Guide to DNA · 2018-06-08 · DNA profile matching that of another just by chance. It is thus not...

Guide toDNA

for Lawyers andInvestigatingOfficers

This booklet is designed to give lawyers andinvestigating officers a basic understanding of DNAanalysis and interpretation. It aims to assist them inthe investigation of criminal offences involving DNA

evidence and when dealing with expert witnessevidence in court.

Contents:page

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

What is DNA?. . . . . . . . . . . . . . . . . . . . . . . 6Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6Nuclear DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6Mitochondrial DNA . . . . . . . . . . . . . . . . . . . . . . . . . 7

How is DNA profiling done? . . . . . . 8

Types of DNA profile. . . . . . . . . . . . . . 10Nuclear DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Y-STR DNA profiles . . . . . . . . . . . . . . . . . . . . . . . 12Mitochondrial DNA . . . . . . . . . . . . . . . . . . . . . . . . 13

Dealing with issues ofcontamination . . . . . . . . . . . . . . . . . . . . . . 15

The National DNA Database® andthe role of the custodian . . . . . . . . . 17Familial searching . . . . . . . . . . . . . . . . . . . . . . . . . 20

Principles for the evaluation ofDNA evidence. . . . . . . . . . . . . . . . . . . . . . 21Assessing the value of the DNA evidence . . . . . . 21Match probability of one in a billion. . . . . . . . . . . . 23Reporting a DNA match to court. . . . . . . . . . . . . . 24Uniqueness probability . . . . . . . . . . . . . . . . . . . . . 25Scientist’s statement . . . . . . . . . . . . . . . . . . . . . . . 26The prosecutor’s fallacy . . . . . . . . . . . . . . . . . . . . 27Hierarchy of propositions. . . . . . . . . . . . . . . . . . . . 28The evaluation of DNA evidence in . . . . . . . . . . . . . database matches. . . . . . . . . . . . . . . . . . . . . . . . . 30

Presentation of evidence andstatements. . . . . . . . . . . . . . . . . . . . . . . . . . 32

Further advice and suggestedreading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

AppendixGlossary of terms and abbreviations . . . . . . . . . . 34

3© Crown Copyright 2004. Guide to DNA for Lawyers and Investigating Officers.

IntroductionThis booklet is designed to give investigating officersand lawyers a basic understanding of DNA analysisand interpretation to assist them in the investigation ofcriminal offences involving DNA evidence and whendealing with expert witness evidence in court.

This is an updated version of the ‘Lawyers’ Guide toDNA’, first published in March 1998, which has beenrewritten to include new topics brought about byadvances in DNA profiling.

The original version is still available from The ForensicScience Service® (FSS®) and as many of theprinciples are still relevant should be referred to if thereader has little or no previous knowledge of DNA. Fora glossary of terms used in this booklet see theappendix on page 33.


What is DNA?DNA stands for DeoxyribonucleicAcid.This is the name of the complex chemical found invirtually every cell in the body. It is like a very long,miniature piece of string containing the codedsequence that determines our physical characteristicsand directs all the chemical processes in the body.

DNA is found in two different parts of each cell - thenucleus and the mitochondria.

Nuclear DNAAlmost all cells in the body contain a nucleus. Thegenetic information coded for by nuclear DNA iscarried in the chromosomes from one generation tothe next, half of a person’s nuclear DNA beinginherited from the mother and half from the father.Thenuclear DNA of identical twins is expected to be thesame. Other siblings inherit different combinations ofnuclear DNA from the same parents and their DNA istherefore somewhat different from one another. TheDNA from unrelated individuals is likely to be evenmore different. Each generation of people is a newand different combination of genetic material from theprevious generation.


If it were possible to examine the entire DNA moleculefrom every person in the world it would be possible toascribe, using DNA technology alone, a sample ofDNA to one specific person (except identical twins).However, it is doubtful that this would ever bepracticable.

Instead, the techniques used by the FSS look only ata set of components of the DNA molecule. Each oneof these components is known as a locus (plural loci).The loci have been selected because they are regionswhere there is known to be considerable variationbetween people caused by short pieces of DNA codebeing repeated over and over again, end to end.Theyare called short tandem repeat loci (STRs). Althoughany given set of components (the DNA profile) will notbe unique to an individual, methods have beendeveloped to calculate the probability of one person’sDNA profile matching that of another just by chance.

It is thus not possible, in general, to express the kindof categorical opinions of identity of source that arepresented from fingerprint evidence. Nevertheless,DNA profiling is highly discriminating and can provideextremely powerful evidence for presentation at court.

Mitochondrial DNAMitochondrial DNA (mtDNA) is passed from mother tochild with no contribution from the father. Each siblingwill have the same mtDNA type and this will bepassed on through the maternal line for manygenerations without significant change. The analysisand interpretation of mtDNA is different from that fornuclear DNA.

This booklet deals primarily with the analysis andinterpretation of profiles obtained from nuclear DNA,(see page 13) has a brief outline of mtDNA analysis.


How is DNA profiling done?DNA profiling has developed rapidly in forensicscience since the late 1980s. Great advances havebeen made in the automation and computerisation ofthe technique, and there have also beenimprovements in the sensitivity and application of themethod. Today, the FSS is recognised as a worldleader in DNA profiling techniques, using advancedequipment and methods.

