The Use of Biochemistry in Forensic Science

Biochemistry is of great utility for Forensic Science investigations, with the biochemical technique of DNA fingerprinting being of particular importance. The development of the biochemical techniques for DNA sequencing allowed the genomes of organisms to be sequenced (Berg et al, 2002: Prelude). As a result, genetic markers can now be used to identify individual members of a population (James and Nordby, 2005: 283). This capacity is clearly beneficial in forensic investigations. However, despite their utility, biochemical techniques must be applied with caution in forensic science. The results of biochemical techniques used in forensic science can have serious implications for the lives of individuals. I will demonstrate both the value and limitations of using biochemistry in forensic science through focussing upon the use of Low Copy Number (LCN) DNA typing in forensic science.

LCN DNA profiling

Introduction

The development of LCN DNA profiling provided forensic scientists with the capacity to analyse minute quantities of DNA. The technique is sensitive enough to analyse “just a few cells” (Gill, 2001: 229). This technique is therefore of particular benefit when investigating serious crimes for which there is limited evidence available (FSS, 2005a: no pagination). An example is provided by the forensic investigation which followed the 2001 murder of Peter Falconio in Australia. The evidence base was severely limited as no body was found. However, very small quantities of DNA were discovered inside the hand ties which had been used during the attack and on the gear stick of the victim’s van (FSS, 2005: no pagination). The use of LCN DNA profiling enabled this evidence to be linked to Murdoch, who was already suspected of the murder (FSS, 2005: no pagination).

LCN DNA analysis was also crucial in solving a documented murder in Northern Italy. Although no trace evidence was discovered on the victim’s body or at the scene of the crime, a search of the victim’s car provided blood stains, sweat and skin samples. The small amounts of DNA yielded by these samples provided profiles which were “identical to that of the saliva obtained from the suspect” (Pizzamiglio et al, 2004: 437). When confronted with this evidence, the suspect confessed the crime (Ibid.).

The biochemical technique

The increased sensitivity of the LCN technique is achieved by increasing the number of polymerase chain reaction (PCR) amplification cycles used (Gill, 2001: 229). Although optimum efficiency is attained by using no more than 28-30 PCR amplification cycles (Ibid.), a variety of studies have yielded useful results using more cycles. Findlay et al (1997) obtained profiles from single cells by using 34 cycles; Wiegand et al (2000) analysed epithelial cells which had been transferred from the assailant during strangulation using 31 cycles and Van Hoofstat et al (1998) analysed fingerprints from tool grips by using as many as 40 cycles.

The increased sensitivity offered by this technique is incredibly beneficial for forensic science investigations. The key tenet of forensics is: “every contact leaves a trace” (Locard, 1910). By enabling the analysis of barely visible samples, LCN DNA profiling increases the investigative power of forensic science (Hoffman Wulff, 2006: 2). However, with this increased sensitivity comes increased risk of misinterpretation. For example, the highly sensitive technique may reveal DNA from sources other than the sample analysed and the results must be interpreted with extreme caution (Gill, 2001: 229). The limitations of the technique will now be explored in detail.

Limitations of the technique

1. Experimental errors

Due to the increased number of PCR cycles used for LCN DNA profiling, there is an increased likelihood of experimental errors, which may significantly affect the DNA profiles obtained (Budowle, 2001). These experimental errors include: preferential amplification of alleles (causing allele drop out), the appearance of false alleles when stutters are preferentially amplified and the preferential amplification of alleles which are present because of contamination (Gill, 2001).

As a result of these experimental errors, it is difficult to validate the results of LCN DNA typing (Budowle et al, 2001: 2). Because experimental errors occur randomly, the results of LCN DNA profiling are not reproducible and replicate analyses can produce different DNA typing results (Gill, 2001). In addition, because the established interpretation thresholds for DNA analysis are too large to apply to the LCN technique, there is no stochastic threshold for use when evaluating the results of LCN processing (Hoffman Wulff, 2006: 2). Thus, the number of alleles required in order to establish likeness is open for debate (Budowle et al, 2001).

