Your DNA on a DATABASE?

[esme]

So if everyones DNA was on it, it would work.

no it wouldn't, the DNA profiles on the database identify groups not individuals, so the more people you have on the database the larger the group it will return for a given sample

 

say you have a sample with a 1 in 10,000 match probability the DNA database hands you 6,500 suspects, and as I showed above most of these will live within striking distance of your crime

 

the more people on the database, the more false positives need eliminating from the enquiry

I thought DNA could be got from things such as strands of hair?

only living cells contain DNA, so flakes of skin or hair roots would work, the main body of hair is completely dead and contains no genetic information

But knowing who was at the scene would be a good start.

I refer you to the 6,500 figure I quoted before

How many burglers, murderers and other criminals would have this unique skill?

very few, however there are hackers out there who would have a go for a fee

Many people such as myself would disagree and would have no objection to a DNA database.

you may not object but that won't stop the database being useless for the purpose of preventing or solving crime and terrorism

[splodgeyAl]

Technically, hacking is bypassing the security of a system, so a hacker would be unlikely to use a username and password. Aside from anything else that is a trail that can be followed, and also makes it obvious you have been in there.

 

Technically, hacking is writing computer software - as in "hacking something together". Breaking into a system is "cracking".

 

/pedantry

 

Anyway, back to the point, no live system is 100% secure so the bloke was talking nonsense. For instance, at the highest level of military security, when a system is not in authorised use, it's drives are removed, and placed in a bomb proof safe in a locked building with armed guards. Thats probably about 99% secure.

 

So, a database accessible from the public internet is about as secure as a wooden door with a yale lock.

[Magilla]

Who would be doing the hacking - a criminal?

 

No, the state. It would be very convenient for many many purposes, not least the discrediting or removal from society of politically "inconvenient" people.

 

Sounds more like sci-fi.

 

You would imagine that with the benefit of hindsight on some of the stuff that's gone on in the previous century it'd be obvious what the implications might be!

[esme]

Same with fingerprints, but that's a tried and tested method.

 

http://today.msnbc.msn.com/id/37232072/ns/today-today_people/

 

tried yes, tested no

 

to date there has been no independent testing of fingerprint matching methods, research is apparently just starting on this 100 years after the introduction of fingerprint matching as a forensic tool (src - new scientist - http://www.newscientist.com/article/mg20527522.600-fingerprint-evidence-to-harden-up-at-last.html, sorry you'll need an NS account to see the full article)

 

there are no false positive or negative error rates published for fingerprint matching

 

different fingerprint examiners can come to different conclusions regarding a match

 

the uniqueness of fingerprints themselves is just an assumption, no one has actually proved it

 

http://www.newscientist.com/article/mg18725173.600-editorial-the-myth-of-fingerprints.html - sorry again you need an NS subscription to see the full article but the first three paragraphs give three examples of places where fingerprint matching has failed, these are 3 of 22 cases collected by Simon Cole, a criminologist at the University of California, Irvine (Journal of Criminal Law & Criminology, vol 95, p 985)

 

then there's Shirley Mckie the scottish police woman who was sent to jail because the scottish fingerprint office said her thumprint was found at the scene of a crime, then at the trial they were very careful to only show that part of the print that actually matched while obscuring the remainder that didn't match

 

that's human fingerprint checking though machine verification of fingerprints does have published error rates although mythbusters pretty much destroyed any faith I had in automated fingerprint recognition systems

[Cyclone]

So if everyones DNA was on it, it would work.

Everyone in the world yes.

 

 

I thought DNA could be got from things such as strands of hair?

I'm sure it can, but don't believe everything you see on CSI.

 

 

 

But knowing who was at the scene would be a good start.

You make the same mistake that gets inoccent people put in jail. Finding DNA at the scene is not 'knowing who was at the scene'.

 

 

 

How many burglers, murderers and other criminals would have this unique skill?

How many does it need before you consider it a real risk?

 

 

 

Many people such as myself would disagree and would have no objection to a DNA database.

Feel free to contribute your sample to the police database then, but don't expect your level of complacency to mean that others should be forced to do the same.

 

I notice that you ignored the point about the potential for misuse of the data.

[Cyclone]

Technically, hacking is writing computer software - as in "hacking something together". Breaking into a system is "cracking".

 

/pedantry

 

Anyway, back to the point, no live system is 100% secure so the bloke was talking nonsense. For instance, at the highest level of military security, when a system is not in authorised use, it's drives are removed, and placed in a bomb proof safe in a locked building with armed guards. Thats probably about 99% secure.

 

So, a database accessible from the public internet is about as secure as a wooden door with a yale lock.

 

They're locked in secure cabinets, those cabinets aren't blast secured, the building is normally blast resistant.

That's upto TS anyway, I'm not aware of any higher classification, although there are plenty of additional tags which alter the way information is handled.

[esme]

Everyone in the world yes...

 

no

 

the more people in the database the larger the pool of people who match the crime scene sample who then have to be eliminated them from your enquiries

 

remember a DNA profile is not a DNA sequence, profiles match groups, sequences match individuals or monozygotic multiple birth siblings

 

the more innocent people there are on a DNA database the less use it is as a forensic tool

 

--EDIT--

 

and even with a database comprising the DNA profiles of criminals caught bang to rights in the act of committing the crime the use of such a database as a forensic tool is pretty low

[Cyclone]

I saw your post after mine esme about the number of matches.

[splodgeyAl]

They're locked in secure cabinets, those cabinets aren't blast secured, the building is normally blast resistant.

That's upto TS anyway, I'm not aware of any higher classification, although there are plenty of additional tags which alter the way information is handled.

 

Ok - I stand corrected on the detail - it was about 15 years ago that I had to worry about such things in my day to day life. I'd also say, that what you've described is slightly less secure than what I described - but I'm guessing that.

 

But we're talking about a live, on-line database, so this is moot, so sorry for derailing the convo onto the technicalities of what is actually considered "secure"

[Cyclone]

lol, yeah, anything online, or even just on a private government network isn't secure. The more people have access the less secure it gets.