|Neutral Citation No.  NICA 7||Ref:||MOR8471|
|Judgment: approved by the Court for handing down||Delivered:||20/04/12|
|(subject to editorial corrections)*|
IN HER MAJESTY’S COURT OF APPEAL IN NORTHERN IRELAND
ALISON MICHELLE MARTIN
Before: Morgan LCJ, Coghlin LJ and McLaughlin J
 The applicant was charged with the murder of George McDowell between 27 July and 30 July 2004. She pleaded not guilty at arraignment on 3 June 2005. On 1 November 2006 she was found guilty of murder and sentenced to life imprisonment. On 13 December 2006 Hart J set a life sentence tariff of 12 years. The applicant seeks leave to appeal against her conviction. She argues that the learned trial Judge erred in admitting DNA evidence and having done so failed to adequately warn the jury as to the need for caution in the circumstances. She further argues that the learned trial Judge erred in admitting evidence of the applicant’s bad character, namely that she often carried a knife, and having done so failed to give the jury appropriate guidance in relation to that evidence. Ms Quinlivan QC appeared with Mr Hutton for the applicant. Mr Ramsey QC appeared with Mr McCrudden for the Crown. We are grateful to counsel for their helpful oral and written submissions.
 On Thursday 29 July 2004 at around 9 p.m. the body of George McDowell was found in the living room of a flat at 14c Crossreagh Drive, Rathcoole, the home of John Nixon. Nixon said that he had woken up and found the victim dead, went to the home of a neighbour and phoned the emergency services. The autopsy report concluded that the victim had died as a result of a stab wound to the head. The entry wound was immediately in front of the left ear and passed fairly steeply downwards through the left side of the neck into the throat severing the external carotid artery. This resulted in brisk and copious bleeding. There was a second stab wound on the outer part of the left cheek which was superficial and another superficial incised wound on the left side of the scalp above the ear which would not have been life-threatening. The victim had a high concentration of alcohol equivalent to almost 5 times the legal limit for driving. Once the main injury had been sustained the victim would not have died immediately but could have carried out actions for a short period of time.
 There were several strands to the prosecution case. The first was the direct evidence of Mr Nixon. He is an alcoholic and it is common case that he had been drinking heavily throughout the period in question. The deceased had been staying at his house and was also drinking heavily in the days prior to his death. Nixon said that the applicant had brought the deceased to his house and asked Nixon to give him a bed for a few nights. In the days prior to the murder the applicant was also drinking heavily and staying with a friend. She initially maintained that she had not visited Nixon’s home after the Monday before the murder but subsequently accepted that she had been there on the Tuesday morning before the murder and also for a few minutes around 9 p.m. on Wednesday evening.
 Nixon’s evidence was that on the night of Thursday 29 July 2004 the applicant entered his flat and told the deceased to get out of it. She then went into the kitchen and came out with a knife in her right hand. She stabbed the deceased in the left side of the neck below the left ear. Nixon got the deceased a towel which he put to his neck and the accused left because Nixon told her to go. Nixon said that he then fell asleep leaving the deceased in the chair. He woke up shortly before 9 p.m. and saw the deceased lying facedown on the floor dead. Although the prosecution relied upon Nixon’s evidence to establish the actions of the applicant the prosecution case was that the murder occurred in the early hours of 29 July 2004 shortly after midnight.
 There were a number of discrepancies in Nixon’s account. Before the police arrived Nixon told his brother that he had been drinking with the applicant and the deceased in his flat and that he fell asleep on the sofa. His evidence was that he had gone to sleep at 5:30 p.m. on Thursday but he gave an account to police arriving at the scene indicating that this event happened about midnight on Wednesday and that he had gone to sleep at 5:30 a.m. There was also evidence that he had been outside the flat between 7:30 p.m. and 9 p.m. on the Thursday evening which contradicted his evidence that he was sleeping during that period. In directing the jury the learned trial Judge suggested to the jury that in light of the contradictions in Mr Nixon’s evidence he did not think that the jury could be satisfied beyond reasonable doubt on that evidence alone about the applicant’s guilt. Indeed the case made on behalf of the applicant was that Nixon had committed the murder or was covering up for someone who did it.
