Identifying a Culprit In a Bloodbath

worromot writes "A group of geneticists published a method to determine if a given individual's DNA is present in a mixture (e.g., in a pool of blood on a carpet). An individual's DNA can comprise less than 1% of the mixture. (The article is in open access on PLoS Genetics website.) While this is a potential boon for forensics, there are more immediate worries about the privacy of the participants of the genetics studies that had been under way for many years. As Science magazine writes, 'The discovery that a type of genetic data that is widely shared and often posted online can be traced back to individuals has prompted the US National Institutes of Health and the Wellcome Trust to strip some genetic data from their publicly accessible Web sites and NIH to recommend that other institutions do the same.' The gravest worry was that an individual who had someone's genetic code could determine, based on the pooled data, whether the person participated in a disease study and whether they were in the disease group, or thereby glean private health information. NIH plans to ask institutions that have posted pooled data on their own Web sites to take these down, too."

Comments

UK government pre-empts this problem

[Anonymous Coward]

Thankfully the British government and the NHS have been leaking private medical records en masse for years, cleverly sidestepping this issue completely.

God Save The Queen.

The Internet is like a permanent stain

[Crazy Taco]

Good luck with taking that stuff down. Posting something on the Internet is like spilling grape juice on a white cloth. If it wasn't made obvious by the age controversy over China's gymnasts, then I'll say it again: once something is on the Internet it stays there, no matter how much scrubbing you do. People need to think first and to not put something up if there is ever a chance it will be an issue.

In the News

[Anonymous Coward]

Here are some more links to news and discussion (thanks google):

Protecting Aggregate Genomic Data Elias A. Zerhouni and Elizabeth G. Nabel (4 September 2008) Science [DOI: 10.1126/science.1165490] http://www.sciencemag.org/cgi/content/abstract/1165490

Science 5 September 2008, Vol.321 no. 5894 p. 1278 http://www.sciencemag.org/cgi/content/full/321/5894/1278

Science Now http://sciencenow.sciencemag.org/cgi/content/full/2008/829/1

Nature News http://www.nature.com/news/2008/080904/full/news.2008.1083.html

Nature News http://www.nature.com/news/2008/080903/full/455013a.html

New Scientist http://www.newscientist.com/channel/health/dn14637-genetic-data-withdrawn-amid-privacy-concerns.html?feedId=online-news_rss20

Financial Times (UK) http://www.ft.com/cms/s/0/92b8eed8-7561-11dd-ab30-0000779fd18c.html

LA Times http://www.latimes.com/news/science/la-me-dna29-2008aug29,0,1478453.story

AZ Republic http://www.azcentral.com/arizonarepublic/news/articles/2008/08/29/20080829biz-dnaswab0829.html

El Mundo http://www.elmundo.es/elmundosalud/2008/08/28/biociencia/1219947993.html

Times (UK) http://www.timesonline.co.uk/tol/news/uk/science/article4649977.ece

Genome Web #1 http://www.genomeweb.com/issues/news/149084-1.html

Genome Web #2 http://www.genomeweb.com/issues/news/149097-1.html

Slashdot http://yro.slashdot.org/yro/08/09/06/1943215.shtml

Chemie.de http://www.chemie.de/news/e/86369/

Medical News Today http://www.medicalnewstoday.com/articles/119794.php

Bioinform http://www.bioinform.com/issues/12_35/features/149237-1.html

There is a invited discussion forum in Plos Genetics:

http://www.plosgenetics.org/annotation/getCommentary.action?target=info%3Adoi%2F10.1371%2Fjournal.pgen.1000167

This will be horrible for false-positives

[Anonymous Coward]

Imagine the number of people who may be implicated merely because they bathed in the blood without actually participating in any murders.

CSI trend

[marco.antonio.costa]

That worries me a bit. Seems as law enforcement is nowadays putting all their chips in forensics miracle technologies and stepping back from doing their ol' homework.

I vaguely remember a story of a case that a guy was wrongly convicted because of a cat DNA sample at the place matched a piece of fur in his jacket, but was a false match, cause cats DNA can be almost identical from time to time. Then that would be possible with humans, a la birthday paradox.

One would imagine a bloodbath would leave other evidence, say, witnesses shocked by the gunfire and screaming. Or chainsaw noises. :P

Re:CSI trend

[Anonymous Coward]

Useful forensic evidence is only available in a small fraction of criminal cases, and genetic forensic evidence is even rarer.

However, the probability of these things being useful goes up with the seriousness of the crime. If your car stereo gets stolen, the cops might not bother dusting for prints because it's just not a priority. If you have a serious crime scene, it makes sense to get as much genetic material as you can to help look at.

The new technique of getting prints off of the micro-corrosions that skin oil leaves on bullet cartridge casings is going to help a lot more here in the U.S. where we have so much gun violence. Shooters aren't going to end up with blood in the pool as often as knife- or blunt-instrument fighters.

law enforcement is nowadays putting all their chips in forensics miracle technologies and stepping back from doing their ol' homework.

Homework is the only thing they got in most cases, so don't worry about them losing their elite interrogation tactics. You should be more worried about them going the way of the FBI's rapport-based interrogations than the CIA's coercive tactics which were originally designed to elicit false confessions for propaganda purposes, I shit you not.

One would imagine a bloodbath would leave other evidence, say, witnesses shocked by the gunfire and screaming. Or chainsaw noises.

