Talk:Quantitative genetics/Archive 1

This is an archive of past discussions about Quantitative genetics. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

I have a problem with the treatment of this section. It seems to imply that any polygenic trait is quantitative. I fail to see how such traits as butterfly wing patterns and coloration are quantitative traits. Even though they are polygenic with interesting epistatic relationships, they can be analyzed using nonquantitative methods.

A quantitative trait is one that

• is continuous, or approximately continuous
• shows variability in the population
• has unknown genetics
• is affected by the environment

In addition, we can use quantitative genetics methods as an approximation for ordinal data (such as the standard bristle number analysis in Drosophila). Further, we can use these methods as a basis for understanding some qualitative traits that appear to have a basic underlying quantitative "cause," such as threshold or multiple threshold models. —Preceding unsigned comment added by TedE (talk • contribs)

You add a few aspects to the defintion that are not part of it. The unknown genetics for example is irrelevant. You can study something using both quantitative and molecular techniques. KimvdLinde 05:19, 27 March 2006 (UTC)

I'm not entirely stuck on the "unknown genetics" part. I use it to help distinguish between normal classical analysis of data and analyses based on ANOVA. Even with QTLs, you search loci involved in a quantitative trait, but you don't know the actual genetics of its interaction with other loci. In those cases where we have discovered single genes (or small gene complexes) that have a large effect on a quantitative trait, we have controlled for those genes, then done the quantitative analysis on the remainder. Still, I don't consider it a big deal and can strike it if you wish. TedE 13:15, 27 March 2006 (UTC)

Ok, what is the point of your argument, I fail to get it.... I have used defintions as used in all major textbooks, including Falconer, Roff, and Lynch. So, what is wrong with the definition as used normaly by major scholars in tis field? My Ph.D. supervisors, as well as my current boss use the two things side by side, and I think you try to insert a seperation between fields that is not there. KimvdLinde 13:24, 27 March 2006 (UTC)

Are all polygenic traits quantitative traits? Ted 13:50, 27 March 2006 (UTC)

Your example above was a perfect example of a quantatative trait. The example of butterfly wing patterns and coloration is also an perfect example of quantitative trait. That you know that a single major gene has a large effect and that you can control for it in the analysis, does not make the trait suddenly non-quantitative. KimvdLinde 14:03, 27 March 2006 (UTC)

OK, that's where we disagree. Simply because it is polygenic does not make it quantitative. Studies of the HLA complex is not quantitative, although it is certainly polygenic. Since you mention Falconer, here is a quote from his 1960 edition of his book, "Variation of this sort, without natural discontinuities, is called continuous variation, and characters that exhibit it are called quantitative characters or metric characters because their study depends on measurement instead of on counting.... A branch of genetics has consequently grown up, concerned with metric characters, which is called variously population genetics [I disagree with this classification], biometrical genetics or quantitative genetics." Or, look at Fisher's 1918 paper, "... biometry in accordance with the Mendelian scheme of inheritance." Or, Kempthorne, "...concerned with the study of phenotypic characteristics which do not fall into a small number of distinct homogeneous classes ... but have the essential feature of being continuous and graded." Well, you get the idea.

While wing shape can be studied as a quantitative trait by quantifying the shape (such as the Drosophila studies by Laurie-Ahlberg and others), the butterfly pattern analysis is a terrible candidate for quantification.

On the other hand, who really cares? I'll leave the article as is. Ted 14:40, 27 March 2006 (UTC)

That studies are done in a non-quantitative manner does not imply it can not be done that way.

Do you have a citation for the Laurie-Ahlberg articles, I can not find them.

And I will tell my (former) Ph.D. supervisor that his butterfly wings are a terrible candidate for quantification :-) KimvdLinde 15:02, 27 March 2006 (UTC)

I suggest you learn to use Google Scholar, it takes about 15 seconds to find her publications, particularly since she gave up the '-Ahlberg' part in more recent publications. Most of her work as Laurie-Ahlberg involves wing parameters and allozymes. As for your supervisor, what can I say? The current research I see on H. erato and H. melpomene suggest to me that quantification would be a big step backwards. This is not my area of research, but from the seminars I've attended, there isn't any quantitative genetics. Instead, what I see is classical epistasis/linkage analysis along with a large EST project. Some really beautiful classical genetics detective work.