 The second strand of evidence related to the forensic material. Throughout the period in question the applicant stated that she wore a pair of Puma trainers. A bloodied footprint had been found on the door saddle into the kitchen. The forensic witness examined the Puma trainers and concluded that the wearer of the right-hand trainer had stepped on the door saddle into the kitchen and that there was blood on the sole of the shoe before the wearer stepped on the door saddle. His conclusion was based on the unique wear on the sole of the trainers which he said provided very strong support for this conclusion. The prosecution argued that this evidence helped to establish that the applicant had stepped on blood and then stepped on the door saddle.
 When the trainers were examined the uppers gave the appearance of being recently cleaned. The forensic witness found staining in three locations, on the stitching, on the inner welt and on the outer welt. A partial DNA profile from the staining was found to have genetic characteristics in common with the deceased’s DNA profile and the likelihood of an unrelated male from Northern Ireland having this combination of characteristics was less than one in 50,000.
 The material recovered from the door saddle was examined by Mrs Cherry, a forensic scientist experienced in the interpretation of low copy number DNA analysis. She concluded that most, though not all, of the DNA characteristics of the deceased were found in a mixture of the sample taken from the saddle board. She concluded that this provided some unquantifiable level of support for the view that the deceased had contributed to this mixture. It was apparent that there was DNA from at least two people present and she was asked to compare the sample with Mr Nixon’s DNA. She found some of his DNA characteristics present and accepted that it was possible that he had contributed to the sample but much less clear. If he had there must have been some contribution by at least one more person. She was asked if her findings were consistent with the trainers stepping on the deceased’s blood and then transferring it to the saddle board. She stated that if the blood was wet and deposited then the findings she obtained might be expected. She emphasised, however, that this was only one of the possibilities. The validity of the conclusion reached by Mrs Cherry that there was some support for the contribution of the deceased to the mixture, its admissibility in evidence and the directions of the learned trial judge on this evidence lie at the heart of this appeal.
 There was disputed evidence that the applicant had washed one pair of jeans and bleached another pair of jeans late on the Wednesday evening, 28 July, or in the early hours of Thursday morning. The jeans were recovered from a house in which the applicant had been staying. Forensic examination of the jeans which had been washed but not bleached established extensive light and diffuse staining on the outside front of both legs and heavier staining on the right knee and the right rear hem both of which give a positive reaction indicating the presence of blood. There was, however, no evidence to show that the blood was from the deceased. The learned trial judge noted that the prosecution case was that this was evidence of an attempt by the accused to cover her tracks and he identified this as a separate strand to the jury.
 The applicant was the habitual wearer of a black jacket. A small drop of blood was found to the right of the zip in the abdominal area. It is accepted that the jacket had to be open when the spot of blood landed on it. DNA analysis provided a full profile that matched the deceased’s DNA and the likelihood of some other unrelated person having the same characteristics was approximately one in a billion. Police also recovered a white Rangers T-shirt from the flat in which she had been staying and found spots of blood on the front centre of the T-shirt and smears of blood towards the waist area. The applicant denied that the shirt belonged to her but the occupant of the house in which she had been staying gave evidence that he found it there shortly after she left.
 The third strand to the prosecution case was the series of lies which the prosecution say she told either to the police at interview or in evidence. We do not need to rehearse all of the issues raised before the jury but there are four matters which appear to us to be material. The first matter related to her contact with the deceased. At interview she initially said that she had left the deceased at Nixon’s flat about 5 p.m. on Monday and had not seen him thereafter. She subsequently accepted in evidence that she returned to the flat for a few minutes on Tuesday to bring cigarettes and for a few minutes on Wednesday evening. She was unable to explain why she did not tell the police about these visits but said that she was confused.
 During her police interview she stated that she did not change her clothes between Monday 26 July and Friday 30 July. She further stated that she did not wash her clothes at all during this period. In her evidence she then asserted that she had washed a pair of white bottoms to remove gravy stains on Wednesday. The prosecution case was that she had washed the jeans on which blood residue was found at about midnight on Wednesday and this was an attempt by her to hide her involvement with the crime scene. She had also washed and bleached a pair of white jeans. She could not explain why the friend with whom she was staying would make such an allegation against her if it were not true. She agreed that she had used bleach on the toilet but denied that she had used it on the jeans.
