As the basis for DNA analysis, there are only a small fraction of sequences that differ between individuals in a given population and the value of the DNA tests lies in the discriminatory ability. Once DNA samples from evidence and a suspect are analyzed they can be compared to a population database and statistically narrowed to a low enough frequency to suggest that both samples came from the same person. This merger of genetic science and applied information substantiates the necessity of statistical application in forensic science.
Testifying in court that a sample match has less than a 1in 3 billion statistic of coming from anyone other than the suspect sampled is in reality an extrapolation from a test pool based on allele frequencies from a given population. The FBI maintains a database of population statistics separated by ethnicity that is available to laboratories performing DNA typing. This database does not contain individual specific information, rather it contains information regarding age and sex from anonymous blood donors (OTA, 1989).
Due to the uncertainty of multiple loci identification there are two basic approaches to utilizing statistics in forensic science, the “window of uncertainty” method and the calculations method. A “window of uncertainty” determines the matching process where two bands must fall close enough in size to be within the limits of their respective measurements. This method is a measurement confirmation step based on the size of the band-producing fragment.
Duplicate measurements are made of the sample and the majority of the measurements should fall within a percentage (< 5%) of the correct measurement value. If two bands, one from the victim and suspect, have a “window” overlap then they are considered a match. Once a match is determined the bands are assigned to a class by size (NRC, 1996). The calculated statistics methods used in forensic science can be determined using either the probability of a random match, referred to as match probability, or a likelihood ratio.
The match probability is calculated from the frequencies of DNA markers in a database, like the FBI’s population statistics database. If a match is made by a search through a large database a calculation must be applied to ensure the validity of the match. A calculation based on the loci not used in the search is a sound procedure if STR or another system with multiple loci is used. Another procedure is to apply a simple calculated correction by multiplying the match probability by the size of the database search.
The likelihood ratio is calculated as the probability of a match if the evidence DNA and the suspect DNA to be the same to the probability they came from a different person. Statistics can easily be measured using the Bayes Theorem approach. It is calculated as follows: Prior odds x Likelihood Ratio (LR) = Posterior odds. This can be read as considering a person’s theoretical odds of the risk of being caught (i. e. witnesses or security devices) multiplied by the theoretical odds of additional evidence (leaving prints, DNA, etc. ) and the outcome is the change in the person’s judgment.
The likelihood ratio (LR) is otherwise explained as the availability of evidence that could either prove a person guilty or innocence (Billings, 1992). Although DNA may seem to be enough conclusive evidence to convict or exonerate a suspect, there exist two fallacies that dismiss DNA in a trial in light of additional non-DNA evidence, also contributing to the likelihood ratio, LR. First, if the LR, or additional evidence, is in favor of innocent, without accounting for prior odds, then this situation is referred to as prosecutor’s fallacy.
This occurs when non-DNA evidence arises and sway’s the juries, regardless of a match with DNA evidence. Second, a defendant’s fallacy could occur if the conditional situation is reversed. This is when the DNA may only match to 1/100 probability, assuming that in a given population anyone with a similar profile or make-up as the evidence sample could be as likely to have left the sample as the suspect. However, the other 99 people were not selected as a suspect and do not have non-DNA evidence that has built a case against them (NRC, 1996).
Laboratory Accuracy When a lab implements DNA typing systems it must also maintain standards based on quality control (QC) and quality assurance (QA). Standardization refers to the regulation of using a specific method chosen and this is the first step in implementing the analysis into a laboratory. The method chosen for DNA typing must be subjected to rigorous trial tests and the analysts must be able to replicate results before applying the analysis to casework (Farley, 1991).
The FBI Standards for Forensic DNA Testing Laboratories (formerly known as the DNA Advisory Board) and the American Society of Crime Laboratory Directors-Laboratory Accreditation Board (ASCLD-LAB) established quality assurance standards followed by the FBI Labs and any lab that wishes to be accredited by these organizations. Part of the standard is to maintain positive and negative controls in processing certain amplification and sequencing procedures (Isenberg, 2002).
At a minimum an accreditation and validation program must have a review board comprised of a sample from biologists, forensic scientist and medical doctors to determine the status of new DNA-based tests. Second, the program must have a proficiency test for the analysts and a licensing program for both government and private laboratories that submit their data in court proceedings. This allows expertise outside of the crime studies field to regulate and monitor the forensic laboratories as a joint effort and safeguards from conflicts of interests (Billings, 1992).
The 1992 NRC Report outlined a proposed QC and QA Guidelines as part of a regulatory program. Some of the major points were that first analysts must have education, training, and a thorough understanding of the principles appropriate with the analysis performed and testimony provided. Second, analysts must pass annual proficiency tests before they are allowed to perform any new analysis test. The proficiency tests are usually a mock scenario where the analyst must determine the common source from a set of samples and their ability to interpret and replicate data is evaluated.
The test could be either open or declared, in which the analyst is informed they are performing a proficiency test, or the test could be full-blind, in which the analyst is not informed. Third, procedures must be supported by published data in the scientific community and the laboratory must have documented clear instructions on the procedures for safety, the handling of evidence and laboratory security as well as the analysis procedure. Finally, case records and data must be retained by the laboratory and available for audit on court orders.
