Talk:Genetic drift

Untitled

See also: Talk:Evolution/Genetic drift Genetic drift looked at in the sense of modern day public policies we will find displaced genetics in new environments. Meanwhile, natural disasters also play a major role in modern day genetic drift. For example, communities must have low income housing making all neighborhoods accessible to all and not exclusive to one ethnicity or culture. This leads to resulting in procreation between genetics that may have never crossed paths, or in general, outside their normal community. As mentioned, natural disasters will displace individuals as well. In our modern time we have lived through wildfires, hurricanes and currently flooding and overall sinking in certain regions of the world. This uncontrollable events also lead people to migrate and settle in new environments. Adapting to new ways of life and ultimately new people. Ralle034 (talk) 18:44, 10 February 2022 (UTC)

Wiki Education Foundation-supported course assignment

This article was the subject of a Wiki Education Foundation-supported course assignment, between 31 August 2020 and 18 December 2020. Further details are available on the course page. Student editor(s): Krsmith09.

Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 22:11, 16 January 2022 (UTC)

Random sampling

Random sampling is clicked and it sends you to the statistical practice. —Preceding unsigned comment added by 98.149.131.179 (talk) 23:02, 18 April 2010 (UTC)

I have disliked the use of random sampling here for a long time on pedagogical grounds. Now found a factual argument that I wasn't aware of and don't quite understand. The process of change in allele frequency due solely to chance effects is called random genetic drift. One should note, however, that random genetic drift can also be caused by processes other than the sampling of gametes. For example, stochastic changes in selection intensity can also bring about random changes in allele frequencies. Li & Graur, Fundamentals of Molecular Evolution, 1991, page 28. --Ettrig (talk) 20:31, 19 February 2009 (UTC)

Theoretical Population Genetics by JS Gale (1990) p 13 uses two different terms for these two processes: "random genetic drift" and "random fluctuation in intensity of selection", but characterizes both of them as stochastic processes. Professor marginalia (talk) 20:46, 19 February 2009 (UTC)

Gale calls "random genetic drift" "random genetic drift" and "stochastic changes in selection intensity" "random fluctuation in intensity of selection". So ... ? --Ettrig (talk) 21:24, 19 February 2009 (UTC)

No, I meant he says there are (at least) two types of stochastic processes, random genetic drift and random fluctuation in intensity of selection--random genetic drift by Gale has the traditional meaning, ie random sampling of alleles. But I found another source that treats it differently, I'm skimming it now. Basically this one wants both to be considered genetic drift, along with other stochastic processes. I'll give its synopsis asap, but this looks like there is no "one view" of this yet. It may require us to describe and attribute all representative views on this. It's clear even in the source I'm reading that genetic drift has a traditional definition (random sample error), and this book is pushing for a wider definition. More soon. Professor marginalia (talk) 21:48, 19 February 2009 (UTC)

OK, here it is. Another one, by Kimura (who I believe is one at the forefront of a new "revival" of drift in contemporary population genetics), Population Genetics, Molecular Evolution, and the Neutral Theory (1994) fills out some of the missing pieces. The discussion requires some development here, I think, especially with better integration with the historical development and status of drift theory in the overall context of traditional evolutionary thought. In terms of evolutionary biology, I don't believe his views have taken firm hold yet, but his work is definitely influential, drawing a lot of interest and attention. In evolutionary thought you find a solid core of "selectionists" who will allow that drift plays only a minor role in evolution-and they've traditionally held to the "slow, gradual change guided by natural selection" model. This view has been strong going all the way back to the early days of the modern synthesis, with probably R. A. Fisher its greatest influence. Then you find a smaller number coming from different angles who push against the idea this is the best "standard model". What Kimura puts forth in this paper about drift is an argument of sorts to broaden the scope and consider how random processes factor beyond sampling error in small populations. The traditional meaning of genetic drift was based on Wright's work and correlates to random sampling. This Kimura calls "the narrow view". He also disputes the traditional view that drift has little significance outside of small populations. We really need more sources on this - but it would be good to cover the subject. I'll see what else I can pull together. Professor marginalia (talk) 22:27, 19 February 2009 (UTC)

Maybe we should call it drift by random natural selection?  ;-) --Ettrig (talk) 10:13, 21 February 2009 (UTC)

Modern mathematical models of drift descend mostly from Kimura's work, which is entirely based on drift as a sampling phenomenon. Other usages might exist but I think they're marginal at best. Graft | talk 23:38, 16 March 2009 (UTC)

I've removed the link to the statistics page. Statistics concerns a method of measurement: "The mathematics of the collection, organization, and interpretation of numerical data", so, unless genes "collect, organize, and interpret numerical data" and this activity then motivates them to act in a certain manner, whatever is being attempted to be explained in the introduction here it seems clear that it is not the statistical technique the link points to. I would like to add here that imo the introduction needs to be rewritten. It oscillates between jargon and kindergarten levels of prose and seems to me to be both clumsy and opaque, even if, in some sense, it might be considered accurate. It is also repetitive. LookingGlass (talk) 07:18, 6 March 2014 (UTC)

To Do

• Generalize example to differing initial frequencies
• Generalize example to combined selection/drift
• Formula for probability of drift
• Formula for expected (magnitude) of drift
• Role in extinction, fixation, substitution
• Probability of substitution
• Balance between mutation influx and extinction/substitution
• Neutral evolution
• Drift -> divergence -> finds new points where selection exists
• Gene duplication -> more drift -> new function
• Loss of unused traits (example: sight in dark cave)
• Diffusion models of genetic drift

Lead

I reworded the lead. It's still a little clumsy, but is better. Are there any defects? I'm thinking of rewriting the second paragraph to briefly explain the "four" in "genetic drift is one of the four processes" (mutation, drift, flow, natsel). If that can't be managed in the lead, it might be better to omit "four". Does anyone know of a good (simple) reference so we could just link it? There should be something in WP, but I can't find it. --Johnuniq (talk) 10:52, 24 February 2009 (UTC)

I would rather remove "four" than add the list. The "four" was there because of a concern earlier that we give the impression that there are only two subprocesses. --Ettrig (talk) 11:21, 24 February 2009 (UTC)

The lead sentence in Population genetics spells it out but it's not sourced there either. I can find a source though--that won't be a problem. I think for awhile the lead contrasted genetic drift (random change) to natural selection (adaptive evolution). Then later it was changed to say there were two evolutionary mechanisms. I think the natural selection comparison should be in the lead, and it shouldn't say there are two when there are four. The article should talk about some of the models involving more than one mechanism along with genetic drift; for example, dynamic models with a mix of mutation, drift and gene flow. Absent some discussion of the mix, there's not much to be said about all of the mechanisms in the lead. These sentences bother me here, "The inter-generational changes are usually small, and genetic drift is usually slow, but in the long run drift can account for a large proportion of the changes in a gene pool. Genetic drift reduces genetic variability, and may cause gene variants to disappear completely." What are "small changes"? What is meant by "slow changes" and "long run"? Maybe it's all relative, but to me the changes look very sweeping and very rapid in genetic drift. The mounting of slow change over a long time period is the classic natural selection model, not classic drift. What are we saying there? Professor marginalia (talk) 15:01, 24 February 2009 (UTC)

I would change to "Genetic drift can account for a large proportion of the changes in a gene pool. It commonly causes gene variants to disappear completely and thereby reduces genetic variability." Please suggest or add a formulation about combined effects. --Ettrig (talk) 16:19, 24 February 2009 (UTC)

I like keeping "four processes" as a link to Population genetics (thanks for the link). I take the point that "small changes", "slow changes", "long run" are rather weasely, and I can't offer an authoritative opinion re alternative wording. Do you know of any clear, sourced statements that shed light on this? It seems obvious that in anything more than a population of 100 individuals, inter-generational changes due to drift will be "small", but I take the point that by evolutionary standards, an allele could be lost from a small population quite quickly due to drift. I haven't seen much clear discussion on the relative significance of the four processes, and I wouldn't support switching from saying that drift gives small/slow changes, to the opposite of making the unqualified statement that it can account for a large proportion of changes, without a source. It seems very plausible that some species would have been strongly influenced by drift, or even that nearly all species have been significantly influenced. But I think we need an accepted source before possibly giving readers the idea that a large proportion of evolution is due to genetic drift (was a large proportion of the evolution of the eye due to drift?). --Johnuniq (talk) 00:57, 25 February 2009 (UTC)

