dcdanko

The academic paper for those interested: https://www.cell.com/cell/fulltext/S0092-8674(21)00585-7

You have the right idea. In some cases the Dead Sea Scrolls were literally shredded- ripped up into small pieces that were sold individually to tourists.

The goal here was to identify parchment fragments that came from the same animal so the original (unshredded) document could be reconstructed.

While we were able to link a lot of parchment fragments we didn't go so far as to actually reconstruct the text on those fragments. While we constrained that task somewhat it's still nontrivial since we don't know what order the fragments should go in, if there are still missing pieces, etc.

Most of the fragments (sorry don't have a %) could be mapped to a genome. In cases where a fragment could not be mapped it's more likely because of degradation than novel biology.

That said we did identify some novel microbial genomes. Not clear if they are actually ancient though vs contamination in the last 100 years or so.

Here's the paper for anyone interested https://www.cell.com/cell/fulltext/S0092-8674(20)30552-3#.Xt...

I'm one of the middle-authors on this. Happy to try and answer questions.

You should be able to get all the supplies online

The standards for authorship in this study were pretty strict. Something like ~2 supplementary figures just for a co-authorship.

Source: I'm in one of the labs on this paper (but not one of the authors)

Still not applicable to Trump because the U.S. Congress has power of the purse, not the President.

His compound rate of wealth increase, post paying taxes on the sale, is around ~7.5%. That beats the S&P 500 over that time, which very few professional money managers can do.

Very few money managers that are accessible to people with less than 10^8 USD can beat the S&P after fees.

Bridgewater, (Ray Dalio's fund mentioned above) regularly gets 12.5% annual returns and they are most notable for being consistent, not high. Pre-2008 some investment banks were getting up to 20%.

A billionaire making 7.5% annually is fairly unimpressive if their goal was to make money (which it might not have been for Cuban).

NASA is only trying to keep Mars sterile for 100 years or so, probably doesn't matter much on the scale of terraforming

Both would be illegal* since your keys are info that could deanonymize patient data. However, you would be allowed to store them on your own PC if you follow proper procedures.

This is more important than it might seem from a purely tech perspective. HIPAA is partly designed to guard against 1) improper use of encryption 2) downstream contractors who are malicious or careless

The one-time pad is probably fine but it's easy to imagine lesser encryption being broken (especially by bugs). This problem gets much worse once PHI is stored on a medium, like a blockchain, where it can never be taken down. After all you probably wouldn't be happy if your health data was on the internet protected only by SHA1.

There are probably better solutions to all of these issues but HIPAA is intentionally conservative. In many respects HIPAA is a financial, not a technical, law.

* HIPAA doesn't make improper storage illegal in the sense that you go to jail if you do it wrong. It exposes people who handle data improperly to massive fines, usually when actual breaches occur. This is part of why BAAs are so important.

mood, circadian rhythm, ancestry, region of the world, diet type, infection, sex, certain diseases, and probably more!

Why? 32 is a big for a human study.

They recruited 32 people that didn't exercise, they didn't choose them after they already had the data.

Even the composition of your microbiome itself can alter your brain chemistry (actually, it's usually called the gut brain axis).

The study here used fecal transplant from lean to obese mice as well as a less invasive test; so it really does seems like a meaningful distinction.

Probably not much at all. SNPs and small indels tend to be have many neighbors with which they're highly correlated. If a variant caller missed a single SNP it's likely that it still called a bunch of others that nearly always co-occur. In most cases downstream association studies would be unaffected.

It's actually possible that DeepVariant is implicitly learning some of these correlations (1). This would make it really really bad for picking out the rare persons that don't fit a trend (and tend to be very important for identifying disease loci). GATK definitely does not know about correlated SNPs.

(1) The paper implies this is not the case, saying that DeepVariant works for other genomes without retraining, but they don't show the relevant results.

The figures in this paper use pretty deceptive scales. To be clear, DeepVariant is 0.5% better than a tool built in ~2010 (GATK), on DeepVariant's best test.

GATK is still the standard, not because better variant callers don't exist, but because it's more important that everyone uses the same tool for comparisons between studies.

Very small transaction outputs can't be used in practice because they're less than the required transaction fee.

This looks pretty cool!

Strongly agree. POS is just a dream (until someone proves me wrong :P)

He provides three examples where an idiom is sometimes bad and sometimes good...

The point of the article is just the classic 'foolish consistency is the hobgoblin...'

Personally I think that the value of (1) Money is extremely dependent on proven track record and the network size. Since the value of (2) and (3) are dependent on (1) I think this has outsize importance. Though as programmers I think we tend to underestimate this value.

Not to say Bitcoin isn't in a speculative bubble, who knows?

Aside: I don't know if it's fair to call bitcoin pseudonymous. Tumblers mean that one can effectively obscure their transactions if that's their goal, pseudo-pseudonymous if you will.

It's shopworn but good tech doesn't always win. Bitcoin has a big(ish) network of users and a proven codebase.

The decentralized nature of blockchains means that they're hard for individual corporations to control and profit from. That hampers adoption of new closed cryptocurrencies (not Monero or Ethereum though) but can serve as a powerful public utility.

Definitely, blockchains aren't secure in the way that many Ethereum apps seem to think they are.

But for certain types of lower stakes applications (say a log running on a mesh network of sensors) blockchains can provide a simple way to do something that could otherwise be fairly complicated. Certain blockchain implementations (I like MultiChain) are really easy to use.

I think a benefit of blockchains is that a single(ish) protocol can support many different distributed applications.

It's sort of like apps that run on Electron, performance might be bad but it's easy on the developer.

A concise and useful list of suggestions.

You mean like the States?

Surely the MID would need to be slowly phased out or just capped?

It can be useful for small scale real estate developers who can use their home to get cheap capital.

Yep, bigger issues are the lack of low cost housing units to buy and older institutionalized discrimination. Not the MID.

Agreed. This article makes it sound like you can get a mortgage practically for free.

The couple of proposals to get rid of it still leave the MID in place up to $500,000 though which roughly matches what you found.

Komodo dragons have a relatively small population which researchers wouldn't want to disrupt on a large scale.

Not that researchers collect their mice from the wild...