On 25th March 2015, 530 years after his death, King Richard III of England will be interred in Leicester Cathedral. This remarkable ceremony is only taking place because of the success of DNA analysis in identifying his skeletal remains. So what sort of genes might a king be expected to have? Or, more prosaically, how do you identify a long dead corpse from its DNA? Several methods were used, and in particular the deduction of the skeleton’s probable hair and eye colour raises some interesting questions about future trends in forensic DNA analysis.
Richard III is one of England’s best known kings, largely due to the famous play of William Shakespeare in which he is portrayed as an evil villain. He only reigned for two years and was killed at the age of 32 at the battle of Bosworth in 1485. According to the historical records he was unceremoniously buried at Greyfriars Friary in Leicester. At some stage knowledge of the exact location of Richard’s burial was lost. But in 2012 excavations under a car park at the probable site of the former friary yielded “skeleton 1”. Suspicion of his royal identity was excited by the fact that the skeleton had a severely bent spine causing the right shoulder to be higher than the left. This well-known deformity of Richard was mentioned in a contemporary source, as well as by Shakespeare. Furthermore, the skeleton was male, the age was about right, it had evidently been killed in battle, and the radiocarbon date was consistent with death in 1485.
This was all very suggestive, but it was the DNA analysis that really proved the case. The work was led by a team at the University of Leicester, with participation by many other UK and European centres. It is important to note that this was not the normal type of forensic DNA identification, which relies on comparing a set of highly variable DNA markers to a database. Such analysis is fine so long as your suspect is in the database, but it is no use for identifying a long dead individual who is not in any database.
By far the best evidence for the identity of Richard III comes from the analysis of his mitochondrial DNA. Mitochondria are bodies found in every cell, responsible for the production of energy. They have their own DNA which is passed down the generations only through the female line. Barring the occasional new mutation, the DNA sequence of mitochondrial DNA should be identical from mother to daughter down a particular female line of descent. Like their sisters, males also carry the mitochondrial DNA of their mothers, but they do not pass it down to their own offspring.
Richard will have shared mitochondrial DNA with his sister, Anne of York. Two complete female lines of descent were traced back to Anne of York, one of 17 generations down to Michael Ibsen, a resident of London, and the other of 19 generations down to Wendy Duldig, formerly of New Zealand. Complete sequencing of their mitochondrial DNA showed a 100% match between skeleton 1 and of Michael Ibsen, and a single base change compared to Wendy Duldig. One change over this period of time is quite likely to be a new mutation. The sequence family (haplogroup) to which the mitochondrial DNA sequence belongs is a fairly rare one, so few other people in England in 1485 would have shared it and in fact the team has systematically ruled out all the other males of the period who might have shared it because of a common female lineage with Richard III. So this match is highly significant and is the best piece of evidence that the “skeleton 1” is indeed King Richard.
Also applied was a newer method which is a technique for predicting the hair and eye colour of someone from their DNA. The most important variants affecting hair colour are mutations of a gene called MC1R, which encodes a cell surface receptor for a hormone. Individuals carrying variants of the MC1R gene with reduced function are likely to have red or blond hair rather than the normal dark hair. The pigmentation of the iris of the eye depends significantly on a gene called OCA2, encoding a protein which transports tyrosine into cells. Again variants of reduced function give less pigmented eyes, meaning that the colour is blueish rather than brownish. Recently a Dutch group created a forensic test based on variants at 24 genetic loci, of which 11 are in the MC2R gene and the rest in 12 other positions including the OCA2 gene. Identification of these 24 variants yields a fairly accurate prediction of hair and eye colour, and in the case of skeleton 1 the prediction was for blue eyes and blond hair. The existing portraits of Richard III all date from some time after his death but the older ones do indeed show light-coloured eyes and reddish-brown hair, an appearance which is consistent with the prediction.
These two types of analysis indicate two rather different senses in which we use the word “gene”. The sequence variants of the mitochondrial DNA, like those used in normal forensic identification, do not, in general, affect the characteristics of the individuals carrying them. The DNA changes often lie outside actual genes, in the regions of DNA between genes. They are better described as “markers” than as “genes”. But the hair and eye colour analysis is based at least partly on actual gene variants that might be expected to generate those visible characteristics.
How much further might this kind of analysis be pushed? Could the height, facial features or skin colour of a crime suspect be deduced from their DNA? The essential issue is the number of gene variants in the population that affect a feature. If it is relatively small, as with hair and eye colour, then prediction is possible. If it is very large, as for height, then it is not possible, because most of the variants affecting height have too small effects to be detectable. Most of the human characteristics that have been studied in this way have turned out to depend on a very large number of variants of small effect. So, contrary to popular perception, there are real limits to what is possible in terms of prediction of bodily features from DNA data. There will doubtless be some other features that are predictable, and these may eventually include skin colour. But unless a completely new approach is invented, it is unlikely that we shall ever see an identikit picture of a suspect generated from DNA at the crime scene.
Featured image credit: Stained glass, by VeteranMP. CC-BY-SA 3.0 via Wikimedia Commons