What it boils down to is population size. The smaller the population, the bigger and faster the changes can be. Large populations don't budge much at all from drift, and most consider drift relevant only to small populations (one source defined n<200). Ettrig may prefer to say drift is smaller and slower in large populations. Many sources I've seen treat drift as so unlikely in a large population, the odds so small it would occur in them, that it's not considered to have much relevance there. These are largely probabilistic conclusions; there isn't much in the way of real world verification about what kinds of phenotypic evolutionary changes (such as the development of the eye) can be attributed to any one kind of mechanism. But there is so much written about the issues they're looking at in evolutionary models of population change. It would be very interesting, I agree, to include some of the issues and dilemmas in the article here. Professor marginalia (talk) 08:02, 25 February 2009 (UTC)

as a casual reader just looking for a quick understanding of the concept of genetic drift, i hit a brick wall before i even finished the first sentence, i got the hell out of here before wasting another minute and found a perfect little page via Google, that explained it to me, quick and dirty, with only a 10 second investment in time; often, Wikipedia articles these days seem so incredibly opaque and poorly written: simple and clear sentences are rejected in favor of complex and ambiguous ones; well, there's my two cents, don't go spending it all in one place, cheers — Preceding unsigned comment added by 69.250.176.112 (talk) 23:27, 30 June 2011 (UTC)

I fully agree with the above of 30 June, 2011. The lede does not meet Wiki's standards. One of the obvious problems is the use of required hypertexting without a clue of it's meaning within the sentence (lazy hypertexting). Certainly one should not be required to begin the treacherous and potentially endless journey to other articles also dependent on lazy hypertexting, those themselves also dependent on lazy hypertexting... — within the opening paragraph! (There lies a systemic Wikipedia fault.) See also: Wikipedia:Manual of Style (lead section), in particular: "The lead section should briefly summarize the most important points covered in an article in such a way that it can stand on its own as a concise version of the article," and "In general, specialized terminology and symbols should be avoided in an introduction."

Notice how the definition is handed off to hypertexed jargon, thus is NEVER defined here: Quoting first paragraph:

"Genetic drift or allelic drift is the change in the frequency of a gene variant (allele) in a population due to random sampling.[1] The alleles in the offspring are a sample of those in the parents, and chance has a role in determining whether a given individual survives and reproduces. A population's allele frequency is the fraction of the copies of one gene that share a particular form.[2] Genetic drift may cause gene variants to disappear completely and thereby reduce genetic variation."

...And are we really to believe that Genetic drift is caused by random sampling? And so forth. We seemingly need fewer experts and more teachers/communicators here.
--69.110.91.50 (talk) 22:37, 18 December 2011 (UTC)Doug Bashford

frequency

relative frequency is probably just a halfway transition from relative abundance to frequency. That is trivial. But my real complaint is about the use of frequency in the first sentence. Allele frequency is explained in that article. But not even frequency (disambiguation) has the meaning used in this context. The closest thing is number of times of occurrence in an experiment. So we cannot expect frequency to be correctly understood by a large majority of the readers. Maybe relative abundance is too cumbersome. But frequency just will not do. --Ettrig (talk) 16:15, 5 March 2009 (UTC)

I don't understand your logic. Abundance isn't going to give any useful meaning for readers trying to understand "relative abundance" either. And I suspect if readers don't understand "relative frequency", they're not going to understand "relative abundance" either. We're bouncing all over the landscape with these edits and reverts. We need to stop throwing spaghetti at the wall and tackle this article like wikipedians. We need to sketch out some basics so we're more on the same page here. A) who are the readers, and what kind of base knowledge do we expect them possess to read this? B) what do the sources say about this topic and how do they phrase it? And C) how do we transition from the historical or mainstream definitions to the more notable "complexions" of it to come along since the early classical work by Wright, et al? Professor marginalia (talk) 17:18, 5 March 2009 (UTC)

If the problem is my use of abundance, I agree, already stated above that it probably isn't a good choice of word. But the criticism of the use of frequency, I still find crystal clear. In the first sentence in this article it is used to refer to a concept which is far from the primary association a non-biologist would make. Please check Wikipedia and Wiktionary. Partial answers: (A) The first sentence should be understandable (accessible) by any normal reader. (B) They normally use frequency. This fact has very low weight in our choice of words in the first sentences of the article. (C) Is misleading. We should describe the present views as well as possible. We should not emphasize or give priority to old views.

The link to Sampling (statistics) also leads the reader astray. Populations do not need to be part of statistical practice concerned with the selection of individual observations intended to yield some knowledge to be subject to drift. I suggest ...is purely random change in proportions of gene variants (alleles) in a population. ... --Ettrig (talk) 20:39, 5 March 2009 (UTC)

I think we risk watering down the description to be vague enough to allow for everybody's strict use of the term, which won't help a reader at all. I think we need to rely on sources here. We don't need to reinvent the wheel. I will come back with a list of definitions from sources and we can discuss those. Professor marginalia (talk) 22:08, 5 March 2009 (UTC)

Futuyma, Evolutionary Biology, 1998, Glossary, defines frequency: In this book, usually used to mean ... the proportion of gene copies belonging to that allelic state --Ettrig (talk) 22:10, 5 March 2009 (UTC)

I've put together a list to look at here. Almost everyone uses "frequency", so I don't think there's a need to avoid it here. Professor marginalia (talk) 23:31, 5 March 2009 (UTC)

This "answer" does not relate to the problem I am bringing up. I have pointed out that the meaning of frequency as used here is a special one. Therefore it is not in accordance with Wikipedia guideline Wikipedia:Lead section, which says specialized terminology should be avoided in an introduction. --Ettrig (talk) 05:42, 6 March 2009 (UTC)

/sigh/ It is not a "special" definition. As per dictionary, frequency can be defined as "rate of occurrence", which fits well I think. Professor marginalia (talk) 07:03, 6 March 2009 (UTC)

ref num	Websters	Penguin	Oxford	Wiktionary
	1979	1969	1963	Current
1	(originally) condition of being crowded, a crowd
2	the fact of occurring often or repeatedly, frequent occurrence	continual repetition	frequent occurrence	the property of occurring often rather than infrequently
3	the number of times any action or occurrence is repeated in a given period		number of repetitions (in a given time)	the rate of occurrence of anything; the relationship between incidence and time period
4	in mathematics and statistics, (a) the ratio of the number of actual occurrences to the number of possible occurrences in a given period; (b) the ratio of the number of individuals occurring in a specific class to the total number of individuals under survey	(statistics) the number of observations having a value between two specified limits
5	in physics, (a) the number of vibrations or cycles per unit of time; (b) the number of cycles per second of an alternating electric current	number of vibrations (of wave) per second	number of repetitions (in a given time)	the quotient of the number of times n a periodic phenomenon occurs over the time t in which it occurs: f = n / t

The table above shows some dictionary definitions. The sense used in this article is Webster 4 (b). This sense is a technical term used in mathematics and statistics. It is not even given as a biology concept. The fact that it is used is natural, given the close relation between population genetics and statistics and probability. This sense is missing in the other dictionaries, possibly very peculiarly expressed in Penguin. Note also how all the other senses (disregarding obsolete number 1) are related to time, which the use in this article is not. The conclusion is clearly that frequency with this meaning cannot be used without explanation in the first sentence in a Wikipedia article. --Ettrig (talk) 18:51, 8 March 2009 (UTC)

The definitions from a survey of sources using the word that I linked earlier include two general encyclopedias and an introductory level textbook. Here I've bulleted them to show which they are. Dictionary.com has seven dictionaries represented. The best definitions for what is meant here are represented as follows:

1. "rate of occurrence: The doctor has increased the frequency of his visits"
   
2. "the ratio of the number of observations in a statistical category to the total number of observations "
   
3. "the number of individuals in a single class when objects are classified according to variations in a set of one or more specified attributes"
   
4. "the ratio of the number of occurrences of some event to the number of opportunities for its occurrence"
   

