Please welcome our first guest blogger: Sue Carney. Sue is an FSS senior forensic scientist. Her areas of expertise include body fluid interpretation and DNA evidence. She has worked on a number of cold cases. You can follow Sue on Twitter.
DNA Profiling in Cold Cases – The UK Process by Sue Carney
DNA profiling often brings new life to cold cases. But exactly how does it help to identify potential suspects in an incident that might have happened 20 years ago?
To understand this process and the unique issues surrounding cold cases, we must first understand DNA profiling in current cases. In the UK a DNA profiling system known as SGM plus is used in the majority of criminal cases submitted for forensic examination. The system examines ten regions of DNA plus a sex test.
One can expect to obtain an SGM plus profile from a body fluid such as a blood, semen or saliva stain, or from other biological samples such as skin, tissue or hair roots. Contrary to popular belief, there is little DNA in sweat or urine although these may carry with them a few cells from the surface of the skin.
The sensitivity of DNA profiling tests has improved over the years since the Forensic Science Service (FSS) used a very early form of DNA profiling in the famous Pitchfork case in the eighties. To obtain a DNA profile using some of the early predecessors to SGM plus required a blood stain similar in diameter to a ten pence piece. Now, it’s possible to obtain a profile from a blood spot as small as 1 mm in diameter.
In current cases, DNA is obtained from biological material which is relatively fresh. DNA on an item in the right conditions can persist for a number of years, although sometimes it may not be intact, especially if the item has been exposed to the outdoors, to weather, or to putrefaction (clothes from a body undiscovered for some time, for example.) Generally speaking, in current cases with sufficient biological evidence, the chances of obtaining a DNA profile are good.
In a cold case, the original items (clothing, intimate swabs, other recovered items) may not have been retained. Certainly in UK cases. The concept of DNA profiling in a forensic context had not been widely anticipated prior to the eighties, and it’s understandable that issues such as lack of storage, might dictate a destruction of items policy that would appear shortsighted today.
In cold cases from this period, the only retained materials comprise items generated as part of the original examination. These include microscope slides, hair samples, fibre tapings and collected debris. Intact DNA might exist in some of these items, in spermatozoa (sperm cells) on microscope slides, in hair roots or in cells adhering to tape lifts. However, the amounts of DNA are extremely small. Considerably smaller than the quantity of DNA required for an SGM plus profile. To detect the DNA in these samples requires a much more sensitive method.
The generic term for this sensitive DNA profiling method is Low Template (LT) DNA profiling, but each of the UK forensic providers offering this technique has its own name. The FSS and Cellmark Forensic Services each refer to it as Low Copy Number (LCN) DNA profiling whilst Key Forensic Services Limited offers DNAXtra, and LGC Forensics offers DNA SenCE. For simplicity, I’ll refer to the technique as LCN from hereon.
LCN examines the same ten areas of DNA as SGM plus, but can detect much smaller quantities of DNA. To understand how, we must digress here for a moment, and consider the stages of DNA profiling:
Following an initial extraction stage during which the DNA is separated from the body fluid or cell sample, a quantification stage measures the concentration of DNA in the extract. This serves mainly to optimise the next, crucial stage, known as amplification. At this stage, the ten areas of DNA we’re interested in (plus the sex test regions) are targeted in a chemical reaction that makes exact copies of them. Only the target areas of DNA take part in the reaction. I’ve heard anecdotes in which sardonic defence barristers ask if the DNA evidence has been manufactured. But it seems obvious to me that if these target regions are not intact or not present in the sample to begin with, then DNA profiling cannot detect them.
The reaction itself, known as PCR, happens in a chemically controlled cycle. Each cycle doubles the amount of target DNA or template. SGM plus uses 28 cycles. Therefore the amount of template DNA has been doubled 28 times by the end of the reaction.
In LCN, the number of cycles is increased to 34. These six additional cycles mean that the amount of DNA present at the end of the reaction can still be detected in the final stage of the profiling process, even if the template DNA was only present at a very low level at the start.
To detect the small amounts of DNA that might be present in the type of sample available in a cold case, the increased sensitivity of LCN works. Also in current cases to detect DNA deposited on items by handling or touching (referred to as ‘touch DNA‘ by some scientists) or in other samples containing smaller amounts of DNA. However, the story doesn’t end there.