 She contended that Nixon had sold her his mobile phone on Tuesday morning for £35 in order to buy more drink. The applicant used this phone to make a call to a friend at 11:10 p.m. on Wednesday 28th July. It seems clear, therefore, that the phone was in her possession by that time. The telephone records also demonstrate that there was a call from this phone to Nixon’s brother at 11:41 a.m. on 28th July and a further call at 8:53 p.m. on the same date to a taxi firm seeking a taxi to and from an off licence from Crossreagh Drive, Nixon’s home. The applicant said that it was possible that while she was at Nixon’s flat he had asked her to ring the taxi firm but this does not accord with her account that when she arrived at the flat on Wednesday evening the occupants were so drunk that she could not converse with them. The prosecution case was that the call to Nixon’s brother and the call to the taxi firm indicated that the phone was still in Nixon’s possession until Wednesday evening.
 In her evidence the applicant alleged that she was in the company of the deceased on Thursday 22 July in the city centre when the deceased’s nose started to bleed heavily. She said that she gave the deceased a handkerchief. This evidence was clearly directed towards explaining the presence of blood on the applicant’s clothing. In cross-examination she agreed that her account of the deceased sneezing and bleeding heavily was brand new. In fact it is clear from her interview with police that she described something quite different. She described the deceased having a trickle of blood coming down from his nostril. She agreed at the interview that it was not a really bad nose bleed but was just a wee trickle. On the basis of this description it is difficult to see how any blood from the deceased could have migrated to the applicant’s clothing.
 The fourth strand of evidence against her related to her bad character. In August 1999 she assaulted a female using a fork. The woman had her hair pulled out and the defendant ended up with blood on her hands and clothing. She was made subject to a hospital order having been charged with causing grievous bodily harm with intent. In August 2000 she struck a male on the head with a bottle, then returned from the kitchen with a knife and slashed his face. She also pulled the hair out of a female’s head on the same occasion. The learned trial Judge directed the jury that these were material to the issue of whether she had a propensity to commit crimes of violence. There is no issue about the admission of those convictions for that purpose in this appeal.
 Evidence was also given by her sister who said that she was fearful of the applicant and that she had seen the applicant carrying different kinds of knives. This evidence was not challenged in cross-examination. Nevertheless the learned trial Judge pointed out to the jury that it was clear that there was no love lost between the applicant and her sister and that they should bear that in mind when considering to what extent they could rely on the sister’s evidence. The admissibility of this evidence is challenged.
 DNA STR profiling is a form of DNA analysis which targets and copies specific regions of a person’s DNA which are known to vary in size between individuals. The system generally in use at present targets 10 different loci. At each location there are usually two bands or alleles. One of the two alleles at each locus is inherited from each parent. The bands of two DNA profiles of known and unknown origin can be compared and if they are different the donor of the known profile can be excluded as the donor of the unknown profile. In the event of a match being obtained it is possible to estimate the chance that a person taken at random from the population would have a matching profile.
 Low copy number DNA profiling is a more sensitive technique which requires a smaller amount of biological material for successful analysis. Due to the greater sensitivity of low copy number profiling mixed profiles indicating the presence of DNA from more than one individual are frequently observed. These can be complex and therefore difficult to interpret statistically.
The applicant’s case on the DNA issue
 Ms Quinlivan submitted that the learned trial judge was wrong to admit the evidence of Mrs Cherry that the DNA findings in respect of the sample taken from the saddle board provided some unquantifiable level of support for the view that the deceased had contributed to this mixture. The applicant contended that the significance and evidential weight of DNA evidence lay in its match probability and in the random occurrence ratio. In support of this submission the applicant relied on R v Doheny  1 Cr App R 369 and the suggestions made in that case as to how experts should deal with DNA evidence. In any event it was submitted that in the absence of any means of quantifying the level of support from these findings no weight could be attributed to the conclusion.