I think it is clearer to say "relative frequency", which dictionary.com defines as, "the ratio of the number of times an event occurs to the number of occasions on which it might occur in the same period". My conclusion would be that "frequency" is fine. Professor marginalia (talk) 21:07, 8 March 2009 (UTC)

The word frequency has several meanings although some of them have considerable similarities. The meaning used in the current version of the article is not misleading. On the other hand this is not the way frequency is normally used in population genetics. The first problem is in the statement "The best definitions for what is meant here are represented as follows...". This statement shows an unwillingness to make the distinctions at all. The argument against this view is that the dictionaries quoted above show different meanings. I am not sure what dictionary.com means, but surely it cannot be interpreted to say that these are the same meanings. The second problem is what meaning is normally used in evolutionary biology and population genetics. I already noted a preference for Webster 4 (b) "the ratio of the number of individuals occurring in a specific class to the total number of individuals under survey". This is very similar to dictionary.com 4, although that is more like Webster 4 (a) with its explicit reference to time. I start with Evolution by Strickberger, third edition, 2000, because this is referenced in here, number 14. In the Glossary it defines Random Genetic Drift as "... random change in frequency of alleles ...". When introducing allele frequencies this book says, p 519, "Gene frequencies are simply the proportion of the different alleles of a gene in a population." Note here that these are not occurrences spread out over a period of time, but the objects that exist at one particular time. The need to use "relative" to qualify "frequency" in the current version of our article is telling. Note that this is not needed with the definition in Strickberger. Neither is "relative" used together with "frequency" in any of the 12 definitions in the list linked above. The book Principles of Population Genetics, by Hartl & Clark, 2007, says on p 48 "Variation in allele frequency owing to sampling error in small populations is called random geneic drift." On page 19 it says "By the allele frequency of a specified allele, we mean the proportion of all alleles of the gene that are of the specified type. Again, note the absence of reference to time and the non-existent need to qualify with the word relative. The relativity is built-in, so to speak. The third problem is whether this use of frequency is so special and technical that it needs introduction before use. Note that the two books just referenced do have such careful introductions/definitions. Likewise, the quotations from Encarta and Britannica (for a general public) do not base the definitions on frequency. To summarize: dictionaries say frequency has several meanings; the meaning used in our article is not the one normally used in definitions of genetic drift; texts that introduce genetic drift to a general public either avoid frequency or introduce frequency before using it. --Ettrig (talk) 20:07, 15 March 2009 (UTC)

You've completely lost me at this point. You seem to be finding "distinctions" between definitions that I simply don't see. English isn't mathematics, and definitions of words aren't precise equalities. Can you spell out in a single sentence why you think "frequency" is a problem? Thanks. Professor marginalia (talk) 20:46, 15 March 2009 (UTC)

The meaning of the word frequency that is used in population genetic definitions of genetic drift is not one of the commonly understood meanings. We need to use other words to introduce genetic drift, then introduce allele frequency with the right meaning, to prepare the reader for what follows in this article and for other, more specialized, texts about genetic drift. I consider this discussion to be rather linguistic than mathematic. There is a very important linguistic aspect of writing an encyclopedia.--Ettrig (talk) 21:19, 15 March 2009 (UTC)

And two more questions. How is "frequency" defined in population genetics? And how is it different from "relative frequency" as it is typically used in normal English? Professor marginalia (talk) 21:43, 15 March 2009 (UTC)

The detailed argument about this is stated above, based on reliable sources. --Ettrig (talk) 06:01, 16 March 2009 (UTC)

You haven't made it clear. Proportion, ratio, frequency are all the same. None of these are "technical terms" in English. The use of the word "frequency" in this article is exactly the same as its usage in population genetics. How is "time" a problem with any of these definitions? Professor marginalia (talk) 14:10, 16 March 2009 (UTC)

You are expressing your personal view and posing a question? I have shown above, based on reliable sources that that view is not correct. I have also answered the question. --Ettrig (talk) 15:09, 16 March 2009 (UTC)

Ok. I'll just say what I think, which I think is also supported by the reliable sources. I think the lead should just spit it out, get straight to the point, and say something like this: "Genetic drift is the evolutionary change an the allele frequency (or gene frequency) of a population as a consequence of the random factors which influence the survival and reproduction of its members." Professor marginalia (talk) 16:19, 16 March 2009 (UTC)

We should absolutely say frequency in the first sentence. "Allele frequency" is an indispensable term when talking about drift, and the sooner the reader is introduced to this usage, the better. What's wrong with the current formulation? It seems relatively unambiguous about what "frequency" means here - what fraction of the population has a particular allele. Is the reader going to get confused and think we're talking about cycles or something? We're not supposed to assume the reader is an idiot, here. Graft | talk 23:51, 16 March 2009 (UTC)

Explanation

The first sentence in the main text has the kind of explanation that I am asking for. A population's allele frequency is the fraction of the population's genes having a given allele variant. Unfortunately this formulation runs a high risk of being misinterpreted. The word population most often is interpreted as all the organisms somewhere. Here it refers to a set of gene copies. So if the allele is A, the only other allele is B, there are 5 organism individuals, one AA, two AB, two BB. Then the allele frequency of A is 40% with the correct definition. With the current definition it could very easily be interpreted as being 60%, three out of five organism units have allele A. --Ettrig (talk) 06:57, 22 March 2009 (UTC)

Yes, you have a good point. What about, "An allele frequency is a measurement of the frequency of a particular allele relative to the total number of alleles at that locus in a population." The misunderstanding is compounded by the "for example" saying a frequency can't remain 50% going from an even numbered population to an odd numbered population, because of course even if population n is an odd number, 2n is an even number and you can still arrive at a clean 50% allele frequency mathematically. I don't like that example for other reasons also. Professor marginalia (talk) 17:24, 22 March 2009 (UTC)

I have argued at an enourmous length that frequency as used in population genetics is a technical concept that needs to be explained. In view, so is allele. So I would like the explanation not to use these words. "Genes often occur in variants, called alleles, with slightly different function. The fraction of all the gene copies that is made up of a particular variant is called its allele frequency."--Ettrig (talk) 17:45, 22 March 2009 (UTC)

It was "population" that led to the possible misunderstanding, not "frequency". Alleles are not required to have any function at all, and sometimes different alleles will share a given function in common. Some of the most valuable findings come from analysis of those allele differences that are thought to have no associated function. Why are we throwing this in? Professor marginalia (talk) 17:54, 22 March 2009 (UTC)

Agreed on different function. This is not part of the concept of allele. (But it was for many decades.) Frequency is currently explained. You suggest that the explanation is removed. I ask that it is kept. I have argued extensively for this above, showing that textbook authors do. I have seen no valid counterargument. --Ettrig (talk) 21:43, 23 March 2009 (UTC)

Here I quote again from the Dictionary of Science, "gene frequency-the relative frequency with which a particular gene is present in a particular population of a species or other group". Professor marginalia (talk) 23:21, 23 March 2009 (UTC)

This is a specialist topic. Requiring readers to understand what 'allele' is before they read it is not a high bar. It's linked in the first sentence. It's used throughout the article. If you don't know what 'allele' means, you shouldn't be reading this thing. Your habit of wanting to "explain" by removing indispensable terminology seems wrong to me. Graft | talk 18:15, 23 March 2009 (UTC)

We have different opinions on what should be explained. In the case of frequency I have backed mine by pointing out that textbooks do. That is, I have presented a valid Wikipedia argument. You have not. Specialist topic is definitely not a valid Wikipedia argument. I have pointed out the Wikipedia standard of accessibility above. I did not remove the word allele earlier, except for the very first sentences. My line is that the first 1-3 sentences should be without allele, frequency and sampling. Then these words should be explained, then used. --Ettrig (talk) 21:43, 23 March 2009 (UTC)

I agree with graft. If a reader doesn't know what an allele is, there is no reason to explain genetic drift in "simple" terms to them. "Allele" isn't the hitch, "frequency" isn't the hitch. Genetic drift is a particular evolutionary mechanism-are we going to explain evolution? No, this isn't an introductory class in biology and genetics, it's an article about drift. So what is drift? It's a purely random change in allele frequency. It's very different than "natural selection". What do we mean "random"? In genetic drift, it's very different than "random mutation". In drift, it's the "random sampling" of existing alleles during reproduction. As in random mating within the population, random environmental episodes which impact individual survival and reproduction, random isolation/removal from the ancestral population, etc. Who "discovered" it? How do they measure/estimate it? What impact can/does drift have in evolution? What are some real examples of it? How are drift models used in research in population genetics? That's it! It's about drift. Drift. Drift. I think we should leave off worrying about explaining what an allele is to the allele article, and let's focus on explaining drift as best we can. For further support, I offer yet another source-The American Heritage Dictionary of Science, touted on the flap as "bridging the gap between overly simplified and the highly technical": "the fluctuation in gene frequency occurring in an isolated population, presumably due to random variations from generation to generation." There is no term more fitting than "frequency", if for no other reason than that's what it is and what it's called. Professor marginalia (talk) 22:17, 23 March 2009 (UTC)

It's hard to get the words right. What about replacing the whole paragraph with:

In a population of organisms, several different alleles may occur at a particular locus. Considering one of the alleles, its allele frequency is the fraction of the number of occurrences of the allele, relative to the total number of all alleles at the same locus. Drift is the accumulation of chance factors that affect allele frequency: variations that result from random sampling.

