I'm reading the paper "Large-scale GWAS reveals insights into the genetic architecture of same-sex sexual behavior". In Fig. 4 shows "Genetic correlations of same-sex sexual behavior with various preselected traits". This sentence confuses me. I understand that you could do a standard statistical analysis to find correlations between same-sex sexual behavior and various traits. What makes the way they are doing the correlations "genetic"?

(Note: This may be more of a statistics question than a bio question. Feel free to move it to a more relevant SE site.)

  • $\begingroup$ They cite reference 16: nature.com/articles/ng.3406 might be useful to you. I think the question is relevant here, we have some regular users who are well-versed in heritability studies. $\endgroup$ – Bryan Krause Sep 8 '19 at 23:00
  • 1
    $\begingroup$ I believe this may be more of an English language question than a stats or a bio question. I shall check the article once to verify. $\endgroup$ – WYSIWYG Sep 9 '19 at 9:15
  • $\begingroup$ Looks off-topic? to me, this is an english language question, not a biology one, correlations are correlations are correlations, the addition of the word genetic is merely there to highlight the data set / type under current reference in the sentence & nothing more. $\endgroup$ – Pelinore Sep 9 '19 at 10:28
  • $\begingroup$ I disagree that it's about English language. The question is really "I understand the meaning of a correlation between phenotypes A and B; what is the meaning of a genetic correlation between A and B." That is, how do you have a "genetic correlation" between Same-Sex Sexual Behavior and Depression? $\endgroup$ – Bryan Krause Sep 9 '19 at 16:42
  • $\begingroup$ @BryanKrause : You've convinced me, where b4 I only thought it likely I'm now sure that this is definitely a language comprehension issue & not a legitimate biology question :) off topic then. $\endgroup$ – Pelinore Sep 9 '19 at 19:50

It is simply a correlation where one of the two variables considered is the genetic state at a given locus. In a GWAS, one loops through all loci and systematically performs a correlation analysis between the genetic states at this locus and the phenotypic trait of interest.

In the example of sentence you cite

Genetic correlations of same-sex sexual behavior with various preselected traits

They looped through each polymorphic locus in their data set and at each locus they performed a correlation analysis between the two variables "genetic states at a this locus" and "same-sex sexual behavior" (whichever way this second variable is recorded).

A study might also be interested in more than one phenotypes. In such case, different phenotypic variables can either be considered separately or can be included in the same analysis through linear models. It is for example not so unusual to include one of the variable as a covariatiate in an ANCOVA type of design.

  • $\begingroup$ I think this answer is not quite complete (in the context of the paper OP is referring to)...I'd have to look closer but it seems there is an additional step, because they are talking about two traits and a genetic state, not one trait. I still think you are much better positioned to write the complete answer here, though. :) $\endgroup$ – Bryan Krause Sep 8 '19 at 23:51
  • 1
    $\begingroup$ I am too lazy to read the paper :D I added a little paragraph about cases where more than one phenotype is considered. $\endgroup$ – Remi.b Sep 9 '19 at 1:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.