The main steps in the technical process are:

u Extraction The DNA is isolated and separated from other cellular material.

u QuantificationThe amount of DNA extracted from the sample is measured to determine the optimal amount of DNA required for the next step of the process.

u Amplification (Polymerase Chain Reaction - PCR)Specific areas of interest (loci) within the DNA molecule are targeted and multiple copies produced.

u ElectrophoresisThe amplified DNA is separated according to size.

u AnalysisThe DNA is measured in such a way as to allow convenient comparison with other DNA profiles and storage of the profile on The National DNA Database®.


STR analysis is the standard method used for bothintelligence and evidential purposes. The STRs haveno function in passing on physical characteristics -indeed, the precise function of the STR loci is still asubject of debate among scientists.

From a forensic point of view, the interesting feature ofan STR locus is that it consists of two alleles, oneinherited from the mother and one from the father, andthe number of repeats for each allele can varyindependently of each other. Each allele has a namethat reflects the number of repeats. “Allele 12” wouldindicate the presence of 12 repeats, “allele 14” wouldconsist of 14 repeats, and so on.

So, at one locus, a person may be designated (12,14)for example - indicating that he/she has one allele of12 repeats at that locus and another of 14 repeats. Itis, of course, possible for a person to inherit an allelewith the same number of repeats for a given locusfrom both parents, so a person who has inherited, forexample, the allele with 11 repeats from both parentswould have a locus designated (11,11).


Types of DNA Profile

Nuclear DNASGM / FSS SGM Plus profilesThe second generation multiplex (SGM) system wasintroduced by the FSS in 1995 for DNA intelligencedatabase purposes and in 1996 for the analysis ofbiological materials from a crime scene. The systemlooked at six STR loci and also tested for the genderof an individual. The probability that two unrelatedpeople would have the same SGM profile is of theorder 1 in 50 million.

The SGM Plus system was introduced in 1999. Thistests for the same six STR loci as SGM, along withthe gender marker and four new STRs. This makes itfar more discriminating than SGM. It is also is moresensitive due to improvements made in the process.The FSS can now analyse samples that would notpreviously have been successfully analysed withSGM because of the increased sensitivity.

Due to the overlap of six of the STR loci the twosystems are compatible. Therefore, samplesprocessed using SGM Plus can still be compared withsamples previously loaded to The National DNADatabase® that had been analysed using the SGMsystem. The SGM profiles can also be upgraded toSGM Plus to improve the level of discrimination andreduce the risk of “adventitious matches”. Anadventitious match is obtained when the DNA profilesfrom two individuals match just by chance.

To illustrate the process, consider this hypotheticalcase. A criminal justice (CJ) sample, taken from acheek scrape or hair root, was analysed with SGM in1997 and its profile was subsequently loaded ontoThe National DNA Database®. In 1999, a crime stainwas processed with SGM Plus and, when its profilewas searched against The National DNA Database®, itmatched that of the the original CJ sample. Theindividual concerned, when questioned about thelater offence, professed his innocence, stating that hehad never even been to the location of the crime. The


FSS can “upgrade” the original profile using SGMPlus. If the profiles still match, the strength of theevidence against him is increased. If they no longermatch then he can be eliminated from being thesource of the DNA.

Full / Partial DNA profilesA full SGM Plus DNA profile is obtained when theanalysis of all ten10 STR loci and the gender markerhas given successful results. A full profile would beexpected from items that are not degraded, such asreference blood samples, cheek scrapes and freshsamples from the scene of the crime. A partial DNAprofile occurs when it has not been possible to obtainresults for one or more of the STR loci or the gendermarker. This can happen when the sample beinganalysed has degraded after being exposed to theenvironment for a period of time, and/or when there isvery little material available.

Partial DNA profiles obtained from a crime samplecan be held on and searched against The NationalDNA Database® as long as they meet certainminimum criteria. The more partial the profile,however, the greater the number of matches that willresult and the lower their intelligence value. Partialprofiles that have only a few components of the DNApresent and do not meet the minimum loading criteriacan still be searched against The National DNADatabase® by carrying out ‘one-off’ speculativesearches. The intelligence value in such searchesdepends on what other information might be availableto help restrict the large number of matches likely tobe obtained.

DNA mixturesA mixed profile will result when DNA from two or moreindividuals is present in a sample. This can oftenhappen, for example, in sexual offences and in violentcrime where the attacker and the victim have bothcontributed.

Where semen has been recovered from rape victimsit is sometimes possible to physically separate themale and female components of the mixture and thusobtain the individual DNA profiles.


Where physical separation is not possible,identification of the possible profiles of eachcontributor may still be possible, depending on therelative proportions of DNA from each that is present.This is simplified where reference samples areavailable from at least one of the contributors to themixture.

For two-component mixtures, where the relativeproportions are somewhat similar, the FSS hasdeveloped a technique - pendulum list search (PLS) -for identifying and prioritising the possiblecombinations of individual profiles that could havemade up the mixed profile.