The ASCLD-LAB requires extensive documentation of all laboratory operations as part of their accreditation, including training, calibration of equipment and validation of methods, to name a few (NRC, 1996). Safeguarding Against Error Regardless of accreditation and solid QC methods, determining the sample type and quantity must first be established for collection. Some common samples that can be used are semen, saliva, hair roots, teeth and white blood cells, because they all contain a cell nucleus.
The outer layer of skin does not contain nuclei but, as Locard’s Principle outlined, there may be a useful sample transferred from sweat or sebaceous secretion. The amount of sample needed for a conclusive result varies on a “case-by-case” basis in any typing system. Factors such as environment, bacteria, or mixed body fluid evidence will influence the quality and quantity required to establish DNA presence in the sample (Farley, 1991). There are generally three areas of error that occur with every forensic laboratory that result in analytical error.
The first and most common errors are due to sample mishandling and data transfer errors. These can occur at any stage from evidence collection to writing the final report. Personnel training and redundancy in checklists and secondary review are a few of policy implementations that would cut the errors rates by a drastic amount (NRC, 1996). Field kits have been a major advancement that has reduced sample contamination. Crime scene personnel are now equipped with kits that contain a blood sample card that only need a drop and automatic finger pricks to break a small portion of skin for blood.
A drop is placed on the card, dried and stored in a plastic container. The alternative is to wait for a trained person to draw the blood with a syringe and transport the sample to have it analyzed. Another field sampler is the buccal scrape that uses a paper “toothbrush” or simply a cotton swab to scrape the inside of a suspect’s cheek to collect the DNA sample and is dropped in its own sterile and compact container (Wilson, 1999). One of the benefits of the discriminating power of DNA analysis is it can detect sample mishandling unlike classic blood group testing in the past.
If a sample is mixed then blood grouping could not distinguish that the suspect is type A and the victim is type O, it only detects type AO, two compatible types. DNA can distinguish because unless the two are blood related there will be two sets of markers when the sample is profiled. A second contributor to errors in the laboratory comes from faulty reagents, equipment or techniques. These errors could easily be avoided with appropriate controls in places and adhering to a QC program.
Running blanks, checking standards on the equipment routinely and monitoring expiration dates of reagents are a few guidelines that should be in every laboratory protocol. The final contributor to error in the laboratory is evidence contamination that results in test failures or unusual spikes. Three kinds of evidence contamination are outlined in the 1992 NRC: inadvertent contamination, mixed samples, and carryover contamination. Inadvertent contamination results from the environment the sample were collected.
Microorganisms, gasoline, chemicals in the area or plant materials are a few environmental occurrences that must be documented when collecting samples. Mixed samples by their nature are contaminated, such as blood from two persons may run together or a vaginal swab that contains semen. Testing separate areas from the sample may better assist in distinguishing the contributors. Carryover contamination occurs when a PCR product of one sample is mixed with reaction primers prepared for a different DNA target template for amplification.
This mix up assigns the wrong genetic type to the evidence being analyzed and results in a false match. The worst-case scenario is the wrong match came from a separate party in the case or a different case altogether (NRC, 1996). A good standard practiced in many laboratories that analyze DNA is to keep a sample of all analysts’ DNA typing on file, just in case. Laboratory Resources After certification and licensing a crime laboratory that has implemented DNA analysis must be prepared to undertake the demands of staying in business alongside technological advances.
Three main issues that affect a forensic laboratory that analyzes DNA are personnel, equipment costs and communication with the local law enforcement agencies. The first issue is with personnel and maintaining a trained staff to meet the caseload. DNA testing is generally used to compliment traditional testing such as detecting and identifying body fluid stains, so there may be a need for coverage that exceeds the 8-5 traditional work schedule. Biological samples degrade and suspects flee so it would not be unforeseeable that a laboratory was staffed to meet these demands.
Also, as part of the accreditation standards the ASCLD-LAB requires that the laboratory staff being properly trained in the procedures in their workplace as well as annual testing of their proficiencies. This training and loss of man hours for testing will periodically put a strain on the staff requiring that an adequate number of employees be hired to cover for this on any shift, determined by the facilities hours of operation. The second issue that is always a determining factor in government and private sectors businesses is the cost of doing business.
To illustrate how it would be a factor in choosing to have an in-house laboratory versus using a “lab for hire” an estimate cost ten years ago for RFLP analysis equipment ranged up to $100,000. Probes and reagents ranged at approximately $50 per sample. Equipment for PCR analysis was an additional $10,000 (Farley, 1991). It is common practice for many jurisdictions to send evidence and even bodies to their county medical examiner’s office or the FBI laboratory depending on the crime.
Some offices will use other county examiners for their expertise in a particular field or simply because they have the equipment for an analysis (Maricopa, 1993). The last main issue on resources need by all laboratories to provide DNA analysis is that the quality of the evidence determines the quality of the analysis and its interpreted results. A liaison between the crime labs and attorneys presenting the evidence is a competent presentation of the DNA results.
Also, a strong ability to effectively communicate is needed by all laboratory staff to be available to explain or discuss the results with any persons in the criminal proceedings, whether it be a judge, jury or attorneys. Training health practitioners and victim advocates are a few of the ways to ensure that the people that will come into contact with evidence will not adversely impact the data, such as in a rape case. Once DNA is entered as evidence in a trail is it of dire importance that the data is presented clearly so as not to leave the jury any doubt in the results or technology (Farley, 1991).