PEOPLE v. NELSON

Filed 8/31/06

CERTIFIED FOR PARTIAL PUBLICATION^*

IN THE COURT OF APPEAL OF THE STATE OF CALIFORNIA

THIRD APPELLATE DISTRICT

(Sacramento)

----

Continue from Part II ………

B

Here, evidence of the random match probability derived through the product rule was presented to the jury. As we have explained in part D of part I, ante, under the product rule, the population frequencies of all measured alleles are estimated through use of population databases and then are multiplied together to generate a probability statistic for the complete multi-locus profile. (People v. Soto, supra, 21 Cal.4th at p. 525.)

Defendant does not dispute that the product rule is generally accepted in the scientific community for use in cases where a suspect is identified through traditional investigative techniques and is then compared one to one with an evidentiary sample. But he argues that there is no general agreement in the case of a cold hit, i.e., where the suspect is identified through a DNA databank.

Before discussing this contention, we must describe the process by which defendant was identified.

C

As we have previously noted, forensic DNA comparisons are performed by measuring alleles at different loci. Individuals inherit one allele at each locus from each parent. In very rare instances, a mutation may cause a person to have three alleles at a locus. More commonly, a person may be a homozygote at one or more loci, meaning he or she inherited the same length of allele from each parent. When a person inherits different alleles from his or her parents at a locus, the person is a heterozygote at that locus. In PCR-STR testing, a heterozygote locus will reflect two lengths of allele. A homozygote locus will reflect one length of allele, but the amount of the alleles in the sample will often reveal the homozygote nature of the locus.

In a sexual assault case, an evidentiary sample will often contain contributions from both the male and female. The male contribution, sperm, is essentially half cells, i.e., it contains one-half of the male's DNA. Collectively, the sperm will include the male's total DNA. The female contribution consists of nucleated epithelial cells, each containing a full compliment of DNA. There is a method, called differential extraction, which can separate the male and female contributions. However, the method is not always completely successful, with the result that there may remain some female DNA in the sperm fraction and some male DNA in the non-sperm fraction.

Degradation of an evidentiary sample can affect DNA testing. Degradation will not change a DNA profile; but with degradation, alleles may become impossible to detect. Typically, the longer alleles are most affected by degradation. There is a characteristic pattern to degradation, which will indicate to an analyst that the sample has partially degraded.

In PCR-STR testing, laboratories use commercial testing kits. During the initial phase of DNA testing in this case, commercially available kits were the Profiler Plus kit, which tested nine loci and a gender marker, and the Cofiler kit, which tested six loci and a gender marker. Two of the loci tested with these kits overlapped. Thus, if both were used, the laboratory could test 13 loci and the gender marker. Eventually, a kit called the Identifiler kit became available and was validated. That kit combined the Profiler Plus and Cofiler kits and was more sensitive. With greater sensitivity, the Identifiler kit would be better at detecting alleles in small or partially degraded samples. However, it might also detect a foreign allele left by something such as sneezing or even talking over the sample.

At the time defendant's reference sample was analyzed and entered into the state convicted offender databank, California's Department of Justice laboratory was using the Profiler Plus kit to test nine loci. The computer search engine that was designed to compare evidentiary profiles to profiles in the databank was not intended to identify the source of the evidentiary sample; rather, it was a screening device. The program would record a moderate stringency match at a locus if a profile in the databank matched at least one of the alleles in the evidentiary sample. A high stringency match occurs where the evidentiary profile reflects two alleles and the databank profile matches both. A person would be identified as a candidate match if there were moderate or high stringency matches on at least seven loci.

To be declared a match for forensic purposes, a suspect's profile must match the evidentiary profile for every allele at every locus that is identified. A single mismatch excludes the suspect. The state's search engine was not programmed to require such a complete match. Evidentiary samples are often mixed with contributions from both the perpetrator and the victim. Thus, for example, if an evidentiary sample reflects two alleles at a locus, it may be that the perpetrator was a homozygote at that locus and contributed one allele, while the victim contributed the other. Further analysis can make the distinction, but for search purposes the state laboratory does not do so. The standard of a moderate stringency match at seven loci was set through experience because that standard best limits coincidental matches without risking exclusion of the actual perpetrator.

If a databank search identifies a candidate match, the person conducting the search examines the data to determine whether there is significance to the match such that it should be reported to the requesting agency. If the candidate match appears significant, the laboratory reanalyzes the original convicted offender sample with those stored on either side of it to ensure there was no laboratory mix up. The laboratory then reports the name of the candidate and the results of the search to the requesting agency.

When the biological evidence from the Ollie case was submitted to the county laboratory, the analyst, Jeff Herbert, did a direct digest on a cutting from one of the semen stains on the sweater. A direct digest means that he did not use a differential extraction to separate the male and female contributions. Herbert tested the sample with both the Profiler Plus and Cofiler systems. He detected complete genotypes, i.e., two alleles, at 11 loci and one allele at another locus. He was unable to detect alleles at one locus. The profile was submitted to the Department of Justice, and the nine-loci profile of the Profiler Plus kit was run through the state's computer search engine. Defendant was identified as a candidate and, in due course, the identification was reported to the county.