That removes the mention of odd/even numbers. Johnuniq (talk) 00:28, 23 March 2009 (UTC)

I know-we are all challenged getting the words right. I like your idea of giving a "for instance", but it's better to remove this one since we know there are problems with it. Professor marginalia (talk) 22:20, 23 March 2009 (UTC)

Futuyma

Ettrig-can you clarify that the glossary definition of allele/gene frequency requires 2 or more allele variants in the population to qualify under the definition of allele frequency? It's quite possible to have an allele frequency of 1 or 100% (allelic homogeneity), isn't it? The case in which there is just one variant of the allele in the population? Thanks. Professor marginalia (talk) 17:22, 28 March 2009 (UTC)

Hmm... Futuyma says several. Not in allele frequency but in immediately preceding allele. I modified slightly to two or more. Are you challenging this or requesting that it is expressed more clearly? --Ettrig (talk) 17:43, 28 March 2009 (UTC)

Maybe it's expressed unclearly, otherwise I don't understand this definition at all. Can you quote it verbatim? The definition seems to be requiring 2 or more allele variants, and I don't think that's a meaningful requirement. Populations can and do have allele frequencies of 100% for certain alleles. There are not 2 or more variants in those cases. Especially pronounced when we're talking about genetic drift in small populations, there is a random tendency toward homogeneity of alleles--and at this point the population no longer has a measured frequency for the alleles? I can't envision any potential value in drawing a definitional line requiring there be two or more allele variants.Professor marginalia (talk) 18:06, 28 March 2009 (UTC)

Definitions of allele: (1) One of several forms ... (2) An alternative form ... (3) (alleles) alternative forms ... (4) One of the different homologous forms ... (5) Any of the alternative states ... The first one is Futuyma. If we go by definitions given, the pattern is clear. Alleles are supposed to occur in plural. But I don't think anyone would protest if an allele frequency reached 100%. Hartl & Clark says on p. 96 Eventually the population will have either all A alleles or all a alleles. Here they are talking about a situation with only one allele, as you expected. --Ettrig (talk) 18:39, 28 March 2009 (UTC)

While you can certainly have an allele frequency of 1, that's sort of irrelevant in the context of drift - you can't have drift if a single allele has fixed, after all. Graft | talk 00:07, 29 March 2009 (UTC)

You say (1) requiring 2 or more allele variants is not a meaningful requirement and (2) allele frequency of 1 is irrelevant. These two comments point in opposite directions. Maybe it would help you to observe that the quote from Hartl is about the transition from a situation with two alleles to a situation with "one". --Ettrig (talk) 06:19, 29 March 2009 (UTC)

Okay, maybe it will be less confusing if I restate this. An allele frequency can be 1 or 0. Genetic drift for Allele A cannot occur if Allele A is 1 or 0. Both true. Allele frequencies of 1 are relevant to genetic drift because in an isolated population, genetic drift has a tendency to eventually push an allele frequency to 1 (or 0). Once it gets there, no drift possible. Whole point being, let's not imply that the definition of "allele frequency" applies only to alleles in populations where there are multiple allele variants. We can say genetic drift is actively occurring only when there are multiple allele variants. But it's a mistake to say that's what the term "allele frequency" itself means. Professor marginalia (talk) 19:43, 29 March 2009 (UTC)

Pedantic question? biological example

In the bacteria example given, the population is allowed to breed for many generations and then the surviving organisms are examined to determine frequencies of A and B. The probabilities for those frequencies are calculated. But can we say the probabilities remain consistent for x, x+1,x+2...x+? generations? Is this accurate? Given that the probabilities are memory-less between the begin point and the end over a span of multiple generations, and given that absorbing boundaries will usually cause alleles to disappear completely in the end, I can't see how these tendencies are consistent with unchanging probability (as given in biological example section) of a fixed allele, 1/16. Doesn't Markov say the probability of fixity in this case is the same as its freq distribution in the original population? Or 50%?Professor marginalia (talk) 22:40, 1 April 2009 (UTC)

Yes, this is an unrealistic assumption made to provide a very simple model. The assumption is that all the bacteria divide at the same time and that none die in this period. This assumption is very similar to the one made in the Fisher-Wright model, that a very big and perfectly representative set of gametes is generated. --Ettrig (talk) 06:53, 2 April 2009 (UTC)

What you're saying in this example is that all bacteria reproduce exactly equally, as do all their offspring, all of them surviving until - poof - they are all simultaneously killed, en masse, except 4? There is really isn't any genetic drift to speak of occurring yet in this example then-we've described a catastrophe and the 1/16 odds all who survive it share the same allele. This is a population bottleneck, but unless and until these 4 survivors (the random sample) breed, there is no real process underway showing drift from the sampling. Professor marginalia (talk) 22:40, 2 April 2009 (UTC)

Founder effect

The current text on founder effect confuses founder effect with bottleneck. Founder effect occurs only at the moment of founding, when the founders are sampled from the population of origin. The current text erroneously also includes the drift that occurs after the founding. Stearns and Hoekstra, p 65, describes founder effect with New populations are sometimes founded by a small group of individuals in which gene frequencies differ considerably from the frequencies in the parent population. Glossary of Futuyma The principle that the founders of a new colony carry only a fraction of the total genetic variation in the source population. Glossary of Strickberger ... Since these founders carry only a small fraction of the parental population's genetic variability, radically different gene frequencies can become established in the new colony. I left some text that contradicts this, because it was supported with sources that I cannot check. Please, could somebody else check those? --Ettrig (talk) 19:43, 13 May 2009 (UTC)

The section needs to give the definition of founder effect but also to describe the after-effects from an evolutionary perspective. The multi-generational consequence, the effect, these founding gene frequencies have upon the population's later gene frequencies is what is of interest to biologists and population geneticists.

Which sources do you want checked? Professor marginalia (talk) 20:38, 13 May 2009 (UTC)

Faulty Probability

The example about 100 bacteria gives incorrect probabilities. The odds of getting 4 A's is not 1/2*1/2*1/2*1/2=1/16, but rather 25/50*24/49*23/48*22/47=303600/5527200=253/4606 All the odds are

• 4A's= 253/4606
• 3A's= 575/2303
• 2A's= 900/2303
• 1 A = 575/2303
• 0A's= 253/4606

JedG (talk) 05:03, 26 May 2009 (UTC)

Agreed, it is faulty. Given that one of the expected 50 As have been identified, the expected remainder is 49. My intention when introducing this example was to be able to show drift with as simple calculations as possible. In somewhat other words, a situation that is possible, that exemplifies drift and that can be analyzed with simple calculations. The first version did not mention the peak population size. Would it be OK to state that the population rises to enourmous amounts, then falls back to 4? --Ettrig (talk) 07:42, 26 May 2009 (UTC)

Perhaps the best thing to do would just say that there is a very large group to begin with. The larger the group the less the difference will be from the values given. No need to state actual numbers at the beginning. I am fine with using simple examples, as they are best for teaching, just make sure that you don't make any inaccurate statements. JedG (talk) 09:03, 26 May 2009 (UTC)

OK, I will make this change. Would appreciate if you keep keeping an eye on it. --Ettrig (talk) 10:27, 26 May 2009 (UTC)

I'm not sure what this example is meant to demonstrate; there's a lot of probability computation, but it doesn't clarify what exactly "drift" is in all those machinations - less calculation, more explanation is required, I think. Graft | talk 18:05, 27 May 2009 (UTC)

Questions

1. How would the marbles-in-a-jar analogy arise in the wild?

The real phenomenon that this is an analogy of is assumed to happen all the time. --Ettrig (talk) 18:44, 19 June 2009 (UTC)

2. Under "Time to fixation or loss", second paragraph says Time to Fixation is inversely proportional to population size. Tfixed equation says it's directly proportional. Which is correct?

