DNA - Analysis and Probability

This paper was written by Peter Zahra SC
Senior Public Defender

DNA - INTERPRETING ANALYSIS RESULTS
Before considering analysis techniques and interpretation some basic principles need to be considered:

DNA and the Basic Cell Structure
Figure 1. Represents the human cell showing the cell nucleus and Mitochondria. Nuclear DNA is extracted from inside the nucleus. There are two copies of Nuclear DNA in each cell, one that is received from an individual's mother and one that is received from their father- ( that is nuclear DNA is inherited from both parents one half from the mother and one half from the father). Outside the nucleus are small structures known as mitochondria, they have their own DNA which is circular and is inherited only from an individual's mother. DNA extraction is possible from either the mitochondria (mitochondrial DNA or MtDNA) and the nucleus (nuclear DNA). As mitochondrial DNA is inherited only from the mother all maternal relatives have the same DNA. Put simply if you have the same mother all your brothers and sisters will have the same mitochondrial DNA as you. Your mitochondrial DNA is the same as your mother and hers the same as her mother.

View Figure 1 [PDF, 266kb]

To view a PDF document, you will need a copy of Adobe Acrobat on your computer. If you do not have a copy of Adobe
Acrobat, you can download a copy for free from the Adobe site.
If you have difficulty accessing this figure, please contact the Public Defender's Office.

Figure 2: Illustrates mitochondrial DNA links (with circles representing females and squares representing males). The chart shows that children have the same mitochondrial DNA as the mother as also do the aunt and the maternal cousins. The shapes represented in black (either circles or squares) are persons with the same mitochondrial DNA. The male mitochondrial DNA is not carried on to the offspring. Any maternal relative can be used for mitochondrial DNA testing.

View Figure 2 [PDF, 234kb]

To view a PDF document, you will need a copy of Adobe Acrobat on your computer. If you do not have a copy of Adobe
Acrobat, you can download a copy for free from the Adobe site.
If you have difficulty accessing this figure, please contact the Public Defender's Office.

Components of DNA
The foundation of understanding nuclear and mitochondrial DNA analysis is an understanding of the basic components of DNA.
DNA is comprised of four major components which are called bases and represented by the letters A, T, C and G.

A - Adenine
C - Cytosine
T - Thiamine
G - Guanine

The structure of DNA consists of two long strands bound in a 'duplex' with the entire structure coiled in a 'double helix'. The two strands of the double helix are identical in structure but differ in base sequence. The two strands are held together by specific and mutual attraction between the bases. Adenine in one strand only attracts thymine in the opposite strand forming an A-T pair and cytosine only attracts guanine forming a G-C pair. The pairs of bases form bridges between the two strands and are analogous to the rungs of a spiral staircase. The sequence of bases along one of the two strands constitute the genetic code. It is these differing combinations of A,T,C and G which make individual DNA unique.
The analysis of both mitochondrial and nuclear DNA involve an examination of the sequencing of the bases of DNA.

Mitochondrial DNA Analysis: Mitochondrial Hypervariable Gene Region Sequence
Determines whether there is a maternal relationship between the bones/ blood sample and the reference sample.

Mitochondrial Analysis-Poymerase Chain Reaction (PCR)
Mitochondrial DNA (PCR) analysis involves the following process: (illustrated in Fig 3)

View Figure 3 [PDF, 225kb]

To view a PDF document, you will need a copy of Adobe Acrobat on your computer. If you do not have a copy of Adobe
Acrobat, you can download a copy for free from the Adobe site .
If you have difficulty accessing this figure, please contact the Public Defender's Office .

Extraction
There are two methods of extraction, either organic extraction or inorganic extraction. They are basically similar procedures in that enzymes are added to the sample which release DNA into a solution. This DNA sequencing chemistry process is an automated resulting in the DNA in the bone or blood sample being released in liquid form.

Amplification
The amplification process reproduces many copies of certain regions of the DNA which are sought to be analysed. This is known as Polymerase Chain Reaction or PCR. This process provides a means of isolating a segment of DNA by reproducing it in large quantities so that detailed genetic analysis is possible. The reaction involves heating the DNA causing the two strands of the DNA double helix to separate (denature). The enzymes added act as 'primers' for the synthesis of new DNA strands. During a single round of PCR, new strands are formed in the target region with the previously existing strands acting as templates. The result is two double stranded copies of the target region for every copy that existed at the start of the reaction. This process is repeated multiple times resulting in millions of copies of the target DNA sequence. This amplification enables the use of samples that may be degraded and only have a few copies of DNA.

