Nasty development in medical privacy

[Peter Fairbrother]

> Brian Morrison at bdm@fenrir.org.uk wrote:

> I would have thought that the more specific (and smaller) the stored
> information becomes, the easier it will be to identify it or connect it
> with another sample.

If medical researchers are looking at individual alleles, which are spread
among the population and not generally identifiable to individuals, no
problem. 

If they are looking at sets of alleles, which they have to do to work on
gene interactions, these sets may be identifiable depending on size etc.
Also they have to know patient history and symptoms etc.

> Currently DNA matching is done by slicing up
> chromosomes and comparing fragment lengths;

Not really - the fragments are sorted by length, but the lengths are the
same from individual to individual - I posted a description of DNA matching
techniques last year if you're interested in the details.

> as soon as it becomes
> possible to be more specific about the sequences produced by this
> technique,

It's possible to be completely specific, just prohibitively expensive.
Sequencing one fragment can be done. From a medical point of view there
isn't a lot of point though, as most DNA is "junk". Medicine looks at the
genes, profiling usually looks at the junk.

> or eventually to cut the strands in known positions,

This is done today as part of DNA profiling, sequencing etc., though
choosing predetermined as opposed to known positions is a bit of a black
art.

> the specificity of data produced will improve.

The data presently easily available from a whole genome (or even one
chromosome) is specific enough already to identify individuals.

I don't know why DNA taken for medical use should be profiled though. As
part of a study into hereditary disease perhaps?

BTW the site mentioned in the original post doesn't have anything to do with
an AIDS case?

Unfortunately, we will have to wait for a better understanding of the genome
before discarding patient-identifying data if meaningful research and
treatment is to continue. As researchers are trying to find out what the
data means, it would be throwing the baby out with the bathwater. They don't
really know what to throw out. Even the "junk" might be important.

Confidentiality can only come from control of access to this data. Or not
including patient's names...

Peter