Mendel was wrong? Researchers say “law of segregation” doesn’t hold

wd400: You keep repeating that the existence of an allele "causes" some outcome. Yet you don't provide any details other than the fact that the allele is correlated with the outcome. Let us be very clear: Unless we know what the allele does to cause result x, which leads to y, which ends up with outcome z, we cannot say that the allele "caused" z. It simply doesn't follow. It isn't some esoteric philosophical distinction that can be dismissed with a wave of the hand. It is basic logic 101 and an elementary aspect of any rational scientific inquiry. And we certainly cannot say that the allele is some "selfish gene," whatever Dawkins' illegitimate personification of genes is supposed to mean. Now it could be that in the paper you cited the authors in fact show what the allele does biochemically to cause x, which leads to y, and ultimately results in z. As I said, that would be most interesting. But it certainly wasn't clear from the abstract. And as is so often the case with announcements about genetic discoveries, we typically have no reason to claim causation with as much enthusiasm and fanfare as many researchers are wont to do. But again, yes, it does make for simple storylines and wonderful press releases. I'm not sure why this basic point is difficult to acknowledge, as though there were some kind of knee-jerk reaction against anything that might in some slight, indirect, minimal sense cast even the smallest doubt on the party line. Why not just say, "Yes. You are right that we don't really know whether the allele is causing the physical result in question. But we do know it seems to be closely correlated and that is an interesting piece of data. Hopefully we'll be able to determine at some point what is really going on in the cell in this particular case." Is that a difficult thing to acknowledge? Is there so much baggage and pride invested in every little pronouncement by evolutionary biologists that heels must be dug in and claims must be supported at all costs? Eric Anderson

No. Like HIV causes AIDs and certain MC1R genotypes cause red hair. wd400

WD400: It’s perfectly obvious these alleles are causative for any useful definition of the word.

Just as 'perfectly obvious' as the muscles in my hands causes these sentences or just as 'perfectly obvious' as the stop sign causes the car to stop in Eric's example. Box

I would hope folks with "half a brain" could read the post. It's perfectly obvious these alleles are causative for any useful definition of the word. wd400

It's ironic that wd400 calls Eric's writings "boring" and "dumb", while it's obvious to anyone with half a brain that he is unable to understand Eric's elementary point: correlation doesn't imply causation. Box

This seems a lot like some kind of boring philosophical discussion. In these cases an allele "A" skews the outcomes of meosis in it's own favour. Another allele "B" doesn't. If you have "A" you get the skewed ratios, if you have "B" you don't. That's causation. You don't have to explain in through sub-atomic particles to know that. This paragraph

On the other hand, we could say, “Look, a “selfish” gene that is “causing” this to happen. Mystery solved. No need to look further into the actual causes, because, hey — look! — a gene that happens to be correlated with the physical outcome. It must be the cause.”

...is just dumb, evolutionary biologists have lots of work to understand how these and other selfish genes work at a molecular level. Knowing how they work in indeed interesting, but it's not going to make them any more or less selfish. wd400

wd400:

It’s like saying does baking powder or heat drive a cake’s rise.

You are using an example of two known factors that contribute to a cake rising. It is known based on independent factors and independent of this particular cake. Associated alleles, at least in most genetic studies, are not like that at all. Now maybe in this particular case you cited someone actually knows what the allele is doing, how it interacts on a molecular level, what specific changes it produces to cause the effect that is seen. If so, great, and I would be very interested in that information. But in the vast majority of cases in genetic research we are dealing with "associated variants" or alleles that happen to be correlated with, or alleles that are in proximity with the physical characteristic in question. That is often the best we can do with our current state of knowledge and I am not criticizing efforts to understand the genome using such tools, but the fact that some allele is associated with a particular physical state does not mean that the allele caused the state in any real sense, nor does it mean we are dealing with a "selfish" gene, whatever that may mean.

It’s clear that these alleles cause skew transmission ratios.

