Talk:Heritability

Figure 2 of Estimating Heritability Needs Revision
Shouldn't this diagram (Twin-concordances.jpg) visually reflect the numbers cited in the

Prominence given to GxE interaction assumption and violation results
User:Everymorning has added not just 1, but 3 recent edits basically repeating the same theme: when GxE interactions are present, interpretation of results from e.g. classical twin design is difficult. Now, everybody working in the field already knows this. Perhaps he can explain his vision for how he wants this point to be communicated in the page. Right now looks to me like he wants to insert it multiple places in the article using different references and slightly different wordings. That doesn't seem like an optimal approach. Why not mainstream textbooks or reviews as sources? This edit is particularly problematic as the cited source is wildly out of date (1990) and is by some seemingly randomly chosen critic (Wahlsten). This edit is problematic as it inserts an extreme minority view as being recognized fact in the page. General opinion in the field is that while large GxE interactions will cause trouble for interpretation, generally tests have failed to identify such large interactions. Plomin et al (2012) discusses GxE in chapter 8, p. 118ff. Their summary of research does not give the picture Wahlsten prefers, rather, they cite a lot of research done GxE done using traditional pedigree designs including twins and adoptees. Some of the research cited has since fallen prey to the replication crisis or later meta-analysis (e.g. the Morfitt GxE finding), but I don't have a copy of the latest version (2016, 7th ed.) of the textbook to see how it changed. Plomin et al does mention that many studies of candidate variant/gene GxE failed to replicate. Deleet (talk) 22:17, 1 July 2018 (UTC)
 * I have since added some other information similar to that which you reverted. Some of what I have (re)-added is based on the same sources you apparently consider inappropriate (e.g. Wahlsten 1990). It is "out of date", but that does not mean the conclusion regarding the validity of ANOVA (or lack thereof) in this context is invalid. The fundamental fact that heritability does not equal genetic determination and that there are serious concerns about the validity of heritability estimates in humans should certainly be included. Naturally behavior geneticists, e.g. in their textbook, will be dismissive of these concerns, but that does not mean that such concerns are not "mainstream". IntoThinAir (formerly Everymorning) talk  02:16, 13 July 2018 (UTC)
 * What is also very important to keep in mind here is that estimates of heritability are based on the assumption that genes and environment exert separable effects on phenotypes, but that is not true: their effects are entangled and inseparable. This recent review article seems an excellent source. Here is a relevant quote from it: "...there are no ‘genes for’ quantitative (i.e., non-discrete) phenotypes, such as common obesity and metabolic diseases (e.g., T2DM). As such, estimates of genetic heritability are misused and often meaningless statistical abstractions derived from attempts to impose an artificial and false dichotomy (i.e., nature vs. nurture) on demonstrably non-dichotomous biologic processes." IntoThinAir (formerly Everymorning) talk  02:17, 13 July 2018 (UTC)

Meta-analysis of the heritability of human traits based on fifty years of twin studies
Polderman et al. (2015): Meta-analysis of the heritability of human traits based on fifty years of twin studies. https://doi.org/10.1038/ng.3285

I'd guess this should be cited. — Preceding unsigned comment added by 131.130.153.43 (talk) 14:27, 24 May 2022 (UTC)