With a warrant, oral swabs were obtained from defendant. Herbert ran a confirmation test to ensure that defendant's profile matched the evidentiary profile. Herbert then developed a 13-loci profile for Ollie from hairs retained during the victim's autopsy. The sweater sample that was initially tested was a mixed sample with a major and a minor contributor. Defendant's profile matched that of the major contributor, and where alleles of a minor contributor were detected, they matched Ollie's profile.

Herbert next analyzed the vaginal swab and another stain from the sweater using the Profiler Plus system. On this occasion, he did a differential extraction to separate the male and female contributions, although without complete success. Herbert was able to detect alleles at eight loci. Defendant's profile matched that of the male contributor at each loci.

When the Identifiler system, with its greater sensitivity, was validated, Herbert tested the vaginal swab and sweater stains with that system. He was able to develop a complete profile at 13 loci from the vaginal swab. Defendant's profile matched the sperm fraction at each locus.

D

The trial court conducted a pretrial evidentiary hearing with respect to defendant's prong one Kelly objection to the statistical expression of the meaning of the DNA testing. The parties agreed the court would take judicial notice of, and consider as evidence in this case, the evidence presented in a prior Kelly hearing in the case of People v. Robinson, Sacramento County Superior Court No. 00F06871 (hereafter the Robinson case). The trial court in the Robinson case had conducted a lengthy hearing and ultimately determined that use of the unmodified product rule is generally accepted in cold hit cases.^[1]

A number of experts testified at the Kelly hearing in the Robinson case. They included Dr. Ranajit Chakraborty, a renowned expert in human population genetics (see People v. Soto, supra, 21 Cal.4th at p. 527, fn. 20); Dr. George Sensabaugh, Jr., a forensic biologist and biochemical geneticist who is an expert in the forensic use of DNA (see People v. Pizarro (2003) 110 Cal.App.4th 530, 589; People v. Axell, supra, 235 Cal.App.3d at p. 849); Gary Sims, who has a master of public health degree with a specialty in forensic science and is director of the case work section of the Department of Justice laboratory; Dr. Dan E. Krane, an associate professor of biological science at Wright State University; Dr. Norah Rudin, a forensic DNA consultant; and Dr. Laurence Mueller, an ecologist and population geneticist who has frequently appeared as a defense witness at Kelly hearings (see, e.g., People v. Soto, supra, 21 Cal.4th at p. 529; People v. Venegas, supra, 18 Cal.4th at p. 72; People v. Pizarro, supra, 110 Cal.App.4th at p. 595; People v. Smith (2003) 107 Cal.App.4th 646, 662).

From the evidence before the trial court, it appears virtually all DNA scientists believe that PCR-STR testing with numerous loci has a tremendous power of discrimination. And most, if not all, agree that the unmodified product rule is a valid and reliable means of demonstrating what it purports to demonstrate, that is, the rarity with which a particular multi-locus profile is expected to appear in the population and thus the probability of a random match.

Some scientists opine that the power of discrimination with multi-locus PCR-STR testing is so great that source attribution can be declared, i.e., it can be declared that the defendant is the source of the evidentiary sample. This view does not appear to have achieved general acceptance. Nevertheless, the minority view does not create a controversy as to use of the product rule. Those scientists simply believe that when the expected frequency of a profile, determined through the product rule, becomes extraordinarily miniscule, then a conclusion of source attribution can be drawn.

Evidence before the trial court indicated that, in addition to the random match probability determined through use of the product rule, there are three possible methods of explaining the statistical significance of a DNA match in a cold hit case. We will discuss each in turn.

One method was suggested by the National Research Council in 1992. (Nat. Research Council, DNA Technology in Forensic Science (1992) (hereafter NRC-I).) At that time, forensic use of DNA was in its infancy, and the idea of using a convicted offender databank to solve crimes was new. (See Annot., Validity, Construction, and Operation of State DNA Database Statutes (2000) 76 A.L.R.5th 239.) The NRC-I report suggested that in a databank search, one set of loci could be used to screen and identify a suspect and then a different set of loci could be used to confirm a match. Statistical analysis using the product rule would be done on the second set of loci.

The NRC-I suggestion was not based upon any scientific or statistical theory. It was concerned with matters of probable cause, which are judicial rather than scientific questions. No scientific or statistical principle requires that competent, relevant information be disregarded. A subsequent report of the National Research Council criticized the NRC-I suggestion for wasting data. (Nat. Research Council, The Evaluation of Forensic DNA Evidence (1996) (hereafter NRC-II).) The evidence in the trial court established that the NRC-I approach has been generally rejected and is not used in any laboratory or in any jurisdiction.^[2]

The existence of such an old suggestion of a method of practice that was never generally accepted, and has long since been generally rejected, does not create a current controversy in the relevant scientific community.

Another approach that has been suggested is the creation of likelihood ratios through the use of a Bayesian formula. â€