LCN (low copy number) DNA profilesLow copy number (LCN) DNA profiling has beendeveloped through extensive research by the FSS. Itexploits the increased sensitivity of the SGM Plussystem and enables the successful analysis ofsamples that would previously not have beenexpected to yield results (e.g. the cellular residue in afingerprint or on a swab from the inside of a watchstrap). LCN DNA profiling has enabled the FSS toanalyse samples from ‘cold’ cases from which it hadnot previously been possible to obtain a DNA profile.These samples have often been stored for manyyears in the hope that improvements in forensictechniques will allow these cases to be revisited at alater date.

LCN DNA profiles are more likely to be partial and/ormixtures. It may also not be possible to indicate thetype of fluid that the profile came from. It is, therefore,important with LCN DNA profiles that discussions areheld between the investigators, lawyers and thescientist to assist with the evaluation of the findings.

Y-STR DNA profilesThe male Y chromosome in nuclear DNA alsocontains STRs. These Y-STRs are passed down themale line, from fathers to sons without change overmany generations. They are different from those usedfor SGM/SGM Plus profiling but they can be used toproduce Y-STR profiles.


Y-STR profiling is helpful in comparing the malecomponents of mixtures containing male and femaleDNA and for the identification of missing relatives inthe male line. Because it runs unchanged throughgenerations it can also provide indications ofethnic/geographic origins.

Y-STR profiles can be useful in the investigation offamilies identified by the technique of familialsearching (see page 20). This will enable police tomaximise the use of collected DNA without causingfurther public intrusion.

Mitochondrial DNAMitochondria are found in every cell of the body andprovide the energy that is required for cellularprocesses. They have their own DNA, which isdifferent from the nuclear DNA and contains muchless information. However it is far more abundantwithin each cell. Partly due to this abundance,mitochondrial DNA (mtDNA) analysis is very sensitiveand can be used when the conventional STRtechniques (using nuclear DNA) would not normallyresult in a full profile (e.g. when analysing shed hairs,aged bones and teeth).

The technique used for mtDNA analysis detectsdifferences in the individual basic building blocks ofthe mtDNA molecule and is known as “mtDNAsequencing”. MtDNA sequencing is far lessdiscriminating than nuclear STR DNA analysis butnevertheless is very useful when taken in the contextof other evidence.

As mtDNA is maternally inherited, and all individualswho are related by a maternal link will have the samemtDNA profile, mtDNA profiling is also ideal for theidentification of bodies (e.g. after mass disasters),where remains can be badly burnt or degraded, usingsamples from maternal relatives for comparison.

Because mtDNA is relatively stable over generations itcan also be used to provide indications of ethnicity.


Some of the disadvantages of mtDNA analysisinclude:

u All maternal relatives being expected to have thesame mtDNA profile, complicating the interpretation of a match between a suspect and crime sample.

u MtDNA profiles being incompatible with STR profiles and not being able to be checked against The National DNA Database®.

u MtDNA analysis not being amenable to the analysis of mixed profiles, so that only items thatare known to have originated from one individual(e.g. bone, hair and faeces) are suitable to be processed.

In a similar vein to Y-STR’s, mtDNA can be used in theinvestigation of families by looking at the female line.


Dealing with the issues of contaminationAny contact between a suspect or victim and anobject or scene prior to a crime being committed canresult in DNA being transferred and later recovered.This is not, strictly speaking, contamination, but it is,of course, an important consideration that has to betaken into account.

The word “contamination” is used to denote thepresence of any DNA in a sample that has beenintroduced post-incident. This can be duringexamination of the scene or during collection oranalysis of the sample, either from the individualsinvolved, from contact with DNA from other items,from the materials used for sample collection oranalysis, or from the surrounding environment. It isnot associated with the offender or victim in question.

The FSS takes enormous care in the analysis ofsamples in the laboratory to ensure the integrity ofDNA profiles obtained from crime and referencesamples. This includes: the use of protective clothingby staff; very strict protocols for the handling andprocessing of samples, and environmental control, inline with international quality assurance guidelines;monitoring of the materials used to carry out analysesto ensure that they are DNA-free; processing of DNA-free blanks; checking for contamination betweensamples during analysis and for contamination bystaff; carrying out a proportion of analyses induplicate; and monitoring of the laboratoryenvironment and equipment.

This allows the FSS to minimise the risk ofcontamination and to take appropriate correctiveaction when it does occur. This action may include re-analysing the sample, if appropriate, or taking thecontamination into account when providingintelligence information to the investigator orpresenting evidence in court.


A similar degree of care must also be taken at thescene, where the risk of contamination is possiblygreater. Detectable levels of DNA can be depositedby investigators and those involved in the recovery ofevidence. By following best practice in crime scenemanagement (i.e. wearing gloves, facemasks, scenesuits and mob caps) the risk of compromisingpotential evidence can be greatly reduced.The FSS iscommitted to helping manage the contamination riskby providing quality controlled sampling kits andkeeping staff and investigators aware of anti-contamination practices through the use of literatureupdates, training courses and the Scenes of CrimeHandbook.

The identification of contamination by individuals inthe investigative and analytical processes is facilitatedby the availability of elimination databases. One ofthese, the police elimination database (PED),contains profiles of those police employees who arelikely to visit a crime scene or deal with thevictim/suspect. Profiles from named individuals on thePED are checked against a specific crime sampleprofile when it is considered that contamination bythat individual could have occurred. Another containsprofiles of FSS staff who may come into contact withcasework items or the facilities used for examinationor analysis. All profiles generated by the FSS areroutinely checked against this staff eliminationdatabase (SED).


