h a l f b a k e r yA riddle wrapped in a mystery inside a rich, flaky crust
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
I had trouble thinking up a good name for this Idea, and may change it if something better is suggested.
The Idea basically consists of collecting samples of DNA of EVERYONE, but only AFTER they die. After collection, the DNA is compared with criminal databases to see if the recently-deceased was
the perpetrator of a crime, who got away with it. While the perpetrator is now deceased and can't be punished, we CAN, very likely, bring closure to a lot of unsolved (in the sense of "whodunnit?") cases.
A Modern Event
https://yro.slashdo...n-state-killer-case
Relevant to something mentioned in the annotations here, the DNA of relatives might lead to the DNA of perpetrators. [Vernon, Apr 27 2018]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
| |
I understand there are boatloads of biological materials harvested from criminals which are queued up for DNA testing. Given that the current method seems to produce a backlog, and that current criminals seem a more promising populace than the random dead to match with old crimes, I think this idea might not be practicable. |
|
| |
//hmm...This is ... intriguing. Sure, why not?// |
|
| |
Fishbone, but not quite sure why. Maybe because dead criminals sometimes aren't criminals. |
|
| |
I'm not sure why, either [AQ]. Vernon suggests taking
samples of everyone after they die. Then, IF some of
them turn out to be criminals, cases solved. This
would not indict anyone who was NOT a criminal. |
|
| |
Wish I had another bun to give... |
|
| |
//cases solved// rather inconveniently for the dead person who can't defend themself. |
|
| |
[flyingtoaster] your comment made me switch from
[+] to [-]. You are absolutely right. DNA is only part
of the case, just because his DNA is there doesn't
mean he did it (also it doesn't mean that the DNA
really was there - maybe it was planted by the person
who really did it) ... this just just too simplistic to
work. |
|
| |
well... that, mitigated a bit by common sense: it could simply result in another couple lines in an investigation file. |
|
| |
Seems to be wanton abuse of human rights. US
Homeland Security would love it. |
|
| |
Some very good points here. I might unbun. Maybe
my comment about "cases solved" is too strong. |
|
| |
True, a dead man's DNA wouldn't necessarily 'solve'
a case. But, it would still be valuable evidence,
would it not? In an active or pending case, it's
evidence in an
ongoing trial, nothing more, nothing less. In a
closed dead-end case, it *might* provide a solution. |
|
| |
Whose rights are being abused? The dead person's?
Does a dead person have rights? I don't know. (I'm
sure someone here does.) |
|
| |
If the DNA only provides 'another couple of lines in
an investigation file,' what harm in that, weighed
against the possibility of a solution? |
|
| |
You cannot really ever collect genetic information on only
one member of a family. You inevitably learn something
(stochastically) about other members of the family from
whom no samples were taken. A database of everyone
who dies would let you do much of what you could do with
a database of everyone still alive. This needn't bother
anyone who *isn't* uncomfortable about a vast government
database of everyone's genome. |
|
| |
For the rest, keeping information on familial relationships
out of the database wouldn't work, since (I think) familial
relationships among the dead could be inferred from the
DNA alone. You'd have to somehow anonymize the
samples so that it was imposible to know who the "donor"
was in life. |
|
| |
//You cannot really ever collect genetic
information on only one member of a family. You
inevitably learn something (stochastically) about
other members of the family from whom no
samples were taken.// |
|
| |
But, only if you look for 'something' about those
family members. I don't understand how charting
(or whatever they call it) someone's DNA
automatically charts anyone else's, similar or not. |
|
| |
Are you saying that since the information is there--
even if only statistically--it will necessarily be used
against someone else? How about a law prohibiting
such a thing? |
|
| |
[Boomershine] //I don't understand how charting (or
whatever they call it) someone's DNA automatically charts
anyone else's.// Because it's not just *his* DNA. Half of it
is his mother's, the other half his father's. And half of it
belongs to each of his children (a different half in each
child). And smaller fractions are shared with more distant
relatives. |
|
| |
Let's take a simple example. John Doe tests positive for
Huntington's Disease, which is autosomal dominant with
very high penetrance; it's incurable, progressive, and
uniformly fatal. His daughter, Mary Roe, now knows she
has a 50% chance of developing Huntington's. She wasn't
tested, and she didn't want to know -- but she knows now.
OK, now suppose Mary has three children. If she carries
the gene (50% chance) then Tom, Dick, & Harry each has a
50% chance of developing Huntington's. If she tests
negative, they can heave a sigh of relief: virtually zero
chance of developing it. Tom & Dick beg her to be tested;
Harry prefers to remain in ignorance: if he knew he carried
a 50% chance of Huntington's he might lose his job, his
health insurance, it'd put a strain on his marriage, and he'd
be unable to enjoy his remaining healthy years because of
that sword of Damocles hanging over his head. Tom & Dick
convince their mother to be tested, and she's positive.