2. Contamination

Alongside awareness of the possibility for experimental errors to reduce the accuracy of LCN DNA profiling, it is important to consider the impact of evidentiary contamination. There is a high risk of DNA contamination before, during and after the forensic event under consideration, which reduces the accuracy of the technique. Although there is also a risk of contamination when undertaking standard DNA analysis, it has less impact upon the results of the profiling. As adventitious transfer and contamination usually involve only low levels of DNA, their effect upon the profile obtained by standard DNA analysis is minimal (Gill, 2001: 231). However, in LCN DNA analysis, the low levels of DNA from contamination pose a far more significant problem. As the essence of the technique is the detection of minute levels of DNA, there is a far greater likelihood of contamination DNA having a substantial effect upon the profiles obtained. Due to the sensitivity of the technique, both background level DNA and DNA from casual contact will be detected (Budowle, 2001: 2). This is most problematic, as these contaminants cannot be removed physically or statistically. Because there is no way that the movements and contacts of the victim before, during and after the crime event can be assessed and accounted for, the possibilities of adventitious transfer cannot be directly ascertained (Gill, 2001: 230).

The possibility of secondary transfer ought to also be acknowledged. Theoretically, secondary transfer means that extraneous DNA could be carried by the perpetrator and deposited at the crime scene. Van Oorschot and Jones demonstrated that DNA can be transferred from objects to hands (1997). Although the likelihood of such transfers is contested, such secondary transfers could result in the deposition of a multi-source sample at a crime scene (Phipps and Petricevic, 2007; Ladd et al, 1999). It may be very difficult to establish whether a true mixture of DNA profiles exists when using the LCN technique (Hoffman Wulff, 2006: 2). Therefore, it is important to acknowledge that evidence may include a mixture of DNA profiles, which may include disinvolved individuals, the perpetrator and crime scene investigators (Gill, 2001: 230). Such a possibility greatly complicates interpretation and means that the results obtained could well be flawed. When the results provided by the technique may affect the liberty of an individual, it is particularly important that limitations and possibilities for inaccuracy are acknowledged.

Difficulties related to contamination are made even more significant as a result of the “considerable lack” of understanding about the issues of the transfer and persistence of DNA, which constrains scientists’ ability to statistically account for DNA contamination (Gill, 2001: 230). There are significant differences in DNA deposition between individuals and as some are “better than others at shedding DNA”, decay rates are unpredictable (Phipps and Petricevic, 2007: 167; Lowe et al, 2002). For example, Murray et al (2003: 780) found that ‘good’ DNA shedders would come to form the major component of the DNA mixture found on a second hand watch strap after only several days. By contrast, ‘poor’ shedders took as long as two weeks to comprise the majority of the DNA in the mixture (Ibid.). Similarly, van Oorschot and Jones (1997: 767) demonstrated that; when a number of individuals handled objects, the dominant DNA profile was not always that of the individual who last held the object. Rather, the dominant DNA profile was dependent on the shedding ability of individuals (van Oorschot and Jones, 1997: 767).

However, identifying individuals as being either ‘good’ or ‘poor’ shedders is not possible, because the shedding ability of a given individual does not remain consistent. Indeed, variable factors have been demonstrated to affect the amount of DNA deposition. Phipps and Petricevic (2007) established that DNA deposition is affected by factors such as whether contact is made by the dominant or non-dominant hand and the time since the hand was last washed. Therefore, as the transfer and persistence rates of DNA are impossible to establish, LCN DNA profiling cannot provide an indication of when DNA deposition occurred. As such, both awareness of and further research into the multiple factors which influence DNA shedding is required (Phipps and Petricevic, 2007; Hoffman Wulff, 2006).

Further contamination can occur during the collection of evidence. Forensic evidence is generally collected in uncontrolled environments, by police officers whose training in preserving the integrity of biological samples is, at best, limited (Lynch, 2003: 96). This factor becomes especially problematic when using LCN DNA analysis, as the small sample size greatly increases the risk of contamination. Given that LCN DNA analysis ought only to be undertaken in sterile environments, where equipment and furniture must be frequently bleached, the quality of the collection of the sample is very important (Gill, 2001: 229). Although laboratory standards cannot be expected, the evidence must be reviewed with an acknowledgement of this limitation.