The National DNADatabase® and therole of the custodianThe National DNA Database® was set up in 1995,following amendments to the Police and CriminalEvidence Act (PACE) 1984 by the Criminal Justiceand Public Order Act 1994. These amendmentsallowed CJ samples (cheek scrapes or rooted hairs)to be obtained for DNA profiling, in broadly the samecircumstances as fingerprints, and for the profilesobtained from these to be searched against recordsheld by or on behalf of the police. Furtheramendments to PACE were introduced by theCriminal Evidence Act 1997, the Criminal Justice andPolice Act 2001 and the Criminal Justice Act 2003.These extended the range of people from whomsamples could be taken, allowed for people tovolunteer to have their profiles added to The NationalDNA Database®, and removed the constraints onretention of samples and profiles when an individualwas not prosecuted or was acquitted.


The FSS was appointed custodian of The NationalDNA Database® by the Association of Chief PoliceOfficers (ACPO) for the first five years of its operationand this appointment has since been renewed. Underthe terms of a memorandum of understandingbetween ACPO and the FSS, the FSS provides the ITfacilities for holding The National DNA Database® anddevelops the software applications and procedures forgenerating the intelligence information required. Theactual profile data remains the property of theindividual police forces that obtained the samples foranalysis.

The operation of The National DNA Database® isoverseen by The National DNA Database Board,which meets quarterly and is chaired by the chiefconstable nominated by ACPO to lead on forensicscience matters.

Other law enforcement agencies outside the remit ofACPO (e.g. in Scotland, Isle of Man and Jersey) havealso agreed to abide by the memorandum ofunderstanding and contribute samples for profiling forThe National DNA Database®.

The FSS exercises the duties of custodian throughthe office of its chief scientist. As custodian, the chiefscientist is accountable to The National DNADatabase Board for maintaining standards of integrityand management of The National DNA Database®

and providing the range and levels of servicesspecified by the board.

The custodian sets the expected “Standards ofPerformance” for forensic science laboratories thatwish to provide DNA profiles to The National DNADatabase® and ensures that these standards areachieved and maintained. In addition to the FSS,there are now a number of other laboratories, in boththe public and private sectors, that do this work. Alluse a compatible technique based on the analysis ofSTR regions of DNA using SGM Plus technology. Thesuppliers meet regularly under the chairmanship ofthe custodian to discuss technical issues and newdevelopments.


CJ samples are classified as non-intimate samplesand can be taken by the police from anyone arrestedand in police detention, charged, about to be reportedfor, cautioned or convicted of a recordable offence.The mouth and hair samples are submitted by thepolice to an approved forensic science laboratory andthe laboratories submit the profiles from these to thecustodian for addition to The National DNADatabase®.

The police also obtain samples from biologicalmaterial left at scenes of crime. These samples aresubmitted for analysis to an approved forensicscience laboratory and the profiles obtained are alsothen passed to the custodian for loading to TheNational DNA Database®.

As each new profile is added to The National DNADatabase® it is routinely checked for matches againstall other profiles held on the database. Such matchesare reported as intelligence information to the police,linking a named individual with an unsolved crime orlinking different unsolved crimes together. Theintelligence matches may also be used as the basisfor charging a suspect, so long as there is sufficientother supporting information. If, on the basis of theDNA match and other information implicating theindividual with the offence, a prosecution is intended,the police are required to obtain a new sample andhave this analysed specifically for use in evidence.

A match between two full SGM Plus profiles fromunrelated individuals has a discriminating power ofabout one in one billion. Matches involving partialprofiles or relatives are more likely to occur by chanceand the discriminating power is thus much less.

The National DNA Database® was the first of its kindin the world and has received widespread acclaim asthe most important advance since fingerprinting in theprevention, detection and deterrence of crime. It is adynamic database, as profiles are constantly added toand removed from it.

Through the DNA expansion programme, managedby the Home Office, the government has investedconsiderable additional funding - £182.6m to the endof 2003/04 - to enable the police to increase their use


of DNA and through so doing to expand The NationalDNA Database®, so that it includes profiles from allactive criminals. The funding has also includedprovision for developments that improve the efficiencyand effectiveness of The National DNA Database®.

In May 2004, The National DNA Database® heldaround 2.3 million profiles from different individualsand around 229,000 profiles from scenes of crimesamples. Around 35,000 new subject sample profilesand 5,000 crime scene samples are added in a typicalmonth. In 2002/03, the matches reported to the policeidentified one or more suspects for 43,904 crimescenes. This was a 36% increase over the previousyear.

The chance of obtaining one or more suspects for anoffence when a crime sample is first added to thedatabase currently stands at about 45%. There is afurther probability, of about 25%, of a match beingobtained later as new subject sample profiles areadded to the database. The chance of identifying acrime scene with an individual when a subject sampleis first added to the database is about 4%.

Familial searchingIf a crime profile is checked against The National DNADatabase® and no match is obtained, it is possible,because there is a tendency for criminality to run infamilies, that the profile of a close relative may be onthe database.