The sentence in itself is inconsistent. According to my memory the equation is correct in this respect, so I will remove inversely. --Ettrig (talk) 18:44, 19 June 2009 (UTC)

3. The values for Tfixed and Tlost are the same when p = 0.5. How is this to be interpreted?

Drift toward frequency 0 has the same speed as drift toward 1. So if the starting frequency is exactly in the middle, the expected time to 0 is the same as the expected time to 1. --Ettrig (talk) 18:44, 19 June 2009 (UTC)

4. What is "mean average" in the para. under the Tfixed eqn? In statistics, mean and average are identical.

Trivial mistake. Expectation value is better concept in this context. Have changed. --Ettrig (talk) 19:10, 19 June 2009 (UTC)

5. The simulation results shown on the right are incomplete. According to the Tfixed eqn, the first should go beyond 55 generations to show fixation, the second should go beyond 550, and the third should go beyond 5500.

6. The progression in simulation results for the three population sizes suggests that genetic drift would not occur at all in large populations. Implications should be discussed in the text.

I have made the same observation. Have seen the same kind of graph in several textbooks. None other with the same "problem". I would like to put other graphs here. But in earlier discussions a strong view voiced was that drift would drow in other effects for large population sizes. So I got stuck. --Ettrig (talk) 19:10, 19 June 2009 (UTC)

7. The simulation results suggest that the genetics of a small, isolated population of Mountain Goats, say, should wander all over the place from one generation to the next. Has this been observed in the wild?

The effect is that many gene variant are lost, which is seen as small remaining amounts of gene variants. Yes, this is consistently observed. --Ettrig (talk) 19:10, 19 June 2009 (UTC)

8. The very last equation in the subsection gives values that are much greater than 1.0 (e.g., 10.8 for N = 20), indicating that mutations take much longer to be fixed and are therefore more likely to be lost. Is this consistent with other theories in genetics?

The mutation is almost always a small minority (1/N). Therefore it has low probability of being fixed by drift and the time for this to happen, when it happens, is long. Yes, I think this is the normal view of drift and I see not problem with it. If there is one, I would be interested to know about it. --Ettrig (talk) 19:16, 19 June 2009 (UTC)

9. The article is purely theoretical. Have examples of genetic drift been observed in the wild? Some sort of observational verification should be cited. Virgil H. Soule (talk) 18:03, 19 June 2009 (UTC)

It is very difficult to determine for a specific frequency change to what extent it was caused by drift and to what by selection. We "never" know all the selective forces. --Ettrig (talk) 19:10, 19 June 2009 (UTC)

10. "In probability theory, the law of large numbers predicts little change taking place over time when the population is large. " Is this statement accurate? Surely the relevance of the LLN is that fitnesses would converge on expected values, reducing the buffeting of drift on non-neutral alleles. If the neutral fixation rate is the mutation rate, then don't all populations change at the same rate (for neutral alleles)? I thought that was the point of the molecular clock? Allangmiller (talk) 00:58, 18 January 2012 (UTC)

I am also unhappy with this sentence. Probability theory does not say anything about genetic drift. Those conclusions have to be inferred or deduced. I think what is hinted at is that the sampling error (=genetic drift) is smaller for larger populations. This is important for each allele. Whether this leads to convergence depends on the combined effects of Genetic Drift, Mutation, Selection and what else. For a larger population, the same selective pressure is more likely to determine whether a particular allele will be fixed or not. --Ettrig (talk) 09:47, 19 January 2012 (UTC)

Fixed the sentence.Joannamasel (talk) 17:46, 21 January 2012 (UTC)

GA Review

This review is transcluded from Talk:Genetic drift/GA1. The edit link for this section can be used to add comments to the review.

Reviewer: Sasata (talk) 05:36, 29 October 2009 (UTC) Hi, I'll be doing this review. I'm going to try and dig up my old class notes on the topic to refresh myself with the subject material before digging into the article. In the meantime, could the editor(s) please ensure that every major paragraph has a citation? It will be required for GA promotion. Hope to be finished a set of initial comments in about 4-5 days. Sasata (talk) 05:36, 29 October 2009 (UTC)

Ok, I've just read the article closely and made some minor copyedits to the text. I must say I'm impressed with the readability of the article, and I think it serves as an excellent introduction to the topic. Just two things I'd like to see:

• more citations... this will be necessary before I can promote the article
• a search for review papers using the term "genetic drift" in PubMed turns up 190 articles, many of which look like they would be excellent sources to add to this article. Please consider using some of these review articles as sources; this will also benefit the reader who is using this article as a starting point for further investigation of the topic. I'll place this review on hold. Sasata (talk) 18:33, 1 November 2009 (UTC)

Hi again, how are we doing with this review? I see some progress has been made in getting the article citation-compliant, but there's still many unreferenced paragraphs. Is more time required? Do you want some assistance? Sasata (talk) 15:36, 10 November 2009 (UTC)

Many paragraphs need no citation (basic and uncontroversial), which ones are you referring to? Narayanese (talk) 23:46, 14 November 2009 (UTC)

I would still like to have a source for the section on Probability and allele frequency. Yes, its basic to those with university training in the subject, but what if someone else wants to read more about it? It's common courtesy to give a link to a page or chapter in a general textbook so they can independently confirm that what is written here is true. How about the 1st paragraph of "Genetic drift versus natural selection". The first and second paragraphs of "Population bottleneck". Just add general refs for those and I think we'll be good to go. Sasata (talk) 05:15, 15 November 2009 (UTC)

Thanks for this. I forgot to "watch" this discussion page. That's why this comment is so late. The requirements sound reasonable to me. The general probability stuff is probably the most difficult for me to find. As you hint above, I considered this so basic that I dared to write it without looking anything up. --Ettrig (talk) 20:59, 17 November 2009 (UTC)

OK, I found one. Not the most orthodox kind, but reasonably authoritative, very pedagogical and follows the same lines of reasoning that I did. Please check it out and say what you think. --Ettrig (talk) 21:22, 17 November 2009 (UTC)

I think those referenses have now been added. --Ettrig (talk) 21:38, 28 November 2009 (UTC)

Ok, it looks good now. Thanks to you (and Narayanese) for helping to make a good article which will benefit countless students of genetics! Sasata (talk) 23:46, 28 November 2009 (UTC)

GA review (see here for criteria)

1. It is reasonably well written.

   a (prose): b (MoS):

   Very clearly written.

2. It is factually accurate and verifiable.

   a (references): b (citations to reliable sources): c(OR):

   A comparison of before and after GAN versions shows that many references have been added during the course of the review, and I believe the current version is now adequately cited.

3. It is broad in its coverage.

   a (major aspects): b (focused):

   Covers the basics of the topic without straying too much into what would be the domain of other related articles.

4. It follows the neutral point of view policy.

   Fair representation without bias:

5. It is stable.

   No edit wars etc.:

6. It is illustrated by images, where possible and appropriate.

   a (images are tagged and non-free images have fair use rationales): b (appropriate use with suitable captions):

   All image have appropriate free-use licenses.

7. Overall:

   Pass/Fail:

Relevance of recombination

What relevance does homologous recombination have to genetic drift? Emw (talk) 04:15, 30 October 2009 (UTC)

??? --Ettrig (talk) 07:55, 30 October 2009 (UTC)

Perhaps the following sentence from the 'Drift and fixation' section is an answer to my question: "It is impossible for a population to gain new alleles from the random shuffling of alleles passed to the next generation, but this shuffling can cause an existing allele to disappear." If that shuffling of alleles is referring to the recombination that occurs during meiosis, then I think the quoted sentence may be inaccurate.