I give up. You've won. I'll leave this article to you and work on something more esoteric. Ted 17:25, 27 March 2006 (UTC)

Dropping the second part explains maybe somethings. So I searched Web-of-Science with the folowing search string: TS=(wing) AND AU=((Laurie-Ahlberg, C*) OR (Laurie, C*)), resulting in zero hits.

I suggest that you run this search string in Web-of-Science: TS=(butterfly AND ((wing AND patterns) OR eyespot or eyespots)) AND AU=(Brakefield, P*). It just gives you around 50 hits, many of them using selection experiments and quantitative genetic approaches to study wing patterns in Bicyclus anynana. KimvdLinde 17:40, 27 March 2006 (UTC)

I have the feeling that the two articles are becoming more alike, and that the heritability article contains substantial stuff that either should be under quantitative genetics, or under a own article of twin studies. Most estimation methods can be used to estimate variances as well as covariances, and the derived measures as heritabilities and genetic correlations are only in some methods estimated directly. I was at the point with this article that I had to add estimation methods, whihc for a large part would be a simple copy-paste from heritability, and then extend it to covariances. So, I propose to merge both. Each estimation method that growths out of control can get its own page, makes more sense to me. --KimvdLinde 06:19, 26 February 2006 (UTC)

I'm inclined to say "No." Heritability has an interest among nongeneticists that you don't see for Quantitative Genetics. Often, the term is used (or, too often, misused) in popular scientific literature and would be a crime to have it as a subsection for quantitative genetics. In the same way, founder effect is separate from population genetics. No, if you want to make the bulk of the variance components estimation in quantitative genetics, then go ahead and make the quantitative genetics part. Once that gels, then the section in heritability can be reduced. Or, you can take the Quantitative Genetics entry off in different directions and refer people to heritability for estimation. There is, after all, a whole lot more to quantitative genetics than just estimation of variance components. Ted 07:05, 26 February 2006 (UTC)

Ok, I see two uses of the word heritability. One is the popular use, for which I agree that the term should remain, and without all the technical stuff is would make a good laymans enrty. QG is than the more technical article dealing with how exactly you estimate it etc. --KimvdLinde 07:19, 26 February 2006 (UTC)

I'm not sure what you would consider technical stuff. A discussion of heritability without mentioning variability is vacuous. My suggestion is to work on the Quantitative Genetics entry however you best feel it should be done. The heritability estimation part has left out the Fisher/Comstock-Robinson/Kempthorne/etc. methods of variance component estimation, which is probably more general. Do the regression/correlation estimation as well. Then, we can see how best to divide up the information between Quantitative Genetics and Heritability. I guess, I'm saying the discussion might be a little too early. You might even want to talk to the person who added most of the material on her/his Talk page. It isn't that old, so she/he might be enthusiastic about moving it over. Ted 21:12, 26 February 2006 (UTC)

I suggest creating a new article labeled "qualitative genetics" and separate the two articles (heritability and quantitative genetics) into their most accurate theoretical and mathematical standpoints between qualitative and quantitative genetics. If this poses too much of a problem (the theory and the math are intricately related, then I would move the articles under the heading "population genetics" and subtitle the various articles on Wikipedia that relate to population genetics and make it more textbook format with links, etc. -E. D. Sperry 19:47, 24 March 2006 (UTC)

I'm not sure what you consider "qualitative genetics," and how that is a part of quantitative genetics. Ted 04:41, 25 March 2006 (UTC)

I have taken away the merge proposal. It has been over a month with only three votes (1 for, 2 against). Not much of a groundswell for either position. Ted 08:27, 30 March 2006 (UTC)

in the credit given in the article? It surprises me that the 3 founders are British or American.-Richard Peterson130.86.14.82 (talk) 23:04, 3 February 2008 (UTC)

No, Haldane, Fisher, Wright, those are the Big Three for quantitative genetics. Two Britons and one American. --Crusio (talk) 23:19, 3 February 2008 (UTC)

I agree. There really are not any people who contributed to this field and were not white. I don't think that it is a bias, just reality. 216.212.225.145 (talk) 03:33, 5 February 2008 (UTC)