No. It is clear that these alleles are associated with the skewed transmission ratios. And it is possible they might be causing it in some meaningful sense, but that is not clear from the mere association. Let me give you an example: A few months ago I had a chance to ride in one of Google's prototype self-driving cars (Lexus RX version). Someone watching from the outside and not knowing much about how things work might observe that every time the car approaches an octagonal sign with the word "STOP" on it, the car slows down and stops. The observer might then proclaim that the stop sign causes the car to stop. But we know that is not really true. What is actually going on is that there is an elaborate system of sensors, controls, analyses, programs and feedbacks that perform various actions toward a pre-specified goal when they sense the presence of the stop sign. Yes, the stop sign is a trigger for a particular sensor that has been designed to detect that trigger and then initiate a particular physical outcome in concert with several other systems. But the stop sign did not cause the car to slow and then stop. With regard to a particular outcome in the cell, what I would like to know is what is actually going on. What is actually doing work. What is actually occurring to cause a particular physical effect in a cell. Things don't just happen, particularly when most alleles are supposedly sitting there in a DNA sequence largely as passive participants in a dance as complex as meiosis. Something is sensing the allele and making a change, or responding, or interacting in a different way. I think that is fascinating and it would be very good to understand. On the other hand, we could say, "Look, a "selfish" gene that is "causing" this to happen. Mystery solved. No need to look further into the actual causes, because, hey -- look! -- a gene that happens to be correlated with the physical outcome. It must be the cause." This kind of thinking is rampant, perhaps not so much in the literature, but at least in the public press reports of genetic discoveries: claims about the discovery of this or that "gene for [insert disease here]." It is too facile, it is too simplistic, it doesn't seek to examine the real interactions that are occurring. But it does make for a simple story and for wonderful press coverage. Eric Anderson

I think you are confused as to what "driving" means here. Driving alleles are those gene variants that get themselves over-represented in the products of meiosis. In the Mimulus case it is a centromeric gene variant (allele) that cheats at meosis. In wheat it's centromere-like sequences, there are many more cases about which I don't know the details. Asking which genes "drive" as in "control" a process like meosis is pretty pointless. It's like saying does baking powder or heat drive a cake's rise. What we can talk about is how variants of the players in these systems change the outcome. Adding more baking powder makes cakes rise more. Adding more repeats to the driving allele discussed in this paper distorts products of meosis in favour of that allele. I'm note sure what your point about correlation and causation is meant to add up to. It's clear that these alleles cause skew transmission ratios. wd400

Thanks, wd400. I couldn't access the article behind the paywall, but from the abstract it looks interesting. It appears to be a case in which the authors refer to the centromeres as meiotic drivers, rather than a particular gene. Presumably they aren't suggesting that the fact a particular gene may be "associated with" or "connected to" a specific physical process (in this case a preferential transmission across generations) that this gene is in fact itself driving the process. Given that we are barely scratching the surface at understanding cellular mechanisms generally and the genome specifically, I appreciate that genetics studies are often reduced to identifying associations of alleles with physical traits, rather than actually understanding what is going on at a deep level to produce the trait. It is a crude and simplistic tool, but unfortunately it is about the best we can do in most cases at present. However, we need to be careful to not allow the old temptation of 'correlation equals causation' to creep in. Eric Anderson

Meiosis is just another biological process that unguided evolution cannot explain. Joe

"It’s hard to imagine a more clear-cut case of a selfish gene than a metotic driver" I assume Dawkins devotes a chapter or two in "The Selfish Gene" to these bad boys? Haven't read that book, seems it would be painfully outdated by now. ppolish

We don't know in this case, but in other female drive alleles effectively hijack the process that ensures centromeres line up in Meosis I. The fact this allele is a copy number expansion of the wild type suggests this is something similar (and certainly something structural). (It is also indeed "the gene itself" that is controlling the products of meiosis. One allele distorts the process, the others don't) Really understanding the molecular basis of driving alleles requires a good understanding of meiosis in the system. wd400

wd400, Thanks for the thoughts. Interesting. How would the gene cheat at meiosis? Does the gene itself control whether it is passed on, or is there something in the process of meiosis that determines which genes get passed on? It seems that throughout the process the gene is being acted upon, not doing the acting itself? Eric Anderson

(a) Meotic drive is well enough known that it is described in the glossary of an intro to Evolution text book. This new paper on one particular driving allele is interesting, but that some alleles don't follow Mendel's law is not at all new (b) It's hard to imagine a more clear-cut case of a selfish gene that a metotic driver. These alleles spread in a population not because they do anything for their hosts (in fact, that can be deleterious to the host/population) but because they cheat at meosis. Laurence Hurst has a discussion on this here (not sure if that link is paywalled...) wd400