First, the shuffling is notably non-random. Recombination rates vary between males and females, and -- by virtue of recombination hotspots -- between different parts of the genome. Second, the possibility of intragenic recombination means that DNA shuffling during meiosis can give rise to new alleles. Emw (talk) 18:52, 5 November 2009 (UTC)

I think that that is just a misleading formulation. The intended image is probably the shuffling of a deck of cards to achieve randomness. In this sense the analogy is only half. It misses the element of choosing a subset of the shuffled cards, the sampling. Think of the giant decks they use for Black Jack in Las Vegas. If you take out a subset of say 25 cards, you are bound to miss some of the card types in you mini-deck. But such a process will never add a card type. --Ettrig (talk) 08:14, 6 November 2009 (UTC)

When you say "that formulation", exactly what are you referring to? Is it the sentence I quoted from the article? I also don't see how the playing-cards analogy applies to my concern. My concern is simply that the article seems to imply that recombination-mediated crossover in meiosis doesn't produce genetic variation. Of course recombination produces variation through creating new combinations of alleles, however it also produces variation my creating new alleles in themselves via intragenic recombination. Emw (talk) 18:35, 6 November 2009 (UTC)

I meant to continue discussing the statement that you quoted from the article. I don't think that that statement was meant to refer to recombination on the molecular level but rather to the creation of new combinations of whole genes. I would like to note that genetic drift was thought of long before there was any understanding of the molecular basis of genetics. They knew the genes were in the chromosomes and that they could be reshuffled by crossover, exchanging parts of chromosomes. If crossover recombination had created a new gene, I think they would have registered that as mutation. That is the mindset in which I think this statement should be interpreted. I think I now can answer your first question. The answer is: None. --Ettrig (talk) 21:38, 6 November 2009 (UTC)

Higher recombination rate -> more genetic drift (Presgraves DC. 2005. Recombination Enhances Protein Adaptation in Drosophila melanogaster. Current Biology 15:1651). Narayanese (talk) 16:04, 8 November 2009 (UTC)

I don't have easy access to research data bases. The summary I found on the web says: Higher recombination rate -> less natural selection. Nothing about genetic drift. --Ettrig (talk) 21:03, 8 November 2009 (UTC)

I think the article I refered to [1] is open access. It's figure 1. I seem to have gotten the direction wrong though, it was lower recombination rate that increased drift (by increasing the size of selective sweeps, if I got it right, since selective sweeps reduce the effective population size). Narayanese (talk) 21:36, 8 November 2009 (UTC)

Yes, I see at last. Thank you! I have earlier mulled over the question whether sampling error is the whole explanation or whether there can be variability in genetic drift. This is a clear answer: Genetic linkage to genes that are exposed to selection increases the variance of the outcome, which really is the same as the drift. This Wikipedia article equates genetic drift with sampling error. This needs to be reconsidered. --Ettrig (talk) 09:29, 9 November 2009 (UTC)

(unindent) I just came across two papers relevant to the discussion:

• Awadalla, P (2003). "The evolutionary genomics of pathogen recombination". Nature Reviews Genetics. 4 (50–60). doi:10.1038/nrg964. PMID 12509753. {{cite journal}}: Unknown parameter |month= ignored (help)
• Mu, J; et al. (2005). "Recombination hotspots and population structure in Plasmodium falciparum". PLoS Biology. 3 (10): e335. doi:10.1371/journal.pbio.0030335. PMC 1201364. PMID 16144426. {{cite journal}}: Explicit use of et al. in: |first= (help); Unknown parameter |month= ignored (help)CS1 maint: PMC format (link) CS1 maint: unflagged free DOI (link)

At the moment I don't have time to adequately summarize those papers here, but thought I would note that they seem useful here because they discuss the relationship between genetic drift and recombination in the evolution of pathogens (with data from P. falciparum). Emw (talk) 05:02, 14 November 2009 (UTC)

Linkage (selection) attenuates drift

At the very top of this talk page (19 February 2009, maybe time to archive) I questioned the limitation of the genetic drift concept to sampling error. My view is that genetic drift should include all stochasticity in alele frequency changes. But there was considerable opposition, supported with authoritative sources, so I gave up. But now I am happy to see that the latest major addition widens the concept. But I am unhappy with details in the formulations. The most important one (imho) is that variability is misssing.

Graur & Li say: genetic drift can also be caused by ... stochastic changes in selection intensity

Desgraves says: selection at one locus increases the variance in reproductive success, and thus the strength of genetic drift

Please note the word variance in Desgraves. The hitchiking pulls in different directions at different times because of recombination. This adds stochasticity in the reproductive success. Such stochasticity is part of the more general definition of genetic drift. In the evolutionary time scale, recombination happens very often, so the effect is almost exclusively random. It seems to me that the variation aspect is not present in the present article. --Ettrig (talk) 15:33, 15 November 2009 (UTC)

I read parts of the referenced article by Sella et al. It clearly states in the abstract that the view presented is controversial and tentative. Such material should not be presented as the view of Wikipedia. I look forward to a very interesting discussion about this. --Ettrig (talk) 12:38, 18 November 2009 (UTC)

I thought I had avoided the controversial parts (which is the strong/widespread hitchhiking effect and associated jumps in allele frequency, not the altered drift)... not that there was anything I used which wasn't also i the textbook I cited. I'd prefereably not have to go through without an open access ref to replace it.

Concerning your changes to the paragraph: I can't see how it is any less a sampling effect than the simple model (i e infinite recombination rate/no linkage), and the papers I've seen treat it as yet another source that influences effective population size what I can see. Narayanese (talk) 23:02, 20 November 2009 (UTC)

Below is a detailed analysis of the recently added paragraph that I find problematic.

Selection causes the effect size of genetic drift to vary across the genome. From the classical view, this is enigmatic, because selection and drift are classically viewed as distinct processes working in parallell so that their respective effects (changes in allele frequencies) are added to each other. Here, selection acts ON drift. An explanation is needed (by the typical reader of this Wikipedia article). I interpret the following sentences as a trial to provide this explanation.

When a locus is under selection, alleles of nearby linked loci will be affected as well OK, selection works on stretches of DNA that travel together. (We are interested in the drift of the nearby loci.)

as if they were under selection, an effect called genetic hitchhiking. OK, repetition and naming, no additional fact.

This reduces the effective population size of those loci This is enigmatic to me. It means selection increases drift. As I say above, selection and drift are normally seen as separate processes. I think a Wikipedia article needs an explanation of this. I see the statement in the open access reference. But the explanation there is not intelligible to me, therefore probably not to most article readers either.

that is, it is as if the population that genetic drift acts upon is smaller. Repetition.

This increased genetic drift decreases genetic diversity near sites under selection. This increase is enigmatic. That drift removes variants is explained earlier in the article.

In addition to the degree of selection, Repetition.

the rate of recombination is an important factor, as Very little factual content. (OK)

a higher rate reduces the the linkage of a locus under selection to nearby loci, OK, true.

thereby reducing genetic drift Repetition of the enigmatic statement, but using the opposite concepts.