Analysing DNA
At this stage the sequence or the order of the bases A, T, G and C are examined. A simple analogy can be found in an individual phone number. Everyone's phone number can be made up of the numbers 0 through to 9 but it is the order of those numbers that make it an individual number. In the same way every person's DNA is made up of the bases A, T, G and C. But it is the order that those bases appear in that makes your DNA individual.

Repeat Process
Extraction, amplification and analysis of the blood reference samples from the suggested maternal relative using the same procedures above to enable comparison with the target sample.

Comparison of Sequences
The Sequences of both the target sample and blood reference sample are examined to determine if they are consistent with each other. That is to see if the order of the bases ATG and C are the same.

Comparison of sequences using the 'Anderson' Standard
A scientist- Anderson, sequenced human mitochondrial DNA and published that sequence in 1981 (Anderson 1981 Nature 290: 457-465).

The published sequence ( also known as the 'Standard Sequence') has been used as a convention or standard against which other obtained mitochondrial DNA sequences are compared. Consequently in the comparison stage one looks for differences between that standard sequence and the sequences that are obtained from the remains/sample or a reference sample. Therefore (by reference to Figure 4) we can see that at the second position the standard sequence has a G but the remains and reference have a base A instead. Similarly at position 5 and 9 the remains and reference differ from the standard. Further the remains and reference differ from the standard in the same way.

View Figure 4 [PDF, 283kb]

To view a PDF document, you will need a copy of Adobe Acrobat on your computer. If you do not have a copy of Adobe
Acrobat, you can download a copy for free from the Adobe site.
If you have difficulty accessing this figure, please contact the Public Defender's Office.

Examination of the Sequence- Hypervariable Regions
Mitochondrial DNA exists in the form of a loop or is circular. Anderson counted the number of bases (A's, C's, T's and G's) and found approximately 16,500 bases. The bases on the standard have been numbered beginning with zero to 16,500. In comparison with the standard, certain points on the standard loop have been found to be particularly significantly in observing deviation amongst individuals. This is also referred to as the D loop region

Since the 'standard' bases have been number zero to 16,500. These known areas of deviation or 'hypervariable regions' can be examined for differences.
Hypervariable Region 1 and Hypervariable Region 2 are both known to have variations among humans

The Comparison Process-Findings in R v Keir (Supreme Court of New South Wales July 1999)
In R v Keir (Supreme Court of New South Wales July 2000) seven small bones were located in the yard of the accused's home. It was the Crown case that the bones were those of his missing wife. The Crown called evidence from expert forensic Biologists from the Department of Defence in Washington U.S. who carried out both MtDNA and nuclear DNA analysis of the bones. The following are the results of that analysis.

Figure 5 illustrates Hypervariable region 1
Looking at hypervariable region marked 1, the standard sequence is picked up at 16024 (T) to 16364 (C). The Certificate details the comparison of mitochondrial DNA sequence results of the [Anderson] standard and the sequence determined from DNA extracted from bone sample of right phalanx (toe) and left patella (knee cap) and reference sample of blood (from the individual C. Strachan believed to be the mother of the deceased.)
A number of observations can be made from the results:
1. To a large extent the sequence for the two bones and the reference are the same as the Anderson standard. (dashed lines indicate base positions that are identical to the published standard sequence)
2. There are however a number of differences. For example at position 16129 instead of the base G in the standard sequence both bones and the reference have the base A. At position 16172 the standard sequence has a base T whereas the bones and reference have a base C. At position 16294 the standard sequence has base C whereas the bones and reference have the base T. Similarly differences appear at 16304 and 13362.

3. The letter N at 16355 indicates that the base could not be confirmed. This may be due to the sequencing chemistry or because the sample is degraded.

View Figure 5 [PDF, 64kb]
To view a PDF document, you will need a copy of Adobe Acrobat on your computer. If you do not have a copy of Adobe
Acrobat, you candownload a copy for free from the Adobe site.
If you have difficulty accessing this figure, please contact thePublic Defender's Office.