So are you saying Mooto Kimura and Masatoshi Nei are white? or that they didnt contribute significantly? —Preceding unsigned comment added by 128.192.47.71 (talk) 17:10, 2 April 2009 (UTC)

No, they would be considered "yellow". —Preceding unsigned comment added by 164.144.248.26 (talk) 19:26, 11 August 2009 (UTC)

And they came much later (kimura was born in 1924 and Nei 1931), and are not considered founders of population genetics (although they made important contributions). You could say that Haldane later became an Indian citizen... —Preceding unsigned comment added by 129.7.227.11 (talk) 22:31, 11 May 2011 (UTC)

I don't know if I would say that Fisher, Wright, and Haldane are the founders of quantitative genetics. They are often referred to as the three main guys involved in the Modern Synthesis of the Darwinian evolution by natural selection with Mendelian genetics, but we also know that there were many people involved in the Modern Synthesis other than those three (they were definitely the main contributors of genetic theory in general though). I'm pretty sure that Fisher published the original paper in 1918, and Wright followed a couple years later with his analysis of coat colors in guinea pigs which utilized his method of path analysis. Whenever I'm asked, I say Fisher, or Fisher and Wright. Ddobias (talk) 06:09, 17 January 2014 (UTC)

Do you have any suggestions of helpful reliable sources for updating this article. This is one of several articles I have in mind to update extensively, and as I plunge in, I'd be delighted to follow where the best sources lead in resolving the editorial questions that have come up in this talk page section. (And I read languages other than English, and have access to sources in other languages, so feel free to suggest sources in any language you know.) -- WeijiBaikeBianji (talk, how I edit) 16:33, 17 January 2014 (UTC)

Hey, I have a couple texts on my desk that I'll go through today. One text that I can think of off the top of my head is Evolutionary genetics: concepts and case studies. This is a volume edited by Charles Fox and Jason Wolf in which Jason Wolf does a brief historical overview of the field. I think he traces the origins of the field to before Fisher, Wright and Haldane when people (like Karl Pearson) were still struggling with concepts of inheritance. I'll start making some edits on this page too. Do you have any areas that you want me to start on? Ddobias (talk) 15:41, 5 March 2014 (UTC)

In the article text, I see the statements, "Low heritability leads to higher environmental influence, while high heritability can lead to more rapid genetic progress. Typically, heritability of reproductive traits is low, disease resistance and production are moderately low to moderate, and conformation is high." Those two sentences next to each other raise some interesting questions, which I hope readers of this article will be able to resolve by further reading. This article already lists some of the standard textbooks in the references, but no reference is cited for these statements. Where can these statements be verified and put into context? -- WeijiBaikeBianji (talk, how I edit) 16:52, 4 June 2014 (UTC)

• I think that's the weirdest definition of repeatability that I have ever seen. --Randykitty (talk) 18:11, 4 June 2014 (UTC)
  • Yes, the terminology looks weird in this section. I found an article about Fisher's work that may explain some of what those statements mean, but that section definitely needs better referencing and fleshing out. -- WeijiBaikeBianji (talk, how I edit) 18:55, 4 June 2014 (UTC)

What's the latest thinking among editors who watch this page on how it relates to other Wikipedia articles about genetics? -- WeijiBaikeBianji (talk, how I edit) 21:19, 18 December 2013 (UTC)

• Frankly, at this point I think the article is pretty bad. As so often with this kind of pages, it's concentrating too much on human populations. It also contains erroneous stuff ("The Environmental variance is much more straightforward. This can be subdivided into a pure environmental component (E) and an interaction component (I) describing the gene-environment interaction") or rather stuff ("Calculating relationship from known relationships"). It talks about pathway diagrams without really explaining those and, in any case, we have already an article about Path analysis (statistics), which is not really a purely genetic method. The sourcing is scanty and there is not much logic flow. In short, I think this should be re-written from scratch... --Randykitty (talk) 16:08, 19 December 2013 (UTC)
  • I agree. It has been tacked together by several editors in an ad hoc manner. Please note my comment in the section above on merging with "Population Genetics": I think it may be a useful way to proceed. Donach 01:13, 13 July 2014 (UTC) — Preceding unsigned comment added by Donach (talk • contribs)
  • I'm willing to do the rewriting from scratch, and I'm glad you've made specific suggestions and that you have this article on your watchlist. I've seen you do a lot of good work on other articles, so I'll try to carry the load here, and will be glad to collaborate with other editors who have suggestions. -- WeijiBaikeBianji (talk, how I edit) 17:47, 19 December 2013 (UTC)

