Hacker News Reader

Comment by grey413 | Hacker News Reader

The champion of the theory attributes the lack of genomic similarity to repeated back-crossing to one of the parent populations. To put it in other words, the droplet of initial hybridization got diluted in the larger gene pool, but the novel genes (and thus traits) remained and underwent selection.

That being said, diluted gene contribution is not the same thing as no gene contribution. To even begin accepting such a theory I would require direct genetic evidence showing considerable horizontal gene transfer between porcine and hominid lines.

View on HN

I saw that claim in the article, yes. I'll admit that I don't know enough about genetics to make strong claims here, but it doesn't sound remotely plausible to me. He wants to simultaneously claim that 1) hominids got so much porcine DNA that they have lots of substantial, recognizable anatomical features as a result, and 2) hominids have so little porcine DNA that it's all but invisible in our genetic code.

I'm not going to say that's impossible, but it sounds like one heck of a stretch. Even just sitting here thinking about it, most genetic inheritance happens one full chromosome at a time. We clearly don't have any full pig chromosomes, so to make this theory work you'd have to have a whole lot of lucky recombination events (chromosomal crossover, etc.) that preserved only the precise genes involved in all these "distinctive pig traits" and got rid of the rest. So what's the selective effect that selects extraordinarily strongly for this random selection of pig-like anatomical traits but against all of the other pig genes that would have usually been linked to them?

In short, this is a very extraordinary claim, and it requires equally extraordinary evidence, especially given how remarkably consistent the known genetic evidence has proven to be.

In general I agree with you, it's an extraordinary claim with less than extraordinary proof to put it mildy. However I thought I'd make one small correction on your genetics.

> most genetic inheritance happens one full chromosome at a time.

is actually not true. Sperm and eggs have only one copy of each chromosome instead of two like most cells in the body. However, that one chromosome is a pretty good mixture of the versions received from each parent due to recombination events that occur randomly during maturation of those cell types. Linkage between nearby genes does exist, but it's not nearly so strong as you seem to imply. Even in a single generation inheritance of two genes on either end of the same chromosome is nearly uncorrelated.

Thanks! I didn't realize that recombination was so frequent. I'm glad to have a better sense of that now.

I don't see why it's that implausible. You are descended from almost every single human being alive 1,000 years ago. Can you find any significant amount of DNA contributed from only a single one of your ancestors?

What he is suggesting is that a single hybrid made it's way back into the hominid population. It had children with other hominids, and those children would have had half as much pig DNA. They had children with even less, and so on. After awhile there would be almost nothing left of the pig ancestor. The only genes that would survive such dilution would be ones that were significantly selected for.

"You are descended from almost every single human being alive 1,000 years ago." Um. No.

Yes actually. Here is an explanation why. http://dgmweb.net/Ancillary/OnE/NumberAncestors.html

It is obviously false. There are many people that were alive 1000 years ago that you have no relation to due to geographic separation. Just because you can analyze something using statistics doesn't make it right. It even explains some of the discrepancy in the description of that chart — your ancestors are not unique and therefore the entire chart is just statistical hyperbole.

I said almost everyone, obviously you aren't (likely) related to remote islanders. However you'd be surprised how much mixing there is between population. All you need is one individual from another population to mix with another and within a few generations he is everyone's ancestors.

I'm not certain how much mixing there was between the new and the old world. However it did happen a lot and it's likely a large number of people do have some new and old world ancestors.

You are overstating the amount of mixing between populations (and somewhat discounting the mixing within populations).

1000 years ago, people in villages and towns were regularly marrying their (2,3,4) cousins and rarely marrying people from 'different lands'.

Another way to look at it: for 1 person to become the ancestor of just several thousand people takes 'a few generations'. Getting to whole populations takes lots of generations.

"You are descended from almost every single human being alive 1,000 years ago."

If a generation is about 30 years then you will have close to 33 ancestors lines in your family tree over those 1000 years. That makes 2^33 ancestors or 10 billion, that's impressive (well there is probably a large overlap)

Yes, but remember the "bottleneck" from 70,000 years ago, when the human race was nearly wiped out. At that time, because of a nuclear winter brought on by a volcanic supereruption, there may have been fewer than 10,000 humans in the entire world.

http://en.wikipedia.org/wiki/Toba_catastrophe_theory

So, based on that, we're descended from fewer people in the past than most people realize. Another way to say it is that most of our ancestors were cousins.

That overlap is more than large, it's actually the most important factor in the calculation. You only need to go back about 6 generations (200 years), and there is practically no interchange between geographically separated groups, which means that I probably don't have any genes from some random asian living 1000 years ago (although we both almost certainly share genes that we inherited from some far more ancient common ancestor).

"You are descended from almost every single human being alive 1,000 years ago."

Not exactly 1,000 years, but more like 120-135,000 years for the male line [1], and 200,000 years for the female line [2].

[1] "Sequencing Y Chromosomes Resolves Discrepancy in Time to Common Ancestor of Males Versus Females " dx.doi.org/10.1126/science.1237619

[2] https://www.princeton.edu/~achaney/tmve/wiki100k/docs/Mitoch...

No it's much closer to 1,000 years. Every generation the number of possible ancestors you have doubles. 2 parents, 4 grandparents, 8 great grandparents. Within a few generations back, you have literally millions of ancestors.

Doesn't that assume no overlap? It seems like the odds of 2^n unique ancestors would be pretty low by the time you get to 10.

This seems like a reasonable discussion of the issue:

http://gcbias.org/2013/11/11/how-does-your-number-of-genetic...

After ~15 generations your genealogical ancestors aren't all that likely to have contributed much to your DNA.

Evolutionary geneticists don't trace individual lineages that far out, nor would they need to. The hypothesis under question is that the porcine/proto-hominind cross provided a variety of important, wide-acting traits that all humans share. That means that every human has them, and that every chimpanzee doesn't have them. That would show up in a comparative genomic assay.

Right, so he's claiming that the genes responsible for this really pretty random list of traits ("protruding, cartilaginous mucous noses"? "Prostate encircles urethra"? "Alcoholism"?) were so strongly selected for that they became 100% ubiquitous in the human genome. Meanwhile, every other trace of pig ancestry got diluted away to nothing.

I understand the idea of dilution, I just don't understand why he thinks that so many random things would survive it to become defining features of the human species. Remember, he's not just saying that 2% of living humans carry Genghis Khan's Y-chromosome or have red hair or something, he's saying that all of these traits became completely universal.

I submitted the article I link in a subthread here:

https://news.ycombinator.com/item?id=7539143

The author posted some replies, one of which points to this article:

http://phenomena.nationalgeographic.com/2013/05/07/charlemag...

That puts it at about 3400 years (the article gets the number from some paper on the topic).

Reply on news.ycombinator.com