Talk:Heritability

From Wikipedia, the free encyclopedia

Quite a bit of work you've added. I'll look at it later in more detail.

• Is "selection strength" used in psychology? The normal term in agriculture/biology is "selection differential."

Ted 15:09, 29 January 2006 (UTC)

Hmmm. You're probably right, "selection strength" is a casual phrase used in place of a correct term. I'll check with the usual sources, but my sense is that "strength of selection" is the common biology term I was thinking of. Cheers, Pete.Hurd 17:14, 29 January 2006 (UTC)

I tossed the images around, the first large piece of text mentiones several time twin studies, so haveing one of the twin study images there would be a nice illustration. It also solves the problems later with to many cramped up iamges. --KimvdLinde 23:09, 4 February 2006 (UTC)

[edit] Estimation methods -> quantitative genetics?

I do not know how other feela bout it, but I think the estimation methods might be better served in the Quantative Genetics article. Reason for that is that the methods are equally usefull for the estimation of genetic correlations, which falls outside the scope of this article? What about moveing those section? --KimvdLinde 23:30, 4 February 2006 (UTC)

I think, in their current states, it fits more in Heritability than quantitative genetics, but I see no reason to make it either or at this point. I can't tell where the quantitative genetics article is going, it's possible that the two articles are future merge candidates. I suggest letting them each grow as seems appropriate and see what subject matter gets duplicated down the line. Pete.Hurd 03:14, 5 February 2006 (UTC)

Thea reason I asked is that I am busy with QG and I am going to add estimation stuff. Heritability is just a term based on some aspects of the estimation methods. We use more variances and covariances anyway. But I leave it for the moment as is, and we see when it is growing at QG. --KimvdLinde 03:21, 5 February 2006 (UTC)

I have taken away merge into quantitative genetics. It has been over a month with a vote of 1 yes, 2 no -- hardly an enthusiastic response of any kind. Ted 13:25, 30 March 2006 (UTC)

[edit] Question

Hi, I have a question regarding the methods of calculating heritability for Kim. I would like to comfirm that in order to calculate the heritability, the data have to have some relaives structure, like parents and offsprings, siblings, etc. Since, recently, someone told me that these relationship is not neccesary. I am confused . So please help me to clarify it. Thanks. Peter Hanon.

I'm definitely not Kimvd, but here is the answer anyway. Yes, you need some data from relatives. Basically, that is the only way to get a handle on genetic variation vs. environmental variation. Ted 00:43, 1 April 2006 (UTC)

[edit] Neutrality (?)

Is this statement neutral? It seems like the writer has already made the decision that intelligence doesn't have high heritability, regardless of the study: "Much the same goes for intelligence tests. The conclusions from studies involving intelligence tests often conclude that intelligence has high heritability. This is probably due to inherent problems with human twin studies, as well as reflecting a high level of genetic variation for many human traits, and corresponding lower environmental variation within the confines of the test." I can almost hear "Once those are taken into account surely the heritability of intelligence will be shown to be lower than these studies mistakenly conclude."—The preceding unsigned comment was added by 69.123.249.121 (talk • contribs).

Actually, the definition of high heritability is a large genetic variance compared to a relatively smaller environmental variance. How would you change it? Possibly, a more expanded version would be better, but this is an article on heritability, not intelligence studies. Heritability is quite plastic for just about any quantitative trait, which also means it is fairly useless for how some (many? most?) people would like to use it. Ted 02:41, 22 April 2006 (UTC)

[edit] Additive Effect and Dominance Deviation

Relationship of phenotypic values to additive and dominance effects using a completely dominant locus.

I had hoped to avoid it, but maybe the concepts of additive effects and dominance deviation needs to be made more clear. I just made the illustration, so it is easily changed if need be. The accompanying text might be something like:

The simplest genetic model involves a single locus with two alleles that effect some quantitative phenotype, as shown by + in the figure. We can calculate the linear regression of phenotype on the number of B alleles (0, 1, or 2), which is shown as the Linear Effect line. For any genotype, the phenotype can then be written as the sum of the overall mean, a linear effect (a), and a dominance deviation (d). The additive genetic variance is the weighted average of the squares of the additive effects:

with a similar relationship for variance of dominance deviations.

I'm not sure it would be that useful, although there is something to spelling it all out.

Comments? Ted 15:06, 4 May 2006 (UTC)

I have one. To me, {bb,Bb,BB} is a categorical variable. It's a bit heavy-handed to treat it as a continuous predictor, and furthermore it's quite arbitrary to suppose Bb should lie halfway between bb and BB in this respect. For example, it could be that Bb is more hearty than both bb and BB. Btyner 14:49, 29 June 2006 (UTC)

Thanks for the comment. There is a difference between Bb being halfway between bb and BB and y_Bb being halfway between y_bb and y_BB. In fact, the diagram shows a situation where B is completely dominant (or, nearly so). The X-axis is really the number of B alleles. In this way, bb=0, Bb=1, and BB=2, which is the way the additive effect is defined. I'll see if I can come up with a way to make that more clear. Ted^Talk/_{Contributions} 15:22, 29 June 2006 (UTC)

The model for the additive effect is: y_ij = a_i + a_j + d_ij + error, not the usual statistical model. Ted^Talk/_{Contributions} 15:25, 29 June 2006 (UTC)

[edit] So... perhaps the introduction can be clearer

Perhaps I am just daft, but does a high heritability correlate to significant or insignificant genetic factors (as opposed to environmental factors)? A clear and plain English statement in the article's introductary paragraph might be helpful to casual "link browsers" like myself. BigNate37_T·C 06:38, 17 July 2006 (UTC)

That question is unanswerable, in part due to the plasticity of heritability itself. Ted^Talk/_{Contributions} 11:55, 17 July 2006 (UTC)

Uhh... I don't think so... from the article, "In genetics, heritability is the proportion of phenotypic variation in a population that is attributable to genetic variation among individuals." Now, my question could be restated to ask "Does this mean hertiability is proportional to phenotypic variation in a population that is attributable to genetic variation among individuals, and does this imply that heritability is inversly proportional to phenotypic variation in a population that is attributable to environmental variation among individuals?" Now, since I asked this question I've read enough from other articles to tell me that this statement is more of less true, so I'm really only asking whether we can't make the introductory paragraph more meaningful to the lay person. BigNate37_T·C 14:36, 17 July 2006 (UTC)

My apologies. I think I see what you meant to ask. In the simplest model, H²=VG/(VG+VE). The answer to your question is no. The relationships are not proportional. However, as VG increases and/or VE decreases, then H² increases while as VG decreases and/or VE increases, the opposite happens. Is that what you meant? Ted^Talk/_{Contributions} 21:22, 17 July 2006 (UTC)

Er, perhaps if I knew what VG and VE were. But I think barring that confusion, you've answered my question. BigNate37_T·C 21:40, 17 July 2006 (UTC)

I think I get it... if this is correct, then I'm satisfied I know what I wanted to know about heritability:

• as environmental variation's influence increases, heritability approaches zero;
• as genetic variation's influence increases, heritability approaches one:

Thus, in the limiting cases where environmental or genetic variation are the sole factor, heritability approaches zero or one (respectively)? BigNate37_T·C 21:46, 17 July 2006 (UTC)

[edit] request for comments

On race and intelligence, please [1] Slrubenstein | Talk 13:16, 31 January 2007 (UTC)