In the UK, Regina v. Hoey in 2007 demonstrated the potential impact of these limitations. The basis for Hoey’s 2003 conviction for 29 murders during the Omagh bomb attacks was the evidence obtained from LCN DNA profiling. However, the conviction was then overturned on appeal in 2007, as the appeal established that the DNA evidence had been handled in a “thoughtless and slapdash” fashion (Weir, 2007: 23). Although this example displays that evidence which has been treated incorrectly may be dismissed at appeal, it is crucial to note that evidence ought to be presented alongside information about the potential limitations of its accuracy, as grave miscarriages of justice may otherwise result.

Due to the limitations of the technique, analysis of the results of LCN DNA analysis must only be done with an awareness of the “special considerations” about the potential inaccuracy of the technique (Gill, 2001: 229). It is crucial that both forensic scientists and courtroom staff are aware that LCN DNA evidence is inextricably linked to a higher likelihood of achieving adventitious likenesses or exclusions than standard DNA profiling (Gill, 2001: 230). This is clearly problematic when the results of the technique are being used legally, as incorrect outcomes “have devastating and untenable consequences” (Morgan and Bull, 2007a: 43). Although it may be possible to statistically account for experimental errors in the future, it will remain crucial to acknowledge the potential for personnel to make mistakes. In U.S. V. Llera-Plaza in 2002, an FBI scientist stated: “error rate is a difficult thing to calculate… to say there’s an error rate that’s definable would be a misrepresentation… the method is one thing, people making mistakes is another issue” (Saks and Koehler, 2005: 894).

3. Problems with interpretation

DNA profiles (both normal and LCN) are often not interpreted correctly. Although DNA evidence ought to only be used to exclude, ‘matches’ are commonly referred to. For example, following the 1981 murder of Marion Crofts, the UK Forensic Science Service contended that a LCN DNA profile found on the victim’s clothing matched that of the suspect Jasinskyj (FSS, 2005). Similarly, in U.S. v. Byrd, a forensic scientist for Pennsylvania State Police testified that it was 99% likely that the DNA obtained from the murder implements matched the DNA of Byrd and his victim (Hoffman Wulff, 2006)

Despite the fact that DNA analysis superseded techniques such as handwriting analysis and lie-detector tests which were less ‘scientific’, it remains important to acknowledge the errors of interpretation which may still occur with the newer techniques (Lynch, 2003). Although the methodology underpinning DNA analysis is scientifically sound and has a “firm theoretical basis” (Broeders, 2006: 152), using procedures which are “commonplace in biomedical research” (Lynch, 2003: 95); forensic science remains an applied science. As such, although the results of the biological process may be sound, the inferences made from these results could still be incorrect. This consideration is especially relevant for LCN DNA analysis, where an apparently ‘matching’ profile can be obtained through contamination of the evidence.

Matches and categorical identifications are impossible throughout the realm of forensic investigation, “unless the number of potential sources is limited and known”, (Broeders, 2006: 153). Although the probability of individuals exhibiting high levels of DNA similarities is considered to be “vanishingly small” (Broeders, 2006: 155), DNA characteristics are nevertheless class characteristics and thus cannot individualise (Thornton and Peterson, 2002). Only where reference to “an indefinitely large set of alternative potential sources” has been made, can the Huberian principle of individualisation be exercised (Broeders, 2006: 153). This invokes the classical induction problem, that individualisation from DNA analysis would require the analysis of everyone who has ever lived, is living and will ever live. As such, DNA can only provide a probabilistic conclusion that the profile matches that of the suspect (Broeders, 2006). However, correct forensic procedure would only ever assess the similarity of DNA profiles after failing to exclude them (Budowle et al, 2001). As Stoney so eloquently highlighted, “what made us ever think we could individualise using statistics?” (1991: 197).