Harry now knows what he didn't want to know -- because
someone *else* gave a tissue sample. |
|
| |
It's stochastic knowledge, and, in order to get probabilities
as high as 50% you have pick an extreme example
(monogenetic, dominant, high-penetrance), but you see
what I'm getting at? |
|
| |
Thanks for the genetics lesson, really. I think I see
your point, but I still think it's different than mine. |
|
| |
In the example you gave, the culprit is a disease, or
the genetic disposition to manifest one. That right?
"John Doe tests positive for Huntington's Disease." |
|
| |
But this DCDatabase would not care, or test for
anything--except for a match in another database;
the deceased's DNA against DNA which was--I assume-
-found at a crime scene, but did not match any
known person. We are only looking for a match, not
a specific genetic characteristic, such as a disease. |
|
| |
No match, end of story. Nothing about this particular
dead person is looked for or shared with anyone
else. |
|
| |
If a match IS found, that is the only criterion
considered--that it matches. Not that it matches--oh,
and, by the way, he had the gene for X disease. Our
sympathies to the Doe family. |
|
| |
Am I still missing something? |
|
| |
[Boomershine], the concern being raised, that you are missing, is that there is such a thing as a "partial match". Which means the criminal could be a LIVING close relative of the deceased whose DNA was compared to the criminal database. I neglected to think of that when I first posted the Idea, but since this is the HalfBakery, I see no reason to delete the Idea on that account. It's simply more evidence that this Idea is HalfBaked! |
|
| |
Crystal clear, [Vernon]. Thank you...(sorry it took so
long to get it through my head). This is basically
tantamount to getting DNA from the whole
family....and that's what you've been trying to tell
me all along...oy. |
|
| |
Keep the bun. It was worth the lessons to me. |
|
| |
It doesn't really have anything to do with diseases, and I
shouldn't have used that for an example. |
|
| |
The idea is, you know something about the DNA of A, by
sequencing the DNA of B, provided you know that A is
related to B. In the Huntington's example, you know, with
50% probability, that A's DNA has a particular mutation (it's
not, strictly speaking a mutation, but never mind),
without ever testing A's tissue -- you only tested B, who
might be dead, and therefore in the dead people DNA
database. |
|
| |
So let's translate this from a clinical to a forensic example.
The security services suspect K of being a terrorist, and
search his home. In the kitchen, they find several
unwashed drinking glasses, and from the traces of spittle
on their rims, using PCR, they obtain some DNA. They'd
like to match those agaisnt a database of the DNA of every
living citizen, but the public objected to such a database
on privacy grounds. They do, however, have a database of
every dead citizen, which includes both your late parents.
We will assume you are male, and all your sibs are female,
or that you are an only child. |
|
| |
Let's suppose your father had 8 comparatively rare
polymorphisms, and your mother had 10 other
comparatively rare polymorphisms, and lets say the DNA
they got from the drinking glasses has 4 of your father's
polymorphisms and 5 of your mother's. Depending on the
number and rarity of the polymorphisms, that could be
enough to get your name on the shortlist for an interview
or something worse. The effect is much the same as if
there *were* a database of living citizens' DNA. To prevent
it, the DNA in the dead people's database would have to
be de-identified, so that no one could tell that the people
those two samples came from were your parents. |
|
| |
Or alternatively //a law prohibiting such a thing// My
anxiety that that's not enough to keep Big Brother at bay
has nothing to do with genetics, so you can feel free to
ignore it. |
|
| |
I'm sorry, this has been long winded and not a little
pedantic, but I couldn't bring myself to leave it half-
explained. |
|
| |
Edit: Oy veh. [Vernon], who has a reputation for
verbosity, expressed it more succinctly than I. |
|
| |
//I couldn't bring myself to leave it half-
explained.// |
|
| |
I understand. It was incumbent upon us both to
persist. Thank you for your patience. |
|
| |
//I had trouble thinking up a good name for this Idea//
How about "Random Trawl to Clear Unsolved Cases Off the Books and Claim that 'Some Dead Person' Did It Despite an Almost Complete Lack of Evidence to Back it Up and Due Process Can Go Hang!"? Sounds like a pretty catchy title to me! |
|
| |
We all share some of our DNA, by proxy, with Big
Brother? (Is that still politically correct? Should we
now call "him" Big Sibling?) |
|
| |
Thanks, but I'd rather not know nor have anyone else
know. |
|
| |
I'm not particularly paranoid about police states or
govt invasions of my privacy. I have nothing to
hide, never been in trouble, etc. OTOH, I am not
in favor of general invasions of privacy at all. I
know we have to watch and guard these things. |
|
| |
But, my failure to grasp the implications of this
idea (until patiently enlightened by [Vernon] and
[mouseposture]) make me feel somewhat
underqualified to a be HalfBaker. |
|
| |
I let the dialectics of the argument obfuscate the
actual points being made. For shame. |
|
| |
I'll be under the porch for awhile. |
|
| |
If we're collecting DNA of the dead, why not fingerprints too? Hell, lets photograph and measure the stiffs so we have the potential of matching eyewitness accounts. |
|
| |
Actually, the dead don't need their bodies anymore... just let the police expropriate everyone's corpse in the name of "public safety." |
|
| |
It should, in theory, be possible to input the deceased DNA into a secure database that only returns a hit on a 100% match. The only problem would be coming up someone you could trust to secure the database. |
|
| |
[ME] I agree. This was really what I was aiming at.