Figure 6 Hypervariable Region 2
Illustrating the Standard Sequence at base 73 to 340.
A number of observations can be made of these results:

View Figure 6 [PDF, 51kb]
To view a PDF document, you will need a copy of Adobe Acrobat on your computer. If you do not have a copy of Adobe
Acrobat, you can download a copy for free from the Adobe site.
If you have difficulty accessing this figure, please contact the Public Defender's Office.

1.Similarly with hypervariable 1 at bases 73; 249; 263 the bone samples and the reference differ from the standard and are the same as each other. At Position 249 the letter D indicates a deletion which is a complete absence of that base in the DNA sequence. Here both of the bone samples as well as the reference blood samples were missing the standard base A at this position.

2. At position 309.1 and 309.2 and 315.1 are 'insertions' or polycytosine. The bone samples and the reference have additional C bases (different to the standard) however this insertion is the same for the bone samples and the reference blood sample.
The analysis results here permitted the expert to give opinion evidence that the results were 'consistent with showing a maternal relationship' between the person C. Strachan and the bone specimens.

NUCLEAR DNA
Short Tandem Repeat analysis
As indicated above the underlying foundation for nuclear DNA analysis is that the DNA type of a child is a composite of half the DNA of the mother and half the DNA of the father.

Short Tandem Repeat analysis is a nuclear base typing system. Throughout individual DNA there are units of DNA bases that are grouped together. These units are repeated a number of different times from individual to individual.

For example Figure 7 illustrates the sequence AATG repeated as two units (Type 2) and as Four units (Type 4). In determining different DNA types, observations are made of the number of times the repeat unit exists. It is the number of times a group of bases are repeated that makes an individual's DNA different.
During analysis different areas of the DNA or markers are examined for length differences using Short Tandem Repeat analysis (STR). Automated DNA sequencing chemistry and gel electrophoresis assay detects fragment length differences (polymorphisms) at various markers. These markers are chosen specifically as they are sensitive to small amounts of DNA or degraded DNA. These markers are also known to have large numbers of length differences in different individuals.

View Figure 7 [PDF, 58kb]
To view a PDF document, you will need a copy of Adobe Acrobat on your computer. If you do not have a copy of Adobe
Acrobat, you can download a copy for free from the Adobe site.
If you have difficulty accessing this figure, please contact the Public Defender's Office.

Examining the Nuclear DNA analysis Results in R v Keir
Figure 8 illustrates STR comparisons between blood from the mother and father with the bone specimen at nine different markers or loci (listed in the first column under the heading 'locus') and a sex marker amelogenen (AMEL).

As can be seen from the chart, the number of repeats in the centre column appears to be a combination of half the repeats from the mother (the left column) and half of the repeats from the father (the right column). For example at D3S1358 the bone specimen produces a profile of 15,18. The bone specimen has taken the 15 repeats from the mother and the 18 repeats from the father and so on. In other words for the bone specimen to be consistent with originating from the parents you would need to account for the 15 and 18 repeats. This is achieved by taking 15 from the mother and 18 from the father.
The sex marker an AMEL indicates the bone sample originated from a female.

The results shown on the chart permit the conclusion that the bone sample is 'consistent with' being the female offspring of the mother and father.

It follows also that if any of the particular repeats in the centre column do not appear to have come from the father at any of the markers then the bone specimen could not come from the offspring of the father.

View Figure 8 [PDF, 44kb]

To view a PDF document, you will need a copy of Adobe Acrobat on your computer. If you do not have a copy of Adobe
Acrobat, you candownload a copy for free from the Adobe site.
If you have difficulty accessing this figure, please contact thePublic Defender's Office.

Probability and Matching
If two samples differ with their genetic types then they are from different individuals. If there is no difference two inferences can be drawn. Firstly, that the samples are from the same individual and secondly that the samples originate from different individuals who happen to match by chance. The more highly discriminating the genetic system used the greater will be the probability that these two individuals are distinguished and in the process decrease the probability of a chance association. If after genetic typing a difference is not found, an estimate can be made of the significance of a chance match. This requires knowledge of the frequency of a particular DNA profile in the population of interest. The New South Wales Forensic Biology Laboratory use a general population database sample from a number of sources with no protocol requiring inclusion or exclusion by ethnic background. American databases are separated into African American; Hispanic American (two types); and Caucasian.