The FSS has developed a technique for searchingThe National DNA Database® for profiles that couldrelate to the father/child and siblings of the offender. Incombination with other information the police mayhave available, it is possible to reduce the results ofthis search to a manageable number of individuals forfurther investigation.

As previously mentioned, the advanced techniques ofY-STR profiling and mtDNA sequencing can bebrought into play to investigate families identified bythis process.

This approach is of particular value in major crimeinvestigations. Further details are available from theFSS Forensic Intelligence Bureau (FIB), for contactnumbers (see page 33).


Principles for theevaluation of DNAevidenceAssessing the value of the DNAevidenceIf it is found that the DNA profile of a suspect isdifferent from that from a crime sample, then it isreasonable (processing errors excepted) and non-controversial to conclude that the DNA in the crimesample is not that of the suspect.

If, however, the two profiles are indistinguishable - amatch - then things are not nearly so straightforward.An obvious question is: Does that mean that the crimesample was left by the suspect? Unfortunately there isnot a simple answer. It is most helpful to view theissues within the context of possible futureproceedings. Then, if the origin of the DNA in thecrime sample is in dispute, prosecution and defencewill have opposing viewpoints.

We could expect the prosecution to invite the jury tobelieve a proposition of the following kind:

The DNA in the crime sample came from thedefendant.

The defence view would be that the DNA in the crimesample did NOT come from the defendant. Since theDNA must have come from someone we couldformulate a defence proposition of the following kind:

The DNA in the crime sample came from someunknown person.

The jurors will, ultimately, have to decide which ofthese two propositions they believe to be the correctone. They will have heard not only of the DNA match,but also of other non-scientific evidence, includingevidence of opportunity (i.e. whether the defendantcould have been at the scene of crime at the


appropriate time) and whatever the defendant says inhis/her defence (e.g. his/her alibi). The importantquestion is how the scientist can assist the jury in alogical, balanced and robust fashion to reach theirconclusion.

It is now widely accepted among statisticians andlegal scholars who have considered the problem thatthe scientist should, in the present context, addressthe following questions of the kind:

What is the probability of a DNA match if theprosecution proposition were true?

What is the probability of a DNA match if the defenceproposition were true?

In this particular case (though not so in the case ofDNA mixtures) the answer to the first question is quitesimple - a match is exactly what we would expect tosee. So we are left with the second question which wecan rephrase as:

What is the probability of a DNA match if the DNA inthe crime sample had come from some unknownperson?

This is not such a simple question because it isnecessary to clarify what is meant by an “unknownperson”. As each person’s DNA is inherited from theirmother and father it follows that people who areclosely related are more likely to share the sameprofile than people who are completely unrelated.Indeed, identical twins will almost certainly have thesame profile. The probability that two full siblings willhave the same SGM Plus profile is of the order one in10,000 and the corresponding probability for a pair offirst cousins is of the order one in 100 million. Forunrelated people, it is current FSS policy to quote amatch probability of one in a billion - for reasonsexplained later.

It follows from this that, if the scientist is going to helpthe jury effectively, he/she needs to be provided withwhatever information is available to sharpen up thenotion of an “unknown person”. If it is known, forexample, that the defendant has close siblings who


might realistically be considered as suspects for theoffence, then the best way to minimise uncertaintywould be to take DNA samples from those siblings tosee if they can be eliminated. If that is not possible,then the scientist can guide the jury with the variousrelevant match probabilities. But it would be wrong tosuggest that the jury have a simple task.

If we assume that the circumstances are such thatclose relatives are not an issue, then the question tobe addressed in relation to the defence proposition is:

What is the probability of a DNA match if the DNA inthe crime sample had come from some unknownperson, unrelated to the defendant?

As the current FSS policy is to quote a matchprobability in answer to this question of the order onein a billion, written in full, the sort of statement thescientist would make is:

If the DNA in the crime sample had come from someunknown person unrelated to the defendant, theprobability of a match would be of the order one in abillion.

Match probability of one in abillionWe have seen that a full SGM Plus profile consists often loci (we leave the gender marker to one side forthe time being) and at each locus two alleles will beidentified - one from the individual’s mother and onefrom the father. If we wished to calculate a matchprobability for a single locus, we would refer to adatabase of samples taken from people from arelevant population to estimate how rare or howcommon the two components are. Then we wouldcarry out a calculation of a match probability for thatparticular locus. The database that we would useneed not be very large - typically one of 100 to 200individuals is sufficient to provide robust estimates forsingle locus match probabilities. The calculation iscarried out using formulae that are recognised assensible by a large international consensus ofscientists and statisticians. The formulae include a


correction factor that allows for the possibility thathuman populations may be, to a small extent,structured by the existence of groups that tend not tomix with other groups.

So, given a suitable database, we could carry out amatch probability calculation for each of the ten loci -these individual match probabilities would be of theorder one in 10 to one in 100.

We then face the question of how we are to combinethese probabilities. There is an internationalconsensus in favour of the practice of multiplyingthem all together and this leads to unbelievably smallmatch probabilities - perhaps of the order one in atrillion (one followed by 12 zeros)) or one in aquadrillion (one followed by 15 zeros).