 In support of this application it was submitted that the court should admit fresh evidence from David Hayward, an expert in the interpretation of DNA profiling results. The proposed evidence was set out in a statement dated 10 June 2011. At paragraph 6.1 of his report Mr Hayward noted that all but two of the deceased’s DNA components were present within the mixture on the saddle board. He stated that this was a significant finding indicating that the deceased may be a source of DNA present within the mixture. He agreed that Ms Cherry was unable to present any statistical basis for the degree of support for her conclusion but accepted that the evidence was clearly greater than no support.
 At paragraph 6.5 of his report Mr Hayward noted that a partial DNA profile matching the deceased was obtained using standard DNA profiling techniques in respect of the shoe that matched the footwear mark on the saddle board. Mr Hayward concluded that this tended to lend support for the view that any DNA attributable to the deceased did in fact come from blood on the shoe rather than by some other route.
 On the morning of the hearing of the appeal Mr Hayward and Mrs Cherry had an opportunity to discuss their differences. In response to Mr Hayward’s report Mrs Cherry had produced a report stating that the DNA results could be described as providing at least limited support for the view that the deceased had contributed DNA to the mixed profile found in the saddle board. Mr Hayward was content with that description and by agreement this evidence was admitted.
 The second argument advanced on behalf of the applicant was that the learned trial Judge had overstated the weight to be given to the DNA evidence in his charge to the jury. The applicant accepted that the judge accurately told the jury that Mrs Cherry’s opinion was that the mixed sample provided some unquantifiable level of support for the view that the deceased had contributed to the mixture. He then said that Mrs Cherry accepted that this meant that this was not a conclusive finding that his DNA was in that sample. This form of words was one with which Mrs Cherry had agreed when questioned by the judge at the end of her evidence but it was submitted that in light of the agreement which was reached between the experts on the morning of the appeal there is a real risk that the jury were misled as to the strength of the inference which they could draw from this evidence.
The relevant case law on expert opinion
 In R v Doheny and Adams  1 Cr App R 369 the Court of Appeal in England and Wales gave guidance on the procedure to be followed in relation to the admission and treatment of DNA STR profiling evidence. That approach has been regularly followed in this jurisdiction and we do not need to set it out here. The court in that case was anxious to demonstrate what has been called the prosecutor’s fallacy. If one person in a million has a DNA profile which matches that obtained from the crime stain then the suspect will be one of perhaps 26 men in the United Kingdom to share that characteristic. If no fact is known about the defendant, other than that he was in the United Kingdom at the time of the crime, the DNA evidence tells us no more than that there is a statistical probability that he was the criminal of one in 26.
 That court also looked at the nature of the expert evidence which the scientist can give.
“When the scientist gives evidence it is important that he should not overstep the line which separates his province from that of the jury.
He will properly explain to the jury the nature of the match (“the matching DNA characteristics”) between the DNA in the crime stain and the DNA in the blood sample taken from the defendant. He will properly, on the basis of empirical statistical data, give the jury the random occurrence ratio—the frequency with which the matching DNA characteristics are likely to be found in the population at large. Provided that he has the necessary data, and the statistical expertise, it may be appropriate for him then to say how many people with the matching characteristics are likely to be found in the United Kingdom—or perhaps in a more limited relevant sub-group, such as, for instance, the caucasian, sexually active males in the Manchester area.
This will often be the limit of the evidence which he can properly and usefully give. It will then be for the jury to decide, having regard to all the relevant evidence, whether they are sure that it was the defendant who left the crime stain, or whether it is possible that it was left by someone else with the same matching DNA characteristics.
The scientist should not be asked his opinion on the likelihood that it was the defendant who left the crime stain, nor when giving evidence should he use terminology which may lead the jury to believe that he is expressing such an opinion.”
The standard JSB direction for identification by DNA is derived from that case.
