Robin Wilton's esoterica

... two steps back.

Over the holiday period, the government's ID Card plans were revised so that instead of a single centralised database, it would be based on three existing databases, each of which would hold specific parts of the citizen's over-all identity profile. In design and implementation terms this appeared to be a rational move. 

A couple of weeks later, though, came news of a "Whitehall super-database", to be founded on new data-sharing principles which would over-ride the existing concepts of departmental charters. We should not forget that the idea of a departmental charter - which
often forbids a department from using the data it holds for purposes
other than the reason it was collected - closely mirrors the principles
of data protection law.

The extent of that data-sharing will apparently be based on input from five one-hundred-person  'citizens' panels', who will act as a focus group to say what degree of privacy 'people' want. It's almost as if, having run into objections in every other forum in which this proposal has been tested, the policy-makers are casting arond for any source which will give them the answer they want. This 'ask the people' approach may sound democratic superficially, but is it? After all, no-one elected these 'citizens' panels', and they certainly will bear no accountability for the course of action they recommend.

It seems to be quite hard to find out what, if anything, is happening to put the 'citizens' panels' into operation. Unless, of course, you know different... 

Pretty much everyone seems to have heard of DNA testing these days, although I suspect few of us have anything but the most rudimentary idea of how it works, or what the statistical matching probabilities mean. DNA testing is generally portrayed as being the 'be-all and end-all' of biometrics... though the time it takes to process a sample currently makes it unsuitable for, say, identifying you at an airport check-in desk.

However, it's quite possible that unless you're a biologist or some kind of forensic specialist, you won't have heard of 'Low Copy Number' testing. Simply put, this is a techique for deriving a result from extremely small quantities of DNA sample. To give you a very rough idea of numbers, here's one slide from a presentation given by someone from NIST at a DNA workshop last May. This slide is pretty clear and simple... but I should warn you, the rest of the presentation pretty quickly turns to phrases like "Stochastic PCR amplification", "heterozygote peak height ratios" and "allele dropout rates"... so approach with caution.

If a chromosome is a string of DNA material, an 'allele' is a point on the piece of string where there's a particular sequence you want to compare with the corresponding point on another string of DNA. So the table above describes the baseline, which is that you get one of each 'allele' (or 'matching-point') per cell in your sample.

The table shows [what mass of DNA you can expect to recover from a given number of cells]; that is, about a nano-gram per 152 cells. If all the DNA in your sample definitely comes from the same person, then even if you only have 0.125 nanograms of DNA, you could expect to recover about 19 cells and therefore 19 copies of each allele you're looking for. The coloured cells in the diagram represent 'Low Copy Numbers'... so here, 19 alleles or fewer.

The presentation then goes on to show how those baseline assumptions need to be adjusted if, for whatever reason, the sample you are analysing contains DNA from two different sources, and if the material from each source is present in differing proportions. So for instance, if you have 0.125 nanograms of sample but it's from two different people in equal proportions, obviously you are only likely to get half the number of alleles from each person on which to base your conclusion. Thus, if 19 alleles is your LCN 'threshold', you would need 0.250 nanograms of 1:1 sample if you are to get a result as reliable as the one shown in the table above.

The current (paper) issue of Private Eye also has an article about LCN DNA testing, pointing out the same concern as the NIST presentation: that the reliability ofthe results can depend very considerably on the way in which you apply the technique. As a result, the FBI and others have, since 2001, 'urged caution when using LCN' as a forensic technique.

In that context, this article from the BBC site does not appear to give a very clear picture. It says that a number of court cases are to be reviewed by the police, because the UK Forensic Science Service (FSS) for a period applied the technique in a way which may have missed valid DNA evidence. It gives the impression that LCN techniques 'have only been available since 2001', and does not make clear how the UK Forensic Science Service's application of the technique differed from anything considered more reliable. According to the NIST presentation, though, LCN has been around since 1996, and the FSS started using it in 1999. The cases underreview date from the period between 2000 and 2005.

What does seem clear, though, is this. For all the public's general acceptance of DNA testing as a cast-iron technique, it is a complicated area and one in which an apparently binary answer may conceal a wealth of interpretative and methodological variance.

The worrying part, from the point of view of individual rights, is that as advances in test technology make it possible to use ever-smaller DNA samples, it becomes proportionally easier for 'enough DNA' to be transferred from somewhere you have been to somewhere you have not. (For instance, through contact with another person, who picks up some of your DNA-bearing material and sheds it again some time later). At that point, the question changes from "Is this your DNA?" to "Can we prove how your DNA came to be here?", which could be far harder to establish.

Out of solidarity, I have decided to give up something for Lent. However, as I find it hard to grasp the deep spiritual significance of giving up chocolate, or coffee, or whatever, for 40 days, I have had to come up with a suitable alternative. After minutes and minutes of searching whatever it is that agnostics search instead of a soul, I have come up with this: 

I am going to abstain from 'days when I only blog critical, sarcastic and/or generally negative posts'. In other words, for every caustic rant I post, I will also find something positive to blog about. I just thought I had better warn you all, lest the sudden, unexplained appearance of feel-good posts makes you fear for my mental wellbeing.

Here goes, then: it is actually nice outside today. The sun is shining, the temperature is mild, the frogs are procreating with style, enthusiasm and their customary epic stamina, and the buds are starting to appear on the amelanchier, 'snowy mespilus' or shad-bush. All in all, spring is in the air.