The FSS, however, does not follow this practicebecause it is considered that current understanding ofthe population genetic factors involved does notwarrant such extravagantly infinitesimally smallnumbers. Instead, if we have a match between two fullSGM Plus profiles we do not carry out a case specificmatch probability calculation. Instead, if we areconsidering the alternative proposition of an unknownperson, unrelated to the suspect, then we quote amatch probability “of the order one in a billion” (abillion is a thousand million, or one followed by ninezeros). This is independent of any particular databaseand we quote the same figure whatever the ethnicgroup of the unknown person. The justification for thisprocedure was published in the open literature severalyears ago and, to our knowledge, it has never beenquestioned.

Reporting a DNA match to courtAs we have seen, interpreting the significance of amatch between two DNA profiles depends on thealternative propositions to be addressed. If, at thetime of writing a statement, the scientist is presentedwith a clear defence alternative, the appropriatematch probability will be quoted. However, in manycases this is not the position and the question ofpossible relatives of the suspect is not clear to the


scientist. In such cases, the FSS will report the matchwithout providing a single match probability. Instead, arange of match probabilities for people of differentdegrees of relationship with the suspect will be given.

If, later, court proceedings are envisaged and theposition with regard to alternative propositionsbecomes clearer, then the scientist will provide anadditional statement that addresses the appropriatepropositions by means of the appropriate matchprobabilities.

Uniqueness probabilityA question that is often asked is: Can a DNA matchprove identity in the absence of any other evidence?But is there ever a case in which there is no otherevidence? At the very least, there will be evidencerelating to time/place of the crime and someinformation, however vague and unspecific, relating tothe defendant’s whereabouts around that time.

There may indeed, as in R v Dennis John Adams, beeyewitness and alibi evidence on behalf of thedefendant. The Appeal Court, in R v G Adams andDoheny, took a rather simplistic view of the problemby suggesting that it was sensible to consider thenumber of other people in some population or otherwho might be expected to possess the same givenprofile as the defendant.

A rather naive approach would be to ask the scientist:How many other people in the world would beexpected to possess the same DNA profile as thedefendant? Putting the issue of close relatives to oneside for the present, a match probability of one in abillion implies that one would expect there to be of theorder ten other people in the world with the sameprofile. It is difficult to see how this helps, however,because there will never be a case in which it wouldbe realistic to consider the entire population of theworld as potential suspects for a given crime. If thecrime is a rape in Huddersfield, would we seriouslyconsider a 70-year-old female resident of Beijing asuspect for the crime? This consideration appears tohave been recognised in R v Doheny and Adams1


where it was suggested that a starting point might bethe British population - or, in a rape case, the malepopulation of the country. But here too, it would seemunrealistic in a rape case to consider all the males ofthe country as suspects.

The judgement in the above case suggestsaddressing a more restricted population, determinedby the case circumstances. The judgement gives thesexually active male population of Manchester as anexample. This brings us to another limitation. Imaginethat the court has decided that it is reasonable, giventhe case circumstances, to consider a population of amillion males. Then, in answer to the question: Howmany other males in this population would beexpected to share the given DNA profile?, the matchprobability of one in a billion suggests the answer isone thousandth of a person. This clearly would besomething that a jury may be expected to havedifficulty in understanding and we need to follow abetter approach to the problem.

It is possible to do this by calculating what is known asa “uniqueness probability”. To do this, the scientistwould need to be given a guide with regard to the sizeand nature of the population to consider. Here it isnecessary to include consideration of the possibility ofclose relatives. If, for example, the defendant has twobrothers, who have not been excluded by profiling, yetwho might be realistically be considered to bealternative suspects, then an allowance for twobrothers can be included in the calculation. Also, thescientist could factor in a generous provision forrelatives such as uncles and first cousins.

1 Doheny and Adams [1997] 1 Cr. App. R 369. Confusingly, thejudgement adopted the phrase “random occurrence ratio” which is notonly inaccurate but also has no technical or scientific pedigree. Thereare a few phrases that are more accurate but the consensus viewamong British statisticians who have studied this field is that “matchprobability” is the most appropriate. From a scientific viewpoint, thephrase “random occurrence ratio” should be avoided.


Scientist’s statementIn the past, it has been frequent practice for thescientist to add some sort of verbal explanation (suchas “extremely strong support”) to convey the weight ofevidence to be associated with a DNA match. Forvarious reasons this practice is being discouraged.The statement that the scientist will initially providewill depend on the information available. If there is aclear pair of competing alternative propositions thescientist will quote the appropriate match probability.In a case where there is no question of close relativesof the suspect to be considered then this will be “ofthe order one in a billion”. If there is no clearalternative, and particularly in the case of nocomment from the suspect, the match will be reportedby means of an abbreviated statement together with astandard paragraph explaining the ranges of matchprobabilities that might be relevant.

In the event of clear competing alternativepropositions later emerging, the scientist can providean additional appropriately worded statement takingaccount of the updated information.

The prosecutor’s fallacyPresenting probabilistic evidence in court has manypitfalls but there is one, in particular, that has led tosuccessful appeals (e.g. in Doheny and Adams).Consider, for example, where there is a full SGM Plusmatch between the defendant and semen recoveredfrom a rape victim and, for simplicity, assume that allconsiderations of possible close relatives have beenexcluded by one means or another. Then the scientistmight say:

“The probability of a match if the semen came fromanother person is one in a billion.”

