Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

We can ID people from DNA that shows up in environmental studies

arstechnica.com

Post reply on HN
Print view View original

51–57 of 57 posts

Re: We can ID people from DNA that shows up in environmental studies

#51

Post by Animats »

perihelions wrote: post #18
SketchySeaBeast wrote: post #15

Earlier quoted context omitted.

Then why aren't they already doing that with CCTV and facial recognition?

We are? https://www.telegraph.co.uk/news/2023/05/07/artificial-intel... - "National Highways has revealed that it is set to launch a trial in the coming weeks which will see the high-tech cameras installed in road lay-bys which will scan the area for any instances of littering." - "Unlike traditional CCTV cameras where enforcement officers would have to trawl through hours of footage to find offenders, the new AI cam…

Here's the product ad for that, from Hikvision.[1] Hikvision's traffic violation management system detects speeding, wrong-way driving, cell phone use, failure to use seat belt...

Hikvision is your full-line Big Brother technology provider. Surveillance for home, work, school, roads, parks, vehicles - Hikvision has a solution for a safer world, a world in which something is always watching.

[1] https://www.youtube.com/watch?v=5U2LWuBFIR0

Re: We can ID people from DNA that shows up in environmental studies

#53

Post by godelski »

jefftk wrote: post #27

Here's the paper this is summarizing: https://www.nature.com/articles/s41559-023-02056-2 In cases where a lot of people are contributing DNA to a sample, which I'd expect for most environmental sampling, it seems like you would need impractically large amounts of sequencing before you could identify anyone specific in sequencing conducted for other purposes: you'd need to get a sequencing read which could have come f…

I know very little about DNA and sequencing, so this may be a naive or even bad question. But aren't small sections of DNA unique? Or do you need a bunch of those small sections to actually create said fingerprint? Maybe a better question is asking how many of those chains you'd need to collect (and how long they are). Or how DNA identification even works.

While the fingerprinting might not be something normal researchers do, there still is the question about if the information exists. Unfortunately you only need a single bad actor. Most people might not go through the trouble to deanonymize this user account, but it definitely is possible and would also be possible (though quite taxing) purely through processing my words alone.

Re: We can ID people from DNA that shows up in environmental studies

#54

Post by chris37879 »

godelski wrote: post #53
jefftk wrote: post #27

Here's the paper this is summarizing: https://www.nature.com/articles/s41559-023-02056-2 In cases where a lot of people are contributing DNA to a sample, which I'd expect for most environmental sampling, it seems like you would need impractically large amounts of sequencing before you could identify anyone specific in sequencing conducted for other purposes: you'd need to get a sequencing read which could have come f…

I know very little about DNA and sequencing, so this may be a naive or even bad question. But aren't small sections of DNA unique? Or do you need a bunch of those small sections to actually create said fingerprint? Maybe a better question is asking how many of those chains you'd need to collect (and how long they are). Or how DNA identification even works. While the fingerprinting might not be something normal resear…

They're probably unique, but the search space is so large you can't be sure that the specific section you have is the section you're checking for from the target.

There's only the 4 base pairs in your DNA, so for any given length of base pairs 4^n combinations that are possible to exist. A complete human genome is 3.2 billion base pairs long. So yeah, it's definitely a length vs information density problem.

Though those numbers may seem very very large, they get cut down to size very very quickly. Across all human's, there's very little variation in nearly all of our genome, something like 99.6% of those 3.2b genes are identical between all humans, which leaves 0.4% (around 12m) base pairs of DNA that make you you out of all other humans. We have identified (or at least think we have) some specific regions of those 12m pairs that control certain things, that's what the field of genetics has been up to with things like the Human Genome Project, so, plausibly, if they caught an environmental dna sample that was long enough to identify _which_ gene site they had, and they knew (or could at least guess) what value your genes would have at that site, yeah, they could probably tell whether it was you that was nearby, or at least someone related to you!

Re: We can ID people from DNA that shows up in environmental studies

#55

Post by oilchange »

SketchySeaBeast wrote: post #7
ehhthing wrote: post #5

Earlier quoted context omitted.

We're a small step right now from the police getting access to genetic databases like 23andMe, so if anyone in your family tree has gotten a genetic test it may be possible for them to identify you based on ancestry.

Then they'd know that that person had a relative there. We're still a lot of dots away from going "Well, we found this DNA by accident, it was clearly and obvious Bob was there."

> We're still a lot of dots away from going "Well, we found this DNA by accident, it was clearly and obvious Bob was there."

Once they have a relative, it isn't that difficult to narrow it down to Bob. Think about it, you went from 350 million possible candidates to maybe 50 or less.

Go to the relative and ask them if they have any relatives in or near state X or city Y. That should narrow it down to a handful. And also you can narrow further by gender. Once you find a relative, you can logarithmically cull the search space.

Re: We can ID people from DNA that shows up in environmental studies

#56

Post by 10u152 »

Animats wrote: post #51
perihelions wrote: post #18

Earlier quoted context omitted.

We are? https://www.telegraph.co.uk/news/2023/05/07/artificial-intel... - "National Highways has revealed that it is set to launch a trial in the coming weeks which will see the high-tech cameras installed in road lay-bys which will scan the area for any instances of littering." - "Unlike traditional CCTV cameras where enforcement officers would have to trawl through hours of footage to find offenders, the new AI cam…

Here's the product ad for that, from Hikvision.[1] Hikvision's traffic violation management system detects speeding, wrong-way driving, cell phone use, failure to use seat belt... Hikvision is your full-line Big Brother technology provider. Surveillance for home, work, school, roads, parks, vehicles - Hikvision has a solution for a safer world, a world in which something is always watching. [1] https://www.youtube.co…

They’re also fully owned by the CCP and have bespoke algorithms for identifying and alerting the authorities to specific racial traits (Uyghurs) for future arrest, enslavement and genocide!

Fun guys, Hikvision

> https://www.bbc.com/news/world-asia-china-22278037

Re: We can ID people from DNA that shows up in environmental studies

#57

Post by sublinear »

sb52191 wrote: post #49
sublinear wrote: post #14

I am no statistician, but wouldn't mass DNA collection reduce the reliability of DNA evidence?

I think it depends on use cases and advancements in our understanding of time's effect on the DNA captured. I.e. I could imagine that in the not too distant future, we know that DNA in air (exposed to sunlight) degrades at a certain percentage over time, and therefore could be used to determine if a person was near a given location recently. Sort of like carbon dating. You could imagine law enforcement using this as…

> Drive around with a device that constantly captures air and checks it for the DNA of a suspect

Why would they drive around when there are undoubtedly hotspots where they would install sensors?

Post reply on HN