 In R v Bates  EWCA Crim 1395 the court was faced with a different problem. A number of partial profiles had been obtained but none of the samples tested showed any alleles at a number of loci. The expert statistician called on behalf of the prosecution ignored the voids and calculated that one person in every 610,000 had the particular combination of the alleles found in the relevant samples. The defence objected on the basis that the voids might have contained exculpatory material which it was not possible to quantify statistically. In any event the jury were not in a position to form a view on something which was inherently unquantifiable.
 The court rejected the submission and set out its reasons at paragraph 30.
“We can see no reason why partial profile DNA evidence should not be admissible provided that the jury are made aware of its inherent limitations and are given a sufficient explanation to enable them to evaluate it. There may be cases where the match probability in relation to all the samples tested is so great that the judge would consider its probative value to be minimal and decide to exclude the evidence in the exercise of his discretion, but this gives rise to no new question of principle and can be left for decision on a case by case basis. However, the fact that there exists in the case of all partial profile evidence the possibility that a “missing” allele might exculpate the accused altogether does not provide sufficient grounds for rejecting such evidence. In many cases there is a possibility (at least in theory) that evidence exists which would assist the accused and perhaps even exculpate him altogether, but that does not provide grounds for excluding relevant evidence that is available and otherwise admissible, though it does make it important to ensure that the jury are given sufficient information to enable them to evaluate that evidence properly. Moreover, as the court observed in Doheny and Adams, the significance of DNA evidence depends to a large extent upon the other evidence in the case. By itself such evidence, particularly if based on a partial profile, may not take the matter far, but in conjunction with other evidence it may be of considerable significance.”
 In R v Atkins and Atkins [2009} EWCA Crim 1876 an expert in facial comparison (facial mapping) gave evidence at the trial of two of the defendants, each of whom denied being present at the alleged robberies, involving comparison of the CCTV image and a photograph of the first defendant. The expert recited the detail of consistent features and the absence of any inconsistencies and explained his techniques, next expressing his conclusions by reference to a range of expressions set out in his written report in tabular form and ranging from “(0) Lends no support” to “(5) Lends powerful support”. He concluded that the similarities identified gave support to the proposition that the man on the CCTV image was the first defendant which he put at “the top of (3) [Lends support] and into (4) [Lends strong support]”. The defence applied to the judge for this aspect of the expert’s evidence to be excluded as inadmissible. The court rejected that submission.
“23. On principle, we accept the caution with which any expression of conclusion in relation to evidence of this kind (and others) needs to be approached. We agree that the fact that a conclusion is not based upon a statistical database recording the incidence of the features compared as they appear in the population at large needs to be made crystal clear to the jury. But we do not agree that the absence of such a database means that no opinion can be expressed by the witness beyond rehearsing his examination of the photographs. An expert who spends years studying this kind of comparison can properly form a judgment as to the significance of what he has found in any particular case. It is a judgment based on his experience. A jury is entitled to be informed of his assessment. The alternative, of simply leaving the jury to make up its own mind about the similarities and dissimilarities, with no assistance at all about their significance, would be to give the jury raw material with no means of evaluating it. It would be as likely to result in over-valuation of the evidence as under-valuation. It would be more, not less, likely to result in an unsafe conclusion than providing the jury with the expert’s opinion, properly debated through cross-examination and, if not shared by another expert, countered by contrary evidence.”
 Atkins was considered in R v Thomas  EWCA Crim 1295. The appellant appealed against convictions for possession of a prohibited firearm and causing grievous bodily harm with intent. The charges related to a shooting incident at a club. Part of the evidence relied on was a bloodstain on a pistol on which all the components of the appellant’s DNA were found to be present. The prosecution’s expert considered the results to provide support for the view that some of the DNA on the gun was from the appellant whereas the defence expert stated that it was not possible to go further than that the results could not exclude the appellant as a contributor to the DNA on the gun. The appellant’s application to exclude the evidence on the basis that the prosecution expert could not give any explanation as to how much support there was for her conclusion was rejected. On appeal the court was troubled by the fact that the expert could not give any indication of the degree of support provided but, following Atkins, concluded that the absence of a statistical evaluation of likelihood was not an automatic bar to the giving of an expert assessment. The credentials of the expert were not in dispute and her opinion was based on her lengthy experience as a forensic scientist. It could not be said to have been so unreliable or lacking in foundation as to make it inadmissible or to compel its exclusion in the interests of fairness. It was of potentially greater assistance to the jury to have that evidence than to be denied it all together. The evidence could be tested in the adversarial process. In that case the expert accepted in cross-examination that she could not say that the appellant had handled the pistol. The court concluded, therefore, that the admission of the evidence did not have any adverse effect on the safety of the convictions.