See, the problem there is that a shocked witness is usually not the witness with the best memory, and chainsaw noises come a lot more often from law abiding citizens trying to keep their storm drains and electrical lines clear. Every little bit helps, and while a potential employer could conceivably scan scientific data to see if a potential employee has a disease, where are they going to get the DNA to match it against? It seems far-fetched and easy to deal with using existing laws saying what you can and can not ask prospective employees.

Race

[MillionthMonkey]

A lot of SNPs and coding regions can be used to identify haplotypes- e.g. we might know that the probability of finding an A rather than a T at a particular base position on chromosome 3 is 90% for Asians and 20% for everyone else, or 40% of people with Huntingdon's and 90% of people without, etc. If you can gather SNP information from locations that are spread out across linkage points on different chromosomes, you can pretty much pin down the phenotype of the guy if any data has ever been gathered specifically mapping the phenotype distribution to the base pair probability. And if you're being genotyped, they'll know your race along with a lot of other phenotypic information about you from the paperwork they'll have you fill out.

This is a weird situation, because race is only one of many attributes you have that you have no control over, but we obviously single it out and make it a sore spot. Now that they can genotype bloodbaths, will we get lynchings of color blind guys to come from this? Probably not, but I can easily imagine something like this igniting racial tensions.

fear mongering ftw

[thermian]

OMG DNA!!!!!11111one

You know what, I can pretty easily say that without a lot of expense, there's not really any real danger of your DNA's 'privacy' (whatever the fuck that is) being violated. Do you have any idea how much DNA analysis costs?

And if it is, if someone gets hold of your DNA? Well, DNA analysis is a resource hungry affair. Without prior knowledge of a reason to try, I can't see that any analysis would be done. It takes experienced people, and there is more than enough work examining DNA from crime scenes to keep them busy, without data mining random DNA as well.

I spent two years working on DNA analysis techniques, particularly with regard to the application of data mining (not for the kind of thing that would be a privacy issue). We, by which I mean the DNA analysis crowd, are a long way from anything which could be applied on a large enough scale to pose a genuine threat to someones 'DNA Privacy'.

Honestly, there are big enough problems to solve without wasting time on sensationalist bullshit like this.

Re:fear mongering ftw

[rwillard]

OMG DNA!!!!!11111one

You know what, I can pretty easily say that without a lot of expense, there's not really any real danger of your DNA's 'privacy' (whatever the fuck that is) being violated. Do you have any idea how much DNA analysis costs?

And if it is, if someone gets hold of your DNA? Well, DNA analysis is a resource hungry affair. Without prior knowledge of a reason to try, I can't see that any analysis would be done.

You're right, of course. Information posted on the internet is never archived, and the barrier to doing data analysis on collected information never lowers over time.

Re:fear mongering ftw

[MillionthMonkey]

It looks like you're essentially applying an evolutionary algorithm to mRNA expression data, which requires a run of many chips over a time series, over experimental parameters, and over sample replicates. These are genotype arrays. They cost less than a thousand bucks each and you only run one per individual (maybe several replicates). Plus the measurements are easier because it's a digital signal, so the scanners don't have to be terribly sophisticated. Already people are making ones as small as shoe boxes.

The "data mining" that would be involved would be the extraction of population data, but this has been done. And that was expensive, requiring lots of chips, but it only needs to be done once per population to derive a large amount of phenotypic information from a given genotype. And a lot of this data has already been published. For thousands of SNPs, we know which ones are associated with which phenotypes now, and with what probabilities. And Illumina just started selling a chip with 1.2 million SNP probes on it, if you're asking for a large enough scale.

How would this be any different from just running a SNP chip?

Re:fear mongering ftw

[thermian]

Actually I wasn't referring to the expression time series work, that was interesting, but all it really demonstrated was that we aren't at the point of being able to do it in a useful fashion.

I was referring to the DNA pattern extraction, that requires a lot of work. I only applied it to promoters, but it has wider uses that are still being explored (including website similarity, oddly enough..). Pattern matching is required for data mining DNA, and we have only just started to get a grip on the very basics of the task.

On the whole my comment was based on the many hundreds of papers I've read on the subject, not my own work.

Re:fear mongering ftw

[worromot]

Do you have any idea how much DNA analysis costs?
About $350, the last time I checked. For about 500,000 genotypes per individual.

There is also a very major technology push for the "thousand-dollar genome", i.e. an ability to get a complete genome for $1000.

The core of the finding, by the way, is that the pooled data that everyone thought was completely safe for privacy point of view, is now no longer so. It is a problem for people who have agreed to take part in these studies (and that's a lot of people: a typical large scale study can involve 20,000 patients or more).

spent two years working on DNA analysis techniques... We, by which I mean the DNA analysis crowd
For what's it's worth, I've spent the last 15 years working on various DNA analysis techniques. I don't know about your "crowd", but the general genetics community takes issues of privacy quite seriously.

More worrying is the potential use in law

[AmiMoJo]

The "low copy count" method of multiplying a tiny sample of DNA to produce one large enough to make an identification has already been discredited in the UK (although the police continue to use it), and I can see this going the same way. A drop of blood will be found, the police will find some tiny sample of some poor guy who happened walk past that spot at some point and they will then have to fight it in court.

It's a shame the police can't be trusted to look at this and regulate their own use of it, but past experience with other DNA and fingerprint techniques has shown that they can't.