Rachael R. E. Frazier¹, Andy Urquhart¹, Rebecca Sparkes¹, Ian Findlay², Alison Taylor², Stephanie K. Watson¹, Peter Gill¹.
¹Service Development, The Forensic Science Service, Priory House, Gooch Street North, Birmingham, B5 6QQ, UK.
²Department of Molecular Oncology, Algernon Firth Building, University of Leeds, Leeds, LS2 9JT, UK.

× Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø × Ø

The UK National Criminal Intelligence DNA Database uses a quick and efficient single tube STR multiplex developed by the Forensic Science Service; the Second Generation Multiplex (SGM), to amplify seven loci simultaneously. Amelogenin, D21S11, D18S51, HUMTH01, HUMVWFA31/A, HUMFIBRA (FGA) and D8S1179 are combined to give a discriminating power in the region of 1 in 50 million (1,2).

Since its inception, the Database has provided the FSS with enormous expertise in analysing a range of profiles. The identification of alleles and their designation will be discussed. The Database is used as an effective tool to screen DNA profiles from suspects and therefore it is necessary to obtain an STR profile from trace evidence left at the scene of a crime. In order to achieve this we are investigating the possibility of obtaining an SGM profile from a single cell. This technique could in future be used to analyse undetected rape cases, where the only evidence remaining may be a few sperm heads on a slide.

As use of the Database becomes standard practice for police forces throughout the UK there is a requirement to provide a further multiplex system. This will be achieved using a system currently under validation at the FSS; the Third Generation Multiplex (TGM).This provides a discriminating power of approximately 1 in 25 million, leading to a combined discriminating power of approximately 1.25 x 10¹⁵ for both SGM and TGM. The FSS has found this two-step method of analysis to be the most cost-effective overall.

Previously, DNA technology was only used on serious crime as it was labour intensive and very expensive. However, with the advent of STR multiplex technology, DNA analysis has become an extremely cost effective means of investigating crime. Seven loci are amplified simultaneously which substantially reduces turnaround time and many of the processes can be automated, thus reducing labour costs. This technology has made it possible to set up a UK National Criminal Intelligence DNA Database. Buccal swabs are taken from suspects and amplified with the SGM. These profiles are entered onto the Database, together with profiles from crime stains and the two are cross-referenced to generate ‘hits’. This is a significant breakthrough as offenders are caught whilst actively committing crime. If a suspect is acquitted the profile is removed from the Database.

The UK National DNA Database is now in its third year of operation and has processed over 200,000 samples. Of these samples, 185,328 are on the Database, the remaining samples being removed due to acquittal. A total of 8,055 matches with crime stains have been detected and 2,087 crime scenes have been linked. Mass screens, where individuals provide samples voluntarily, have resulted in 15 ‘hits’. (August 1997 figures).

Null alleles result from point mutations within a primer binding site or a deletion of all or part of the locus. Partial nulls can also be observed where the allele is present, however there is a reduction in the peak area.

It is thought that true null alleles are rare, indeed none were seen in a study of 600 meioses, and we have not seen any double nulls. However this does not mean that double nulls do not exist and they may have simply been attributed to locus failure. A null allele has been observed at the amelogenin locus, where they are easier to ascertain and a primer binding site mutation has been confirmed as the source by amplifying the sample again with amelogenin primers outside the original primers.

More than two allelic peaks are usually observed in mixed samples; however, we have demonstrated that genetic anomalies can lead to multiple bands. Triple peaks can arise from trisomy, gene duplication and translocation. Frequencies of occurrence are detailed in Table 1. Nine samples have exhibited three peaks at D8, three of these being brothers. We suspect this must be due to a duplication or translocation as D8 trisomy is not viable. Four trisomies have been observed at the D21 locus which may be indicative of Downs Syndrome; however, only two of these individuals are known to exhibit the condition. One hundred and seventy samples have exhibited XYY profiles, so-called ‘supermales’ which is the most common form of triple signal. XXY (Klinefelters Syndrome) has been seen in 70 samples.