Databases established within the various states of Australia have revealed statistically insignificant differences.

Presently every Australian forensic laboratory has adopted a common system referred to as Profiler Plus which is a system of nine different DNA markers and Amelgenon. A database presently exists for Profiler Plus within New South Wales of approximately 480 individuals. These samples are derived from criminal case work (suspects and victims in criminal cases). A database using Profiler Plus of 348 was used in the trial of Keir .

ADMISSIBILITY OF EVIDENCE OF PROBABILITY- NSW DECISIONS
R v PANTOJA(1996) 88 ACrimR 554
In Pantoja the Crown sought to establish a motive for the accused's wife murder by leading evidence that the appellant had had sexual intercourse with his wife's sister a week before the murder. The accused was a member of a racial sub group of South American Indian known as Quechua Indians. Two DNA profiling tests RFLP and PCR were carried out on vaginal swabs from the wife's sister and semen stains from her nightdress and were compared with a blood sample from the accused. The forensic biologist concluded in evidence that statistically the probability of a chance match was one in 792,000.

The Appellant argued that the evidence of DNA testing and statistical evidence should have been rejected by the trial judge as the databases did not include DNA from the accused racial group.

Hunt C J at CL in considering the general question of the admissibility of DNA evidence said at (at page 558):
'DNA testing has been accepted by the Courts for some years as an acceptable scientific technique for the identification of the source of bodily tissues, in accordance with the approached two scientific evidence generally adopted by this court in Gilmore (1977 2NSWLR 935 at 939 to 941). It is unnecessary to describe the technique for DNA profiling yet again in this judgement. It is adequately described in a number of other cases. See, for example: Elliott (unreported, Hunt J, 6 April 1990) at pp 2-6; Tran (1990) 50 A Crim R 233 at 234; Lucas [1992] 2 VR 109 at 110-112; (1991) 55 A Crim R 361 at 362-364. Green (unreported, Court of Criminal appeal, NSW, File No 60408 of 1991, 26 march 1993) at pp 6-8; Jarrett (1994) 62 SASR 443 at 445-447; 73 A Crim R 160 at 163-165.

The line of American Authority stemming from FRYE v United States upon which Gilmore was based has now been reversed; this has been because of the acceptance by the US Supreme Court that the principle started in FRYE has been superseded by the Federal Rules of evidence. There has not been a reversal of the principle which that case stated. I consider that the Courts of this state should continue to adopt the approach accepted in Gilmore until that decision has been further considered by this Court or the High Court. It was not challenged in the present case'.

His Honour however went on to note (at page 559) 'it is important to emphasise that a match obtained by any blood test - DNA or otherwise - between the suspect and the offender does not establish that the two are the one and the same person. It establishes no more than that the accused could be the offender. However any blood test which positively excludes the suspect as the offender, if there is a reasonable possibility that the test is correct, must necessarily exclude the suspect completely not withstanding that a match has been obtained by other blood tests'.

Statistical Evidence of Probability
His Honour noted the following scientific foundation for the admissibility of evidence of probability (at page 561):
'where a number of blood tests - each of them operating differently from or independently of others - have established a match between the suspect and the offender (in the sense that they have not excluded the suspect as such), the inference is that the greater the number of such tests the less the chances are that the suspect is not the offender. Very little more could be gleaned from such matches alone. The significance of those matches becomes far greater where it is possible to calculate the probability of the match having occurred by chance or coincidence. Databases have been built up which are said to demonstrate the probability that another person would share the particular blood type or would match the DNA of the suspect at a particular marker. The greater the number of such matches, the greater the probability, so that the statistics are calculated by multiplying the probabilities for each match'.