Now, I'm just too tired. |
|
| |
Security is an issue with any evidence database,
though, isn't it? |
|
| |
As I understand, the DNA records used in evidence are partial, partly because much of our DNA is common to all of us (and to chimps, shrimps and daffodils) and partly because the sheer volume of data would be overwhelming. |
|
| |
I don't trust DNA matching. I live in Norfolk - 'that' sample could match half the population of my village and several surrounding villages.
Seriously - I often hear statistics quoted along the lines of "the chance of the defendent's DNA matching the evidence sample by chance are a million to one against". This ignores the fact that in a country of 60 million, the other 59 are likely to live fairly close to the defendent. |
|
| |
The analogous situation is fingerprints. I am not sure how these are dealt with. |
|
| |
I am sure that unknown persons found dead undergo fingerprinting, in hopes of identifying the dead person. Certainly if a fingerprint is on file as belonging to a person (not necessarily a criminal - federal employees are fingerprinted) the dead person will be so identified. |
|
| |
What if the dead perso's prints are a match for a fingerprint found at a crime but never matched to a live individual? I wonder if such prints come up in a search. |
|
| |
//As I understand, the DNA records used in evidence are
partial// Yes, and I think I can think of other reasons why
matches would always be partial at best -- If so, a
requirement for 100% match would be unhelpful. Of course,
one could insist on 93%, or 86%, or whatever ... but how
would one select, and justify, a threshold? |
|
| |
Well, once you drop below about 93% match you are in the territory that includes all other primates. Below 85% and you have to rule in all mammals. Below about 70% and you're looking at a broad range of invertebrates as potential suspects. |
|
| |
Sure, if you have the perp's complete, uncontaminated
genome. However, what if the DNA in your sample is
degraded to the point where you only have, say, 60% of the
criminal's complete sequence, making 60% the maximum
possible match? Now consider that the sample may be
contaminated with other organisms' DNA. This is what seems
to me to make difficult the problem of choosing a threshold
for a "match." |
|
| |
Matching is more precise than that. If you get a 99%+ match, it is almost definitely going to be unique. Close relationships are unlikely to produce a false positive for a high criteria match(except in the case of identical twins). Probabilities come into it much more when you are trying to determine relationships, but this wouldn't do that. (oh, and a 99%+ match does not mean the DNA is 99% identical, it means the specific areas checked are 99%+ identical, like fingerprinting, only certain key features are checked). |
|
| |
Sorry [Vernon], me no likey. I don't like the idea of the bureaucrats getting my DNA (posthumous or otherwise). |
|
| |
Vaguely recalling stats now, so no promise of accuracy, but the principle stands... |
|
| |
Current DNA testing checks a number of parameters and can distinguish with a resolution of 1 in 50,000,000. Any more than that becomes prohibitively expensive. The population of the UK is about 60,000,000 so at least 10,000,000 people have an exact match to someone else under those tests. In the US it gets worse, with a population of 300,000,000 then *everyone* has the same DNA as 6 other people. Spread to India or China and you're into populations in billions. |
|
| |
Taken to extremes, you could find dozens of dead people with matching DNA to that found in a crime scene. The perpetrator gets off on a technicallity every time. |
|
| |
Don't go looking for random matches. That will only cloud the issue. |
|
| |
The theoretical probability is closer to 1 in 13 quintillion, although the actual probability seems to be lower. |
|
| |
No matter what, if you find a match it's not case closed. If the person in question was in the right area at the right time, however, it suggests highly that they are of interest. |
|
| |
Please note I haven't voted on this, becuase while I think it would be useful to clear cold cases, I also do thinkit would be an invasion of privacy, and probably impossible to keep secure. |
|
| |
//means the specific areas checked are 99%+ identical// I
was assuming some better-than-current technology. As
DNA
matching is currently done, i.e. only specific areas
checked, the estimates of probability of a
false positive depend on your model of the statistical
variability of those specific areas checked, in some human
population. The operative word here is model -- that's
why
these estimates are disputed. (Of course they're disputed
in
court, where it's routine to dispute even things that are
obviously true.) |
|
| |