I recently started reading that book have been constructing a drinking game based on it. I'd definitely recommend the book so far, but wouldn't necessarily recommend this drinking game:

* Every time a new process is described as being a factor of x times faster/cheaper/better than its predecessor, take log(x) sips of your drink

* Take a sip of your drink for every unique 3-Population Test or a 4-Population test that is described (where uniqueness is defined as the set of populations going into the test).

* Take a shot every time the difference between Population X's DNA and East Asian DNA is used as a unit of measure. Take an additional shot for the quantity if it's more than one, e.g. "The difference between Denisovan and European Hunter Gatherer DNA is twice as large as the difference between Denisovan and modern East Asian DNA." would result in two shots.

* Take a shot every time certain events are described as happening in sequence, but the years also have overlapping error bars

I'm surprised that the prize wasn't awarded jointly with Reich.

The Nobel Prize tends to favor initial breakthroughs. Svante's lab sequenced the very first complete Neandertal genome, showing it was possible to reconstruct genomes from ancient DNA, with high-throughput sequencers, and developing the necessary lab protocols and software to put it all together (with his lab). Svante was senior author on the Denisova paper as well. So, I am not surprised. There was no major advance needed for Reich to do his work further elucidating the origin and dispersion of various hominin species and ancient human populations.

Paabo told me (when he visited to give a talk at my institution) that the first few years of Neandertal sequencing, he was just sequencing his own genome accidentally, since he didn't have truly stringent lab technique. He was out ahead of everybody else for at least 2-3 years at the time, and could demonstrate to people trying to replicate his results that they were just using contaminated samples. Sequencing is absurdly sensitive- it can pick up a real signal from just a few individual molecules.

Does this mean that inexpensive DNA tests you can buy online are likely to contain contaminated samples? As the person taking the sample at home hardly has stringent lab tech available?

No, tests like that work with an abundance of DNA and so contamination is less likely. A DNA test typically works with saliva or blood which contains a ton of DNA> That DNA can then be amplified, which amplifies both the signal and the noise, but the noise is very low and can be trivially ignored after the fact.

With paleogenomics, you're collecting a tiny amount of DNA, mostly degraded, from ancient samples, and so the "noise" (contamination DNA) ends up being close to or more than the signal you get from the sample.

Details here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3926038/

The issue with paleo DNA is that the target is much lower in concentration than potential contaminants from the lab worker. etc. This wouldn't be the case in DNA tests where the concentration of the sample would be comparatively high.

That was a common problem in Paabo's early work. He also "sequenced an Egyptian mummy" that turned out to be himself as well.

Seriously? Nobody was verifying that end to end?

This reminds me of another story - BGI was sequencing the chicken genome, and for a while, a lot of people who did other sequencing projects kept receiving chicken DNA in their sequencing data files (as contamination).

As it turns out, paleogenomics takes next level cleanliness, far more than anyone expected, particularly when you're trying to sequence something so close to human DNA.

Sometimes verifying the little details can get in the way of publication velocity. It's a risk many advanced scientists take because the chance of getting caught (publishing something that is later shown to be obviously false using careful control experiments) is low, and the risk to your career of being second to publish.

My feelings exactly. (That aside - delighted that Pääbo and his field are getting this recognition.)

>"Who We Are and How We Got Here" by David Reich

IMHO, I'd take the conclusions offered here with a grain of salt. It's applying statistical methods to very improbable events (ancient bones surviving with DNA) to determine other very improbable events (evolutionary DNA change). The conclusions at least somewhat rest on the assumption that those two things have relatively consistent probabilities. That could be true but it's probably not and if it's not it's somewhat unclear how that would bias the results that the book draws its conclusions from.

> other very improbable events (evolutionary DNA change)

Sorry, I don't understand this part.

Are you suggesting that evolution is a "very improbable event"?

Among serious scientists evolution is already settled and beyond dispute.

I don't see any basis for an argument if you're suggesting that.

Denying evolution belongs to religion, not science.

No, they are referring to the fact that the events that lead to permanent changes in the genome are extremely rare- the vast majority of mutations are not fixed.

> statistical methods to very improbable events

What is the alternative?

If I had one I'd be clearing a space on my shelf for my Nobel Prize

Seriously though, I don't think there's any way to know for sure. Like imagine if it were dice rolls. You can look at the results and draw conclusions to see if it's a weighted dice. But what if my table biased rolls or if some results were lost? It's probably reasonable to conclude that they weren't but no way that I can see to be sure.

Random walks.

Here, have a UTF8 code point for the next comment: ä

You're welcome!