In Pantoja a database of 256 was used in order to establish the probability of a chance match. His Honour concluded that as their was no evidence at the trial that the size of this database had been shown to be statistically valid then the evidence should have been rejected by the trial judge. His Honour concluded (at page 561) 'without any greater knowledge as to the validity of the size of the databases, that evidence would have overawed the jury by the seemingly scientific garb in which it was presented, with the very real risk that they would have thought that it had greater weight than it may have been capable of bearing'.
Ultimately at the re-trial evidence was called by the Crown as to the validity of the database and the evidence was admitted without objection. (Prior to the re-trial the decision in R v Milat was handed down-This decision is discussed below)

Directions to the Jury
His Honour considered that in directing the jury as to the use to be made of DNA evidence that (at page 564):
'The significance of a match between the blood type or DNA of the offender and the suspect (or accused person) must be clearly explained to the jury; that (as I said earlier) it establishes no more than that the accused could be the offender whereas any blood test which positively excludes the accused as the offender - if there is a reasonable possibility that the test is correct - must necessarily exclude the accused completely notwithstanding that there is a match obtained by other blood tests which operate quite differently or independently, and however strong the other evidence in the case may be'.

R v Milat
In R v Milat the Crown sought to lead evidence of DNA profiling in order to establish that the blood on a piece of cord found in the garage of the premises occupied by the accused at the time was consistent with the blood of a murder victim.
Eight DNA tests at different genetic markers were performed on the blood. The evidence established only that the blood on the cord could be that of the deceased (in the sense that no test has excluded the blood as having come from the deceased; that evidence does not establish that it is in fact the deceased's blood).

DNA profiling was carried out in accordance with two different procedures RFLP and PCR. Hunt C. J. at CL accepted the evidence of a forensic biologist that the DNA tests at eight different markers (2RFLP) and (6PCR) were each independent of the others and that the Crown would be entitled to demonstrate what the probability was of a match at all eight markers by chance.

Three databases were used. The database for RFLP results were compiled from testing of 500 people who had donated blood at the Red Cross Blood Bank. (Selected over a short period to avoid resampling the same person).

The database for the PCR results at one marker were compiled from testing of 409 people who were involved in criminal case work (convicted and suspected persons and victims).

The databases for the remaining PCR results were compiled from 402 people; 205 from the Red Cross and 197 in criminal case work. His Honour noted that from enquiries made by [the forensic biologist] from similar laboratories around Australia and overseas and from reports of DNA testing in the United States and the United Kingdom that the size of the database used here 'compared favourably'(see page 440). Further at page 449 'the statistical validity of databases compiled from as low as 100 to 150 people is supported by a number of eminent scientists and scientific bodies'(These studies were considered at pages 449 to 451 and listed in the Appendix to the Judgement).

His Honour further noted at page 450 (in relation to databases used in Milat) that: 'the statistical validity of any database also depends upon whether it can be said to be representative of the general population, and the degree by which any database can be said to be so representative depends in turn upon whether the persons tested have been selected at random - whether each person in the population has had an equal chance of being selected'.
His Honour went on to say: (at page 450)

His Honour noted that comparisons with the U.S. Caucasian Data base and data bases of Aborigines and Asians revealed variations in DNA patterns. His Honour noted that 'it is accepted that there is a variation in DNA patterns between [Aborigines, Asian] and Caucasian. No figures have been made available, but the population of [NSW] is accepted as being predominantly Caucasian

His Honour found that the validity of the databases used in the Milat case had been established. (No issue on the admissibility of the evidence of probability was argued on appeal to the Court of Criminal Appeal)

Proviso / Directing the Jury
His Honour went on to qualify this conclusion by heading the proviso noting (at page 451): '... unless the whole population were to be tested, no estimate could ever be 100% correct. The confidence limits for any estimate depend upon the size of the database. ...'In Milat the forensic biologist indicated a confidence level of 95% in relation to his assessment of the probability of a chance match. The probability was obtained by multiplying the probabilities in relation to each of the eight independent matches obtained.
His Honour however went on to make the following comments in relation to directing the jury(at page 452):

R v. LISOFF [1999] NSW CCA 364 23/11/99]
The Discretion to exclude Evidence.
This was a Crown application under S5F of the Criminal Appeal Act1912 where it was argued that the trial judge erred in his exercise of discretion to exclude scientific evidence relating to the DNA identification of the complainant blood on the accused's clothing.