In an effort to simplify things for the benefit of the jury,counsel might be tempted to invite the scientist toagree to the following paraphrase:

The probability that the semen came from anotherperson is one in a billion.


The second sentence is undoubtedly simpler but itdoes not follow from the first sentence. This iswhat is known as the “prosecutor’s fallacy”, thoughstatisticians know it as the “fallacy of the transposedconditional”. This is, undoubtedly, what the learnedjudges had in mind in the judgement in Doheny andAdams where they say:

“The scientist should not be asked his opinion on thelikelihood that it was the defendant who left the crimestain, nor when giving evidence should he useterminology which may lead the jury to believe that heis expressing such an opinion.”

As a general rule, the role of the scientist is to advisethe court of the probability of the evidence given theproposition, in this example, of the probability of amatch if the semen came from some unknownperson. It is properly the role of the jury to considerthe probability of the proposition given the evidence,in this example, the probability that the semen didcome from some unknown person.

Hierarchy of propositionsOver recent years, the FSS has given much attentionto forming a clear logical framework for presentingscientific evidence in a balanced and robust fashion.This is being taken forward through a project entitledCase Assessment and Interpretation (CAI). One ofthe key ideas to arise from this project work has beenthe formulation of a “hierarchy of propositions”.Consider the following set of propositions:

i) The defendant assaulted Mr Z.The defendant had nothing to do with the assaultof Mr Z.

ii) The defendant kicked Mr Z in the head.The defendant was not present when Mr Z waskicked in the head.

iii) The blood on the defendant’s clothing came fromMr Z.The blood on the defendant’s clothing came fromsome unknown person.


Here we see a progression. The first pair are clearlythe level of propositions that might be addressed by ajury and are thus called offence level propositions.The pair of propositions at (ii) are termed activitylevel propositions. Finally the progression movesdown to a level at (iii) that is comparable to thoseconsidered so far in this guide, termed source levelpropositions.

Clearly, the scientist has a role in assisting the courtin explaining how the scientific evidence relates to apair of source level propositions. But there might alsobe cases where, because of a combination ofcircumstances, the scientist might be able to assist inaddressing activity level propositions. In the abovecase, for example, the scientist might be able to bringexpertise to bear on the interpretation of the pattern ofthe blood staining on the defendant’s clothing.

In general, the higher up the level of propositions thatcan be addressed, i.e. closest to offence level, thegreater the assistance the scientist can give to thecourt. However, it is essential to realise that the levelat which the scientist can assist is determined to alarge extent by the information that is made availableto him/her. In particular, it is important to knowanything that the defendant might be saying that isrelevant to formulating the propositions.

With the increase in sensitivity in DNA profiling thathas been seen in recent years, scientists have had toallow for another eventuality. In certain circumstancesit may not be possible for the scientist to be confidentthat the DNA profile that has been generated hasactually come from the biological material that it isthought has been tested. For example, if this materialis a degraded or weak stain and has not given a DNAprofile but an underlying body fluid or a trace ofcellular DNA has given a profile.

Hence we also have sub level one propositions:

The DNA on the defendant’s clothing came from Mr Z.

The DNA on the defendant’s clothing came fromsome unknown person.


The evaluation of DNA evidencein database matchesThe overwhelming success of The National DNADatabase® in the UK is now well known, not only by allthose involved in the criminal justice system, but alsoin the public eye, due to numerous convictions in highprofile cases.

All new profiles (from CJ samples, suspects, orundetected crime scene stains) loaded onto TheNational DNA Database® are searched against all theprofiles already held and also against each other. Thepolice are then automatically informed of anymatches. The term match rather than hit is preferred,as the latter tends to imply a profile being connectedto a crime(s) which may not always be the case. Theinformation from The National DNA Database® is forintelligence purposes only, so any matches linking asuspect to a scene must be followed up by taking acontrol sample (usually a buccal scrape) from thesuspect to confirm the result and to be used inevidence.

Given the size of the The National DNA Database® itis quite possible that SGM profiles from two unrelatedindividuals will match by chance, this is a so-called‘adventitious match’. However, given the greatlyincreased discriminating power, such occurrencescan be expected to be very rare with SGM Plus.

Some press reports have suggested that theseadventitious matches are mistakes that may have ledto the conviction of innocent people. They are notmistakes, they are predictable occurrences. Themistake occurs if the possibility of the match beingadventitious is not considered and individuals arearrested and prosecuted without first investigatingalibis, motives, geographical opportunity or, whereappropriate, upgrading the profiles involved to SGMPlus profiles to reduce the risk of the match beingadventitious.

Police must therefore be aware of the consequencesof adventitious matches and that DNA evidence is notused in criminal prosecution without careful


consideration of the surrounding circumstances.Indeed, this was one of the findings in the AppealCourt hearing of R. v Lashley (2000), where DNAevidence had been relied upon for a conviction. Therewas no evidence that the appellant had been in thevicinity of where the crime had been committed andhence the DNA (SGM only) profile was just as likely tohave been left by any one of five or six other men inthe United Kingdom.