The increased sensitivity of LCN brings with it a selection of other issues that don’t impact on SGM plus to the same degree. Firstly, such a sensitive system is also sensitive to potential contamination. For this reason LCN labs are strictly environmentally monitored and scientists adhere to extensive anti-contamination procedures. However, even in the most stringently controlled environment, it is possible for minute fragments of DNA to drop into a sample. If such a fragment happens to contain one of the ten areas of DNA used in profiling, then it could show up in the LCN results. Such an occurrence is known as a drop in event, for obvious reasons. If a DNA fragment from elsewhere in the genome (a person’s entire DNA) drops into a sample, it is not targeted in the profiling process, no drop in event is detected, and the LCN results are unaffected. But how likely is a drop in event?
Consider the genome described as a large library. The ten regions of DNA examined in forensic DNA profiling could be described as ten books from the library. Therefore, the chance of a random ‘drop in‘ fragment, which could be from any part of the entire library, containing one of these ten books, is limited. Drop in events are relatively rare. To ensure that LCN results are not spurious drop ins, samples are DNA profiled in duplicate. Only if a result appears in both tests, can it be considered a genuine result since exactly the same drop in event would be extremely unlikely to occur in both tests.
The second issue requiring consideration is the principle of sampling error. That when sampling something present at a very low level, each sample contains different parts of the whole. This sampling error is called stochastic variation, or more simply can be described as as drop out. Consider having a number of identical jigsaw puzzles. Each jigsaw is a picture of a white wall with a portrait on it. The portrait is small, made up of a few hundred pieces, whilst the white wall is huge and comprises a few thousand pieces.
The jigsaw enthusiast grabs a handful of pieces from one of the boxes and assembles them to form part of the complete jigsaw. Some of the pieces might fit together. Some may even reveal parts of the portrait in the centre, but most will be of the wall. If this process is repeated for the subsequent puzzles, he or she may reveal different parts of the picture. LCN DNA profiling results are often only part of the picture. Because the template is present at such a low level, only a few pieces are sampled for the amplification stage of profiling. The missing results are often described as having dropped out, and those that have been selected by chance in the duplicate tests, may not be representative of the whole profile.
Thirdly, depending on the history of the item from which the template DNA has been recovered, the sensitivity of LCN often tends to detect mixtures of DNA from more than one person. It may not be possible to determine how many people’s DNA is present in a mixture, and this brings other interpretational issues.
Due to drop in, drop out and mixtures of DNA, interpretation of LCN results is a carefully considered process. It is very unlikely that the forensic scientist will be able to identify the specific cell type or body fluid that the LCN results are from. The amounts of DNA are so small that no initial chemical tests for body fluids are usually carried out.
It is not usually possible to determine how or when LCN DNA might have been deposited, or indeed to be sure that it relates to the incident being investigated. These questions are of vital significance in a criminal trial. The presence of matching DNA alone does not necessarily prove guilt or innocence without contextual information as to how and when the DNA might have been deposited. However, the LCN reporting forensic scientist may not be able to offer any opinion on how and when.
In 2007, in R v Hoey, the Omagh bombing case, the judge questioned the validation of the LCN profiling system. LCN profiling was suspended for a while, at the recommendation of the Association of Chief Police Officers (ACPO), whilst an expert review of the technique was carried out by Professor Brian Caddy of Strathclyde University. In April 2008, the Caddy Report concluded that LCN DNA profiling is scientifically robust and suitable for use in police investigations.
LCN is often of much more use as an intelligence tool. The detection of a best estimate, partial DNA profile, from a consensus of at least two tests, may not contain enough information to be added to The National DNA Database and held as a permanent crime stain record. It would be unwise to do so, given that these DNA results might not relate to the crime being investigated. However, they can be used to carry out a one off, speculative search of the database, which compares the results to every permanent record held on the database once. In comparison, permanently held records are compared to every other record, every 24 hours.
A speculative search of LCN results might identify a list of possible matches. Some are eliminated based on other, non searchable information in the LCN results. Searches might also be restricted based on geographical area or age. It’s no use having a list of 18 – 25 year old suspects, when the incident happened twenty years ago and the perpetrator was a man in his thirties, for example.
A list of potential suspects in an intelligence report may be the outcome of the successful LCN investigation of a cold case in which useful DNA information is detected. At that stage, there is still a great deal of police work ahead in the case before any individual might be brought to trial, if it gets that far. In a few cases, the DNA results identify a single suspect. These are the cases that hit the news, such as the conviction of Leslie Marshall for the 1989 rape of a woman in Burnley, Lancashire. Marshall was sentenced in August 2010. In the remaining cases, the LCN DNA profiling might play only a small part. But that’s how it’s done.