The frequency of somatic mutation seen per locus is detailed in Table 2. This is probably a gross under estimation as they may go undetected as stutters or be swamped by other alleles. The result is three uneven peaks at a locus where the signals are additive: the peak areas of the two smallest peaks will tend to add up to the peak area of the largest peak. The occurrence of somatic mutation is important to note in a forensic context as variation will be observed in two ways; i) tissue to tissue; a buccal scrape may not match a blood sample and ii) time; a 1996 buccal scrape may not match a 1997 buccal scrape.

The Database is used to link crime scenes with suspects and therefore it may be necessary to obtain an STR profile from trace amounts of evidence. In order to achieve this we are currently investigating methods of obtaining an SGM profile from a single cell. We are collaborating with the University of Leeds, where STR markers are used to provide a means of detecting contamination in pre-implanation diagnosis. Results have previously been obtained using a developmental STR multiplex from the Forensic Science Service [3]. Buccal cells are currently analysed as they are easy to obtain and relatively easy to recover. Cells are recovered using micromanipulation techniques and amplified with the SGM using slightly adjusted conditions; modified primer concentrations, the use of AmpliTaq^® Gold and an increase in cycle number to 34.

Preliminary results have been obtained from 195 buccal cells taken from 4 different individuals (Table 3). Results were obtained from 92% of cells and a full SGM profile from 57% (Figure 1). An acceptable profile was obtained from 71% of cells; this being amelogenin plus 4 or more STRs, a result that can be entered onto the National DNA Database. Allelic drop out was seen in 35% of cells and was observed to occur at a rate of approximately 10% for each locus. Additional alleles were observed either in place of or in conjunction with the true alleles. These can be attributed to i) somatic mutation, ii) PCR-generated non-allelic bands or iii) contamination.

The results obtained are encouraging and show the potential for the system, however there are several points that need to be taken into consideration when analysing very low amounts of DNA; single cells in the extreme. Contamination is an important issue and strict controls need to be employed to minimise any effects. Currently the cells are recovered in Leeds and sent down to Birmingham to be analysed. Decontamination regimes are in force in all laboratories, however the results have shown that the potential for contamination still exists and the situation must be monitored effectively. It is important to note that with low copy number DNA the technique is extremely sensitive and therefore transfer by innocent means is more likely to occur. Allelic drop out, preferential amplification, somatic mutations and the presence of additional alleles may lead to potential typing errors (Figure 2). It is therefore vital to use a multiple tubes approach and analyse a sample repeatedly to obtain a consensus profile. Stringent rules of interpretation need to be applied and conservative statistical criteria used when interpreting the results. This system has the potential to be used on a great many evidence types where an STR result was not previously obtainable, an example being slides from undetected rape cases where only a few sperm are remaining, and other low copy number samples. Although the results are promising the technique is still only at a developmental stage and a great deal more work has to be done before it can be used for forensic casework. The techniques must be shown to be robust and reproducible and fully validated before use.

As more samples are loaded onto the National DNA Database there is a need to be able to distinguish between random matches that may happen purely by chance. After extensive cost benefit analyses it was determined that the most cost effective method of achieving this was to develop a third multiplex system. This system is comprised of D3S1358, GGAA3A09, D1S518, D10S516, D14S306, GATA4F03 and HUMTH01 (known by their Genbank designations, Figure 3), the latter being a QA check between SGM and TGM. TGM has a discriminating power of approximately 1 in 25 million and combined with the SGM it provides an increased discriminating power of 1 in 10¹⁵.^. It is envisaged that it will primarily be used as a ‘hit’ confirmation tool for the Database. Once a ‘hit’ is generated, the sample will be amplified with TGM to confirm a true match or an adventitious match. It will also be used in casework and paternity although these exact roles are undecided as yet.

REFERENCES

Frequencies are quoted as the number of repeat units a sample is found to vary from the original.

¹Additional allele present in conjunction with the true allele.

2Additional allele present in place of the true allele.