This article suffers from some incorrect terminology:

-Mixing up genes and alleles (children share all of their genes with their parents and all other human beings, except in the cases of chromosome duplications or deletions; they share on average 50% of the alleles with one parent, this being limited to those genes where their parents carried different alleles). -Dominance is NOT interaction between genes, it is interaction WITHIN a gene, that is, when the heterozygote's genetic value is not simply the mean of the two homozygotes. -Interaction BETWEEN genes is called epistatic interaction and is generally indicated by "I". -Interaction between genotype and environment is generally NOT indicated by "I" but by G*E. -If this is to be complete, then genotype-environment covariation (cov(G,E)) should also be included.

The concept of alleles should be introduced to this article and the above incorporated, unfortunately I currently lack the time for this.... --Crusio 10:46, 29 September 2007 (UTC)

Thank you for the above. This distinction should also be noted in the "Twin Studies" article which states that dizygotic twins share about 50% of their genes, followed by formulae supposedly used for calculating the relative influences of genes and environment in various traits. 60.234.217.14 08:02, 12 October 2007 (UTC)

Many of the issues raised by Crusio (above) have been addressed recently. Some historical overview has been added also, which includes criticism of the "iconic" additive genetic variance. A key recent revisionist paper is cited (I.L.Gordon). Donach (talk) 08:49, 22 June 2010 (UTC)Donach

I don't agree 100% with the above arguments. I think if we are trying to use consistent terminology it would be better to use locus/loci and allele/alleles. Often in the descent measurement (inbreeding coefficients, coancestries, etc.) literature authors will talk about the probability of a relative possessing "gene copies" which I assume means allele since that is how identity is defined. Also, quantitative geneticists rarely identify the genes (protein coding regions) causing trait variation without the help of other sciences like molecular biology, and sometimes the cause of trait variation isn't a protein coding region, but an intergenic region of the chromosome. Either way, the term locus works. Ddobias (talk) 06:49, 17 January 2014 (UTC)

• Is it nearly time that this comment was deleted ? There has been so much revision since the comment was made it is barely relevant anymore.125.237.12.175 (talk) 04:05, 6 September 2014 (UTC)

• We don't delete old comments because they have been addressed. If a talk page gets too cluttered, we can archive them, but that is not yet necessary here, I'd say. --Randykitty (talk) 11:32, 6 September 2014 (UTC)

population genetics ??? Michael Hardy 03:42, 30 Jan 2005 (UTC)

No! I think the problem here is that the quantitative genetics page is not actually about quantitative genetics! In broad terms, quantitative genetics deals with the genetics of continuously varying characters - this may be in natural populations, or may not. It is often couched in terms of selection, but need not be (e.g.QTL analysis). In fact, Population genetics and quantitative genetics, although linked in some ways, each represent truly massive fields of their own.--DJO 22:05, 19 Apr 2005 (UTC)

• I agree with DJO. In fact, there are useful points of view and statements in the Population Genetics article, and I do not think it useful to re-state them here. [I liked particularly the Lamarkian overview made in the Population Genetics article.] I think the two should be wiki-linked. Concentrate here on the things which are different: namely the continuous variability of the phenotype. This adds a whole layer of complexity which Population Genetics avoids! But the need to consult both needs to be stated. Donach 01:17, 13 July 2014 (UTC)
• I also think that Quantitative genetics should remain distinct from pop-gen page. I'm glad to see that it has it's own page, however I do think this page could use some improvements. Can the equations be summarized or re-written into equation blocks rather than bolded text body? I feel that it is currently hard to follow. This page my even be improved by deleting the equations with '?'s. Second, I think this page would benefit from a History section that outlines the seminal researchers who started the field before DNA was even discovered (e.g. Fisher or Wright). Third there should be a section for multivariate trait analysis. I can add summeries of G, B and P matrices to the correlated attributes section. Personally this is the portion of Qgen that I need the most help understanding. Unfortunately wikipedia lacks most of the basic information about this major aspect about quantitative genetics. Petersoapes (talk) 16:33, 19 October 2015 (UTC)