 The last case on this issue is R v Broughton EWCA Crim 549. The appellant was convicted of conspiracy to commit arson. The case against him depended on Low Template DNA evidence where the quantity was at the very bottom of the scale where a DNA profile could be reliably analysed. The prosecution’s expert concluded that the DNA consisted of a single profile and that three apparently inconsistent results were the product of contamination or stutter. On that basis she concluded that there was a match probability of one in 32,000. The defence expert contended that the findings supported the view that this was a mixed profile. The Court of Appeal held that the judge properly directed the jury that if they did not accept the prosecution expert’s evidence as to the inconsistent results that would destroy the match probability statistics. The judge went on, however, to direct the jury that in those circumstances they could reach their own conclusions on the DNA evidence although he urged them to exercise caution. The court allowed the appeal. The prosecution had not provided any alternative statistics on match probability in the event that the jury did not accept their expert’s evidence that this was not a mixed profile. The jury should have been told to acquit if they did not accept her evidence on the interpretation of the components of the profile as there was no evidential basis upon which they could interpret the DNA findings in any different context.
 From this review of the authorities we consider that the admissibility of expert evidence on DNA findings is subject to the ordinary tests of reliability and relevance (see R v Broughton at paragraph 32). Relevance in such cases is often determined by the other evidence before the court. Where such DNA evidence is admitted it is the duty of the court to ensure that the jury has sufficient guidance to enable it fully and properly to evaluate the evidence. In many cases the expert may be able to provide match probabilities and the task of the court will be to ensure that the jury are alert to the meaning of those statistics. Non-statistical opinion evidence can be admissible whether or not this is referable to any informal scale of probability if relevant and reliable. In appropriate cases it may be necessary to warn the jury not to attempt to carry out any statistical analysis of their own. When considering questions of admissibility it may also be necessary to consider whether the evidence may give rise to unfair prejudice as a result of which it should be excluded pursuant to Article 76 of PACE. In our view it is a good discipline when dealing with the admissibility of such evidence to consider with the parties how the jury should be directed in relation to it if the evidence is admitted. Such a discussion will often provide clarity as to whether the evidence should be admitted and if so how it might be used.
Conclusion on the DNA issues
 The prosecution case was that the applicant had murdered the deceased late on 28 July 2004 by stabbing him on the head resulting in brisk and copious bleeding from which he died. As part of that case the prosecution relied upon DNA recovered from the spot of blood on the applicant’s jacket, the blood stains on the applicant’s trainers and the partial DNA profile on the staining. All of this was intended to support the case that the applicant was in Nixon’s flat at the time that the deceased sustained the injury from which he died and was sufficiently proximate to him after he was attacked to be contaminated with his blood. The evidence of Mrs Cherry lent support to the conclusion that the deceased had contributed to the mixture of DNA found in the bloodied footprint on the door saddle. We can be satisfied that the evidence was reliable as the experts are now agreed that it provides at least limited support for the view that the deceased had contributed to the mixture. It was also relevant to the issue of whether the applicant stepped on the deceased’s blood thereby supporting the prosecution case that she was present after the injury was sustained. We are satisfied, therefore, that the evidence was admissible.
 When he came to charge the jury on this evidence the learned trial judge first set the context. He reminded the jury of the evidence from the forensic scientist who dealt with the footprint on the door saddle and gave his opinion that it provided very strong support for the conclusion that the wearer of the right-hand trainer had stepped on it. He noted this expert’s evidence that there was blood on the sole of the shoe before the wearer stepped on the door saddle and he set out the evidence linking the applicant to the trainers. The only evidence of blood in Nixon’s flat was that of the deceased. None of these conclusions was challenged. The next part of the context was the finding of staining in three locations on the right shoe which was the one that made the footprint. The staining was identified as blood and surrendered a partial DNA profile relating to the deceased. The likelihood of an unrelated male from Northern Ireland having this combination of characteristics was less than one in 50,000. The learned trial judge reminded the jury that the inference which the prosecution asked them to draw from this material was that the trainers had been cleaned to remove traces of the deceased’s blood.
