNA. It was 1,870 times more likely to be a mixture of the victim's and defendant's DNA than a mixture of the victim's and an unknown Caucasian person's DNA, and was 11,700 times more likely to be from the victim and the defendant than from the victim's and an unknown Southeastern Hispanic person, and was 6,720 times more likely to be a mixture of the victim's and defendant's DNA than a mixture of the victim's and an unknown Southwestern Hispanic person.

Clement testified very specifically that two databases are employed to calculate statistics and explained each carefully. Clement also testified that the statistical calculations at issue are not novel or new and are used in many areas other than forensics. While Clement agreed that the use of these statistical estimates was new to forensic laboratories, she explained that statisticians have been using the statistical estimates for years to report statistics for mixture calculations. Because the forensic scientific community has endorsed them, laboratories now embrace statistical calculations for mixed contributor samples. The FBI's computer program has also been accepted by the scientific community and approved by statisticians and human geneticists. Clement testified that the two calculations used to calculate the statistics in this case are generally accepted in the scientific community in the field of statistics.

Frederick Bieber, a medical geneticist employed by the Harvard Medical School and Bringham and Women's Hospital also testified as an expert at the evidentiary hearing. Bieber explained likelihood ratios, their...