To summarize: My problem with this paragraph is only the statement that selection increases drift. --Ettrig (talk) 08:22, 22 November 2009 (UTC)

If I had a better grasp of the mathematical model used to explain the observed effect of recombination and nearby functional regions on diversity in presumbptive neutral regions, I would have written it better than just the current way where selection 'magically' reduces a genetic-regional population size. Perhaps it is better to describe the data (e g recombination--diversity correlation) more and be light/more concise on model? Narayanese (talk) 07:09, 24 November 2009 (UTC)

The following excerpt is problematic: Selection causes the effect size of genetic drift to vary across the genome. The effective population size is reduced by selection (by a factor f0 which is the equilibrium frequency of alleles that are not eliminated by selection), increasing the genetic drift at the locus. Also nearby loci are affected as a result of genetic linkage. The statement in the last sentence should be first. The linkage is part of the explanation of the reduction of effective population size. (By linkage, deleterious mutations remove subsets of gametes, leaving only f0*Ne for our mutations to spread in.) It is not reasonable to expect the Wikipedia reader to understand what equilibrium frequency refers to here. My interpretation is that it refers to the frequency of non-mutated gametes that is maintained by selection when a stream of deleterious mutations remove the other gametes. --Ettrig (talk) 21:58, 28 November 2009 (UTC)

Primary sources

The history section could use full references to the works that are mentioned ("Sewall Wright ... 1929", "the work of Ernst Meyer" etc) in addition to the secondary sources. Narayanese (talk) 21:21, 3 November 2009 (UTC)

Unexplained step in jar experiment

The description of the jar experiment does not make clear that at each step, the marble has to be returned to the original jar before selecting the next one. When I first read it, I puzzled over the process, and just couldn't understand how it could work... until I realized that the this was what was intended and it just was not stated explicitly when describing the experiment. I've added a sentence to make it clear. (If by chance I've misunderstood, please fix it.) Omc (talk) 07:48, 3 February 2010 (UTC)

I remember puzzling over the same problem. My "solution" was to notice that the text does not say that the selected marble is removed from the jar. I tried to clarify this. I do not want the text to say that the marble is removed. That would make it more complex, in my view. --Ettrig (talk) 07:47, 4 February 2010 (UTC)

To me it's more natural to describe the process as "remove one marble from the jar, examine it, and return it". Any experiment like this is typically described by saying "remove one marble from the jar", I think. The image of selecting one marble but leaving it in the jar (pushing your fingers deep into the jar at each step to select a random marble?) seems awkward. But this is quibbling - my main point is that we needed to make it clear that at each step we're selecting one marble from the whole population (20), rather than removing a marble and setting it aside and then selecting from the remaining marbles (19, 18, 17, ...). Your version also makes that clear, and it's fine. Omc (talk) 17:41, 4 February 2010 (UTC)

Hmmm. This illustration/experiment seems overcomplicated. The process being decribed would seem essentially to be:

There are two colours of marbles in a group.  
1. Choose one of these colours >  Record your choice >  Repeat x20.  
2. Compare the proportions of colours in your record and the original.

From the article it seems that no more is being said than that the proportion of alleles in a future generation is determined by random selection as well as by natural selection. If this is the case, then rather than dive straight into abstruse mathematical expression, some real examples of this mechanism of chance (picking of the colour of marble for the new group) would help make the point better. LookingGlass (talk) 08:02, 6 March 2014 (UTC)

The jar model could be explained more clearly. It could include sentences like these: "Because each marble is returned to the jar, the percentage that are blue does not change during the initial sampling round (which chooses marbles randomly). Therefore, each time a marble is chosen, it is equally likely to be blue or red. Replacement of the "parent" marble in the first jar is necessary to keep the probabilities constant. It's therefore possible (indeed, likely) that by chance more blues will be chosen than red (or the converse), just as N successive coin tosses can lead to sequences with more heads than tails (or the converse). In both cases the initial outcomes are given by the binomial distribution with p = 0.5. In the Figure, the number of times particular marbles are selected are indicated by the number of dots within the marbles. Some marbles are not chosen, others are chosen 1 or more times (in the first set of 20 samples illustrated, one marble happens to be chosen 5 times). However, when the sampling process is repeated, to create the third jar, the probabilities of drawing blue and red marbles changes (p > or < 0.5 for the binomial distribution) because there may be more (or less) blue marbles in the second jar as a chance result of the first round of random sampling. In the case illustrated, there was a majority of blue marbles in the second jar, so the new samplings needed to create the third jar are now biassed in favor of blue selections. This can eventually (and in the case illustrated, after only 5 rounds) lead to an entirely blue jar. Additional sampling rounds will no longer change the outcome, which is "fixated"."

This is rather long-winded and not necessary for those who already understand what is going on - but those people don't read Wikipedia! Paulhummerman (talk) 03:48, 14 November 2017 (UTC)

Yes, surely some of this could be added. If you are unsure, make the change in small steps and allow some time for reactions in between. Please avoid saying the same thing twice or more, such as the percentage that are blue does not change, equally likely, probabilities constant. --Ettrig (talk) 11:47, 14 November 2017 (UTC)

Who is Moran?

Can someone provide information about Moran from "Moran model"? Maybe a link to a page (stub) with that name? Is Moran model identical with or related to Moran process? 132.66.40.82 (talk) 16:07, 28 October 2010 (UTC)

Yes it is identical to the Moran process, I added the link, and that page in links to Pat Moran who developed it.Joannamasel (talk) 17:52, 28 October 2010 (UTC)

Can changes due to genetic drift be beneficial or detrimental to reproductive success?

The following statement has been marked with "citation needed" and as "dubious - disputed" in the main article:

“In contrast to natural selection, which makes gene variants more common or less common depending on their reproductive success, ^[2] the changes due to genetic drift are not driven by environmental or adaptive pressures, and may be beneficial, neutral, or detrimental to reproductive success.”

The reason is the following: if the changes due to genetic drift were other than neutral (viz. beneficial or detrimental to reproductive success) then they would be made "more common or less common depending on their reproductive success". Therefore they would be indistinguishable from random mutations upon which natural selection operates.

The warnings should not be removed until the present dispute is resolved.
Miguel de Servet (talk) 10:41, 17 January 2011 (UTC)

The statement should be clarified something like this: Since the effect of genetic drift is random, it may make a detrimental allele more common as likely as less common and vice versa a beneficial allele less common as well as more common. For alleles that are not neutral, the total change is the effect of selection plus the effect of drift. Hope this helps. --Ettrig (talk) 11:20, 17 January 2011 (UTC)

"if the changes due to genetic drift were other than neutral (viz. beneficial or detrimental to reproductive success) then they would be made "more common or less common depending on their reproductive success"--not necessarily. For example, in very small populations the magnitude of the drift can be greater--outpacing the lesser effect the allele has on reproductive success. Take the evolution of alleles in a population which are detrimental only in individuals who are homozygous for it--such as Tay-Sachs disease which is fatal before the individual reaches reproductive age. Drift and natural selection may both operate on the allele, but if drift overshadows the selection then the evolution, the change in frequency of the trait, results from drift. I don't quite understand the latter point about mutations. Random mutation is a source of new alleles in a population. Natural selection and random drift are two very different mechanisms that increase or decrease the frequency distribution for the mutation in a population.Professor marginalia (talk) 17:07, 17 January 2011 (UTC)

Random mating

Both Hartl & Clark and Hartwell et al state the H&W conditions first, including random mating, then state that frequencies remain constant under these conditions. If we are to remove the random mating condition, which seems reasonable, we need to refer to another source. --Ettrig (talk) 19:02, 12 February 2011 (UTC)

The usual 5 HW assumptions lead collectively to two HW results, ie constant allele frequency, and genotype frequencies of p^2, 2pq and q^2. But taking only one of the two results, not all 5 assumptions are needed. There is an excellent and authoratitive discussion of HW in Ewens' "Mathematical Population Genetics", although this source is perhaps a bit inaccessible to a casual reader. I am not sure which other books cover this subject carefully enough. Shall we cite Ewens?Joannamasel (talk) 22:25, 12 February 2011 (UTC)

Yes, certainly. There is no requirement that the cited source is accessible on the web. I found this but didn't have the patience to determine whether it supports the reduction of assumptions. If it can be used, please specify one or a few pages. --Ettrig (talk) 06:53, 13 February 2011 (UTC)

Picture

Hello Genetic Drift Editors. I thought i would add the Sewall Wright picture to the article because 1) he was a major contributor in our understanding of the topic and 2) I thought that after alot of the charts that are in the article thought it would be nice to add a face in their to kind of associate with article. This such a well done article I just wanted to give it an extra kick.--Jraffe0404 (talk) 04:54, 6 March 2011 (UTC)

Paint

I'm not very fond of the paint analogy. I'll think about it and perhaps try to be more constructive. 212.126.224.100 (talk) 13:57, 23 August 2011 (UTC)

You might be interested in looking at the version that was here before user:Graft rewrote it completely, December 23 2010. --Ettrig (talk) 14:50, 23 August 2011 (UTC)

This sentence The effect of genetic drift is larger for alleles present in a smaller number of copies, and smaller when an allele is present in many copies. is criticised. Copies of what? The answer is copies of alleles. When the question is posed, I think this could be interpreted as alleles in copies of something. We want to say When there are few copies of an allele, the effect of genetic drift is larger, and the effect is smaller when there are many copies. I came here to ask for help with the formulation. But I will try this and hope it provokes improvement. --Ettrig (talk) 06:23, 19 December 2011 (UTC)

The "cans of paint" analogy makes no sense because there is no "copying" or "replication" of the paint in the process. Genes aren't a handoff, a give and get, between parent and offspring (in the example paint is being poured from one pail to another, with the original remaining full as if by magic). It's a weird and confusing image, imo, this oddly behaving farmer and his magic paint cans. Professor marginalia (talk) 19:09, 27 January 2012 (UTC)

I changed it back to the GA version. I'm not sure that the analogy section (in either form) fits the rest of the article well, but at the very least if we are going to have one the alleles need to be copied, not handed-off like the baton in a relay, from one generation to the next. And at least marbles are discrete units, like alleles. Symbolizing them as "ever-filling paint cans" is a bit more confusing. Professor marginalia (talk) 19:34, 15 May 2012 (UTC)

Other models

I find the text about other models very difficult to understand.