The Respondent was charged with one count under section 33 of the Crimes Act(NSW) of maliciously inflicting grievous bodily harm with intent to intent to inflict grievous bodily harm. At his arrest police took possession of his track suit pants that he said he was wearing on the day of the assault. They were then forwarded to the Crime Scene Examination Unit where an officer identified spots of blood on the tracksuit. They were then forwarded to the government laboratory for DNA analysis. That DNA analysis established that blood on the track suit had the same genetic profile as the victim's blood.
At trial during a voir dire, evidence was called by the defence from a DNA expert to the effect that blood on the track suit was post transfusion blood (the victim had a transfusion at a hospital subsequent to the assault) therefore implying that the blood had been 'planted' on the tracksuit. There was conflicting evidence called from both Crown and Defence expert witnesses concerning the sensitivity of DNA testing to the detection of foreign DNA matter said to be consistent with blood from a transfusion.

Counsel for the Respondent at trial argued that the DNA evidence should be excluded by the Trial Judge in the exercise of discretion under Section 135 and 137 of the Evidence Act(NSW) (1995).
In the trial Judge's judgement His Honour noted: (at page 5)
'I take that your starting point that disputes about facts are ordinarily for the jury to decide, and if those facts depend on expert evidence, it is still the case that the Jury should decide that, and there might have to be specific directions which make it clear what the dispute about the expert evidence is'

His Honour then went on to refer to remarks made by Hampel J in R v Lucas (1992)2 VR 109 to the effect:
'DNA testing is widely regarded as extremely reliable and discriminating. Its limitations and particularly limits as to the conclusion which can be drawn from the tests are not generally appreciated. The Jury has no basis upon which it can evaluate the evidence. There is no way the jury can properly weigh the value of such evidence if there is no evidence before it as to the frequency of a match in the general population'.

The trial Judge in excluding the evidence went on to say
'Now here of course [evidence of probability] is not an issue because it does not apply. But when I apply the principles behind that reasoning to this case, I find that there are two questions, namely, that the effect of the DNA profile evidence and that of the evidence relating to the custody of the clothing and the blood and if they are to have probative value cannot be separated. Taken together, if accepted, they would assist in proving the presence of the accused at the crime scene. That is clearly probative value. While I have said taken separately the issue relating to the expert evidence and that relating to the security of the exhibits might be left to a jury, in my view there is a real danger that because of the combination of the circumstances, namely, the complex nature of the scientific evidence, the complex relationship of that evidence to the evidence relating to how the blood might have come to be on the track suit pants and the blood spatter evidence, a jury, even if properly directed, could fail to appreciate either the complexity of the inter-relationship of the various pieces of the forensic evidence or the limits that might be placed on the forensic evidence. There is a real danger that the fact finders might be unduly swayed by the 'scientific' nature of the evidence to make a decision on an improper basis, particularly to require a lower degree of probability than they would otherwise require. I therefore rule that under Section 137 of the Evidence Act the evidence must not be admitted'.

The Court of Criminal Appeal considered at length the circumstances in which the discretion to exclude evidence might be exercised under Sections 135 and 137 of the Evidence Act (see pages 9-14 of the Judgement).

The Court ultimately concluded that His Honour applied the wrong test in the exercise of his discretion arising under Section 137. (see pages 13 to 15). It found that predominantly the questions for determination were questions of fact falling within the capacity of the jury to decide.

The Court went on to say (at page 14) 'there is nothing so extraordinary about the conflict and the evidence presented in this case which would justify the conclusion with a careful and sensible jury, properly directed as to the relevant evidence, could not decide in a reasoned and responsible way whether or not the Crown had demonstrated beyond reasonable doubt that the body of evidence supporting the Crown case should be preferred to the opposed body of evidence.
In relation to the scientific DNA evidence the Court held (at page 14):

'The test that a jury 'might be unduly swayed by the 'scientific' nature of the evidence to make a decision on an improper basis' was derived from judgments quoted by his Honour earlier in his decision. Those references were concerned with the undue weight that may be given to scientific evidence by reason of its very concreteness. They were not concerned with a situation such as that presented to his Honour where there was a real conflict of scientific evidence. That is a quintessential jury question. It is not, in our opinion, appropriate to infer that a jury even 'might' be affected in a prejudicial way by reason of the 'scientific' nature of the evidence, where such evidence is contradicted by other 'scientific' evidence.'

The Court concluded that the statutory discretions exercised by the trial judge miscarried.