Hacker News Reader

It wouldn’t be you in any case. Basically the same as an identical twin.

Also, why would they bother?

Maybe if you were a celebrity of some sort, maybe, but even then..

I see a near future scam where people seek out genetic material of the wealthy to create children so that they can sue them for child support.

Imagine someone doing this to the judge that ruled they owe child support for their child from a previous marriage out of spite.

They could sell the organs back to you when you need one. Or the whole body for a brain transplant.

(This was a plot point in a sci-fi story I vaguely remember.)

You're probably thinking of "The Island"[1]. Good movie, I saw this myself a long time ago.

[1] https://m.imdb.com/title/tt0399201/

It’s a different one and I actually remember it better, but I didn’t want to spoil the plot twist for that story by naming it.

Err, so what was the one you were thinking of? I've seen The Island before, so would be potentially interested in something else along the same lines... :)

Maybe https://en.wikipedia.org/wiki/Immortality,_Inc. ? (there was a movie too: https://en.wikipedia.org/wiki/Freejack)

Thanks. :)

Even that is unlikely. We don't yet know what exact information is enough to recreate a human (or his twin) but by now we know that DNA alone is not.

This data would be of much more interest to your health insurance company. The ultimate preexisting condition pricing guide.

Mostly a US concern, as usual.

Health insurance companies are prohibited by law in the US from considering preexisting conditions.

You mean, “at present”..

No way to know how the regulation will change in the future depending on the amount of lobby dollars that are thrown.

If you want to hand wring about future changes you don't need to look at dubious inference of risk from DNA. Just consider cancellation of insurance policies if you come down with some illness that requires expensive long term treatment.

The point of this attack vector is to drop patients who are high risk BEFORE they get diagnosed so they can claim it isn't for any existing condition.

This is especially scary in small, closely related communities like Iceland or Utah where a relatively low number of collected samples can be extrapolated to large swaths of the population that didn't give their genetic info away.

Also that protection is relatively new in the US and is constantly being attacked politically so it isn't unreasonable to think about a world where it is removed like other longstanding health related protections.

The Genetic Information Nondiscrimination Act is almost as old as 23andMe and older than the prohibition on discrimination for preexisting conditions, so it would be strange to come up with a hack where you break the first to try to get around the second.

This is a great reference and it's a good point - there are multiple lines of defense here because the danger is so high.

If the implication is that it is so old it won't ever be challenged it's worth noting that roe v wade was in '73.

US Federal Genetic Information Nondiscrimination Act (2008) (https://en.wikipedia.org/wiki/Genetic_Information_Nondiscrim...)

Also, since 1991, many US states had already passed laws that prohibit insurers' use of genetic information in pricing, issuing, or structuring health insurance.

And they were prohibited from considering genetic information by the Genetic Information Nondiscrimination Act before they were prohibited from considering preexisting conditions more generally by the Affordable Care Act.

23andMe being a Google spin-off I wouldn't be surprised if they do a full sequencing regardless of whether you pay for it or not. At the very least I would expect them to store the samples indefinitely so they can do it once it's economically viable.

As I understand it, full sequencing is a _lot_ more expensive. 23 and me currently shows the genotyping is like $100 while sequencing is $1000. It’s probably too expensive and time consuming for them to sequence everybody’s samples unless they pay for it.

Only a couple years ago full sequencing was closer to $10,000

Everyone knew price would drop in time, but, assuming you’re correct, it’s amazing to hear that’s happening

499 euro was the cheapest that popped up on google https://www.dantelabs.com/products/whole-genome-sequencing

Sequencing a whole genome at 30x has been around 1k for the last 10 years at least.

Housing the samples in a controlled environment for an extended period of time is likely way more expensive than sequencing. I suspect the samples are discarded once genotyped.

I'm no biologist but the samples seem to survive being shipped by standard mail so it doesn't look like they would need very strict storage requirements. If for any reason they do, since they take very little space it shouldn't be too expensive per-sample. What am I missing?

It's not technically a Google spin-off. Google and Sergey Brin invested in it, but it was never part of Google.

Well the rightful answer is "it's complicated" DNA wise yes there are 3 billion bases and they sample about a million of them, so 1/3000 bases.

THing is about the bases

1. most of them are the same in everyone

2. the ones that are different tend to be correlated with each other locally and thus captured by their 1Mb assay

so

3. you can infer all the stuff they didn't sequence with pretty high confidence. Not a "in theory" but more like "has been a typical thing to do in the statistical genetics field for at least a decade"

4. outside of the genome there is the epigenome which may or may not be relevant, it undergoes very specific resets short after fertilization

also worth noting

5. bioinformatics is an imperfect and algorithm based science. Reads are aligned according to error and difference profiles (i.e. string mismatches and the most and least types of mismatches). So unless things are finely calibrated, a level of analysis most bioinformatics don't do, deeply study their read alignment penalties for particular data sets, even a robust bug-free read aligner correctly applying penalties will have bad alignments, false positives, false negatives

anyways I for one hope the futurepopele will clone me from the bytestream