If the variance in the number of offspring is much greater than that given by the binomial distribution assumed by the Wright-Fisher model, then given the same overall speed of genetic drift (the variance effective population size), genetic drift is a less powerful force compared to selection.[9] If higher moments of the offspring number distribution exceed those of the binomial distribution then again the force of genetic drift is substantially weakened.[10]

Should the variance effective population size be changed to the variance in the effective population size? This form normally means that the variance is an alternative to overall speed. But they are not of the same kind. Should we also add ... is the same? I am also completely lost on moments of the offspring number distribution. --Ettrig (talk) 10:18, 27 January 2012 (UTC)

"Variance effective population size" is a technical term, and I edited the link to go directly to that. Ditto for moments of a distribution. I agree that the text is difficult, but genetic drift really is an technical mathematical subject, even if that's not how it is taught in intro classes. Joannamasel (talk) 16:44, 27 January 2012 (UTC)

lead sentence

I've proposed a change to the lead sentence of the natural selection page, and I think a corresponding change should be made to the lead sentence of this page.

My proposed lead sentence for natural selection is "Natural selection is the differential survival and reproduction of individuals due to differences in phenotype."

My proposed lead sentence for this page is "Genetic drift is the differential survival and reproduction of individuals due to chance". (the reference can stay)

These edits would make the descriptions more accurate, and help highlight the parallels (and by doing so also the fundamental differences) between natural selection and genetic drift.

I can see that some sort of joiner sentence would need to follow; perhaps "From a genetic perspective, it can be defined as change in allele frequencies in a population due to chance." 129.127.101.245 (talk) 06:05, 25 September 2015 (UTC)

"Differential" raises the question of "differentiated with respect to what", the answer to which is "phenotype". There is no such thing as differentiation with respect to chance. Also, the current definition does not encompass all chance mechanisms (including mutation and recombination) only random sampling, so this would be a complete alteration of the definition of genetic drift. I realise that normally my identity is irrelevant on wikipedia, but in this case I happen to be the author of the reference in question: I assure you that this was not the definition used in that publication. For a different definition, you would need a different ref. Joannamasel (talk) 15:40, 25 September 2015 (UTC)

It only raises the question "with respect to what?" if you don't finish reading the sentence. It is very clearly referring to individuals -- a difference in survival and reproduction of INDIVIDUALS; which very well can be determined by chance. As for "chance" vs "random sampling", I don't believe using "chance", in the context of the whole proposed definition, has the problems that you suggest because "chance" in that definition is only relating to survival and reproduction, which excludes recombination and mutation. Furthermore, I don't understand how you could object to the proposed definition but be fine with the sentence following the current definition, as they essentially say the same thing. I also just realised there is some discussion about this in the first talk topic and it seems a number of people have objection with the use of "random sampling". Having said that, the only problem I have with using "random sampling" is that it's not clear what that involves and is likely to be confusing to the uninitiated; it's sort of clarified in the following sentence, but the proposed definition would remove the need for that sentence.

I do concede that there is a problem with my hastily constructed suggestion for a tie-in sentence "From a genetic perspective, it can be defined as change in allele frequencies in a population due to chance." This clearly is confounded by the issues you mention, and could be fixed by changing to "From a genetic perspective, it can be defined as change in allele frequencies in a population due to random sampling." With the proposed definition leading this there should be no ambiguity about what random sampling entails.

I don't see why the reference couldn't be kept for the proposed definition -- it may not be a verbatim quote, but the proposed definition and the original are essentially the same, so if it's suitable for one it should be suitable for the other.

As stated, the proposed definition would help highlight the parallels (and by doing so also the fundamental differences) between natural selection and genetic drift, and I think would also make it less open for interpretation.129.127.101.245 (talk) 01:28, 29 September 2015 (UTC)

I continue to oppose. The proposed definition also shifts the focus from allele frequencies to individuals. I know of no definitions in the literature that are not defined with respect either to allele frequency nor to genotype frequency but instead to individuals. There is no non-genetic concept of genetic drift. The proposed definition and that in my cited publication (=current sentence) are most definitely not the same. Joannamasel (talk) 04:19, 29 September 2015 (UTC)

Okay, I see your point. Perhaps instead the proposed sentence can be modified to replace the sentence following the lead, to make it more comparable with the definition of natural selection (if only the two phenomena were defined in the same period we may have some more congruity between how they are defined). I would suggest replacing "The alleles in the offspring are a sample of those in the parents, and chance has a role in determining whether a given individual survives and reproduces" with something like "This change is due to differences in survival and reproduction of individuals based on chance" and maybe ", in contrast to natural selection where change is due to differences in survival and reproduction of individuals based on differences in phenotype.", although these proposed changes do seem a bit clunky...

At the very least, I think the second sentence should be changed because "The alleles in the offspring are a sample of those in the parents" makes it focus too strongly on reproduction and not enough on survival.129.127.101.245 (talk) 10:03, 29 September 2015 (UTC)

"Sampling" is Wright's original terminology. It is a statistical term that covers both survival and reproduction. I don't agree with a need for parallelism in wording between selection and drift. Do you also want to introduce parallelism on the mutation page, because this is another evolutionary force? I think it is best to stick to well-established wording in the opening. If you feel that someone reading more of the article could fail to understand what sampling means, I would make changes in the bulk. Joannamasel (talk) 13:46, 29 September 2015 (UTC)

I agree with Joanna on this - I wouldn't want to try to redefine the historical and accepted definition of genetic drift on this page. (Aaronragsdale (talk) 17:47, 29 September 2015 (UTC))

Out of style passage near end of history

Could someone please remove (or thoroughly rewrite) this. It is completely out of style with the rest of the article. (I am out of reverts.) --Ettrig (talk) 16:53, 24 November 2015 (UTC)

Unclear wording

"which claims that most instances where a genetic change spreads across a population (although not necessarily changes in phenotypes) are caused by genetic drift acting on neutral mutations.[6][7]"

I found the relationships between the three concepts, "phenotype," "genetic drift," and "genetic change" impossible to unravel! I understand the concepts themselves, I just don't understand what this sentence is trying to say about them.

OmneBonum (talk) 07:25, 21 February 2019 (UTC)

most instances where an allele becomes the most common variant in a population it is not because it is selected for, but because of random changes in allele frequencies. Most of the allele changes that become common this random way have no or negligible effect on the phenotype. --Ettrig (talk) 08:11, 21 February 2019 (UTC)

Wiki Education assignment: First Year English Composition 1001

This article was the subject of a Wiki Education Foundation-supported course assignment, between 23 August 2023 and 30 November 2023. Further details are available on the course page. Student editor(s): Amaner222 (article contribs).

— Assignment last updated by RuthBenander (talk) 14:05, 28 September 2023 (UTC)

Marble analogy

“To represent this reproduction, randomly select BUT DO NOT REMOVE a marble from the original jar and deposit a new marble with the same colour into a new jar.”

I struggled a little with the logic of this analogy until I understood it as amended above. If I select a kitten from a litter I take it with me.

Someone with more authority could make the change if deemed helpful. =avallone (talk) 15:02, 5 May 2024 (UTC)