 Having set the context he then turned to the evidence of Mrs Cherry. He noted her expertise. He correctly pointed out that most, though not all, of the DNA characteristics of the deceased were found in the mixture of the sample taken from the saddle board. He correctly advised the jury that the expert conclusion was that this provided some unquantifiable level of support for the view that the deceased had contributed to the mixture. He noted that she accepted that this meant that this was not a conclusive finding. He pointed out that Mrs Cherry agreed that one of the possibilities was that the DNA may have been transferred on the sole of the applicant’s shoe if the applicant had stood in blood that was wet and then deposited it on the saddle board. He correctly reminded the jury, however, that this was only one of the possibilities.
 We consider that it is also important to take into account the questions asked by the learned trial judge at the end of the evidence of Mrs Cherry in order to try to summarise it for the jury. Mrs Cherry agreed that she had found most but not all of the DNA characteristics of the deceased in the sample from the saddle board so that her finding was not therefore a conclusive finding. He developed this further to establish with her that she could not contribute any degree of statistical certainty to whether the deceased’s DNA was present on the saddle board.
 The criticism of the learned trial judge’s charge principally relates to his assertion that the opinion of Mrs Cherry was not a conclusive finding and the risk was, therefore, that the jury may have overestimated the weight that they should give to it. It is important not to isolate a particular passage within a charge but to look at the charge as a whole within its context. The reference to the finding not being conclusive had been debated at the end of Mrs Cherry’s evidence where the discussion had centred on the absence of statistical certainty and the absence of some DNA characteristics. The evidence of Mrs Cherry was relevant in the context of substantial other evidence indicating that the applicant’s right trainer had been contaminated by the deceased’s blood and that the sole of the trainer had carried blood onto the door saddle. In our view, when examined in context, the reference to the absence of a conclusive finding in the learned trial judge’s charge properly reflected the limitations of Mrs Cherry’s evidence and does not give rise to any concern about the safety of this conviction.
The bad character evidence
 There is no dispute about the admissibility of the convictions set out at paragraph 15 above which were introduced to establish propensity on the part of the applicant to commit crimes of violence. There is no criticism of the direction of the learned trial judge on these. It is submitted, however, that the learned trial judge was wrong to allow the evidence of the applicant’s sister that she had seen her carrying different kinds of knives on the basis that evidence that a person carries knives cannot be relevant to a propensity to commit murder.
 Support for that submission is sought in the decision of the Court of Appeal in R v Lafayette EWCA Crim 3238. That was a case in which the victim was the former boyfriend of the appellant’s girlfriend who had turned up at her house. An altercation took place during which the victim was stabbed. Bad character evidence of a domestic dispute was allowed in the course of which it was alleged that the appellant had shouted a threat through a letter box and was found to be in possession of a small knife. The Court of Appeal doubted whether the evidence should be admitted as it did no more than show a propensity to angry outbursts.
 This case is different. The bad character evidence which is not in dispute is evidence of a propensity to use weapons in the commission of violent crime. That is the critical element of the context. Although we do not have the advantage of a transcript of the learned trial judge’s ruling we consider that against that background evidence that the applicant was carrying knives of different kinds was clearly admissible as evidence of propensity. We do not accept that the evidence should have been discounted because the witness was believed to be hostile to her sister or that there was anything inappropriate about the learned trial judge pointing out that it was not challenged in cross-examination. We note, however, that the learned trial judge alerted the jury to the fact that the sisters were not on good terms and advised them to take that into account when considering this evidence. We reject the submission that the admission of this evidence or the manner in which it was dealt with by the learned trial judge rendered the conviction unsafe.
 For the reasons set out above we do not consider that this conviction was unsafe. The appeal is dismissed.