Thus, using DNA profiling for identification rather than exclusion overlooks the very nature of DNA profiling as a classification process and also contradicts one of the key tenets of forensic science: “when undertaking comparison of samples, exclusion should be sought rather than a match” (Morgan and Bull, 2007: 86). As a result of the increased sensitivity of LCN DNA analysis and the risks detailed above, strictly adhering to the principle of exclusion is especially important. However, it is evident that current use of LCN DNA typing does not always fulfil this key philosophy of forensics.

The expectation of obtaining DNA ‘matches’ has been further complicated by the CSI effect, which has led to juries placing increased trust in the ‘expert’ witness and contributed to the incorrect idea that forensic science is infallible (Morgan and Bull, 2007a). Although LCN DNA typing uses scientific techniques, in the courtroom credibility is “fashioned and undermined in testimony” (Lynch, 1998: 829). As the judge and jury are unlikely to be familiar with scientific theory and practice (Morgan and Bull, 2007a), the jury’s status as a “susceptible body of individuals”, whose judgment may have been affected by media portrayals of powerful and successful forensic techniques becomes most significant (Morgan and Bull, 2007a: 44). Although courts tend to place their trust in the ‘expert’ witness (Lynch, 2003), incorrect expert testimony has been cited as a contributor in 63% of wrongful convictions (Saks and Koehler, 2005: 893). Adherence to the exclusionary principle is therefore particularly important, especially due to the sensitivity of LCN DNA profiling.

The CSI effect has also increased juridical expectation for evidence to be presented. Juries now often “demand unreasonable levels of physical evidence” in order to reach a verdict” (Morgan and Bull, 2007: 84). ‘Negative evidence’ expert witnesses may even be called upon to explain an absence of evidence in a trial (Hoffman Wulff, 2006). Indeed, the increased sensitivity of detection provided by LCN DNA analysis may act to further such expectations. However, it is important for forensic scientists and courtroom staff to remain mindful that contacts that are unrelated to the forensic event may have transferred enough DNA to be detected by LCN analysis.

4. The Courtroom

Although DNA profiling utilises scientific techniques and may thus appear to be an objective procedure, the evidence itself remains silent and must be given a voice in the courtroom (Jasanoff, 2006: 330). As such, the objective science has to be represented. This need for representation renders the courtroom a “sociology of knowledge machine”, within which uncertainty can be produced (Lynch, 1998: 829). Indeed in 1995, U.S. v. Simpson, saw the defendant being exonerated after his “star-studded” legal team exploited every weakness in the process of evidence translation from crime scene to courtroom. (Jasanoff, 1998: 715). As there are so many limitations to consider where LCN DNA profiling is used, it is possible for lawyers to use strategically deployed language and powerful visualisations of evidence to dramatically influence legal proceedings (Jasanoff, 1998).

There is therefore a strong argument for controls on evidence integrity and expert quality to be implemented, as seen in the U.S. legal system. Frye v. United States, 1923, constitutes the principal control on evidence in the American courts, defining expertise as: that which has “gained general acceptance in the particular field in which it belongs” (Saks and Koehler, 2005: 894). Regulations such as these are urgently required in the UK, where “novel scientific techniques” are currently accepted, without special scrutiny (Ormerod, 2002: 774). It is perhaps telling that LCN DNA evidence is considered admissible in UK trials, but used only as a last resort in a US criminal case (Hoffman Wulff, 2006: 4).

Conclusion

This essay has argued that, although biochemistry is undeniably of great utility for forensic science, the bridge between a laboratory science and an applied science must be carefully negotiated. This argument has been demonstrated through a focus upon the limitations of the use of LCN DNA profiling. However, although convicting a suspect solely on the basis of LCN DNA evidence would not be wise, doing so would also contradict a key tenet of forensic analysis: the need “to employ a number of independent techniques” (Morgan and Bull, 2007: 86). The limitations of LCN DNA analysis would be greatly reduced in significance if the findings are supported or contradicted by evidence from other techniques, as dictated by the philosophy of forensic science. This paper has demonstrated that the limitations of LCN DNA typing are considerable, however adhering to the tenets of forensic investigation will mean that these limitations are highly likely to become exposed or negated.