Presentation of theevidence andstatementsDNA evidence has been presented successfully incourts since 1988. Over that time the techniquesinvolved in DNA profiling have been continuouslychanged and improved.

The primary means of communicating the findings ofthe forensic scientist to the court is throughstatements. The format and style in which thestatement is written is vitally important and these arealso kept constantly under review by the FSS. Theintroduction of new techniques as well asdevelopments in the criminal justice system may wellrequire changes to be made in the future.

Areas of challenge initially concentrated on thecontinuity of sample handling, the reliability of theactual techniques and the strength of DNA matchprobabilities. Now, however, they are more to do withthe relevance of the DNA evidence in relation to therest of the case when taken into account with theother, non-DNA, evidence.

A change to Crown Prosecution Service (CPS) policythat provides for suspects to be charged on the basisof an intelligence match on The National DNADatabase®, providing there is some furthersupporting evidence, has been introduced (CPSPolicy Directorate, August 2004). The implementationof this policy is supported by guidance provided byboth the CPS and the Association of Chief PoliceOfficers. Where this prosecution process is beingfollowed further evidential material will only berequested from a forensic scientist where “noindication of plea” or a “not guilty plea” has beenentered. Such additional work will be performed instages at the request of the prosecution team. Onsubmission of the reference sample an abbreviatedstatement confirming the match will be provided.Further statements dealing with continuity andhearsay can be requested where admissions onthese matters are not obtained from the defence. Full



evaluative statements will normally only be providedwhere the DNA evidence is to be contested at trialand after a defence statement is available.

The FSS recommends that if it becomes apparentthat the DNA evidence will be a significant issue attrial the scientist is contacted. We believe that it is vitalthat the scientist knows the background information,and any other relevant evidence, in order thatappropriate alternatives can be considered. Caseconferences are an ideal way of talking through theissues, setting the DNA evidence into context,deciding if further work or analysis is required anddiscussing how the evidence could be presented. Ifvisual aids are required for the presentation of theevidence, the form of these and the resourcesrequired can also be discussed at the caseconference.

Further advice andsuggested readingFurther advice on any particular case can be obtainedfrom the scientist concerned.

General enquiries and requests for further copies ofthis booklet can be directed to the FSS customerservices helpdesk on 0121 607 6996/7/8.

Specific enquiries relating to The National DNADatabase®, e.g. match reports, sample retentionpolicies and submissions should be directed to TheNational DNA Database® helpdesk on 0121 606 2950.

Other booklets available from the FSS, whichsupplement this booklet, include:

u DNA Present and Correct - A guide to DNA Profiling

u The Scenes of Crime Handbook 2004u DNA Profiling: A discussion of issues relating

to the reporting of very small match probabilities - Criminal Law Review; 2000,341-355

u Lawyers’ Guide to DNA (previous version of this booklet)

AppendixGlossary of terms and abbreviations

Adventitious match DNA profiles from twoindividuals, who are not identicaltwins, that match by chance.

CJ (criminal justice) sample Non-intimate sample (cheekscrape or hair root) lawfully takenfrom a suspect arrested, chargedor convicted for a recordableoffence for intelligence purposes

Reference sample Sample taken from a person forcomparison with a specific crimesample and used as evidence ina court of law.

Crime sample Sample recovered from a sceneof crime or victim from whichDNA can be obtained forintelligence or evidential use.

DNA Deoxyribonucleic Acid.

Familial searching Utilises the fact that elements ofDNA are inherited. Unknowncrime profiles can be searchedagainst The National DNADatabase® to return profiles ofpotential relatives of the offender.

FSS* The Forensic Science Service.

LCN Low Copy Number - analysis andinterpretation process used forsamples containing extremelysmall amounts of DNA.

Likelihood ratio The probability of the evidencegiven the prosecution propositiondivided by the probability of theevidence given the defenceproposition.

Match probability The probability that an unknownindividual will have a particularprofile, given that a knownindividual has that profile.

MtDNA Mitochondrial DNA.


The National DNA Database®* Relational database holding DNAprofiles from CJ samples,reference samples, volunteersamples and undetected crimesamples that are searchedagainst each other and all newprofiles that are loaded onto thedatabase.

Pendulum list search (PLS) A technique for identifying andsearching individual DNA profilesthat are contained in a two-component mixed profile.

PCR Polymerase chain reaction -technique used for targeting andamplifying specific regions ofDNA.

SGM Second generation multiplex -profiling system developed andutilised by the FSS from 1995 -1999.

SGM Plus** Profiling system utilised by theFSS since July 1999. A moresensitive and discriminatingsystem than SGM but compatiblewith SGM.

STR Short tandem repeat - non-coding region of DNA targeted inthe SGM/ SGM Plus profilingmethods which varies in lengthbetween individuals.

Y-STRs Short tandem repeats containedin the male Y chromosome innuclear DNA.

* The Forensic Science Service®, FSS® and The National DNADatabase® are UK registered trademarks.

**AmpFlSTR® SGM Plus is a registered trademark of AppliedBiosystems, a division of the Perkin Elmer Corporation.

© Crown Copyright 2004. All Rights Reserved.produced by the FSS communications department.

Date post:	17-Jan-2020
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

Guide to DNA · 2018-06-08 · DNA profile matching that of another just by chance. It is thus not...

Documents