I has a strong undeniable correlation to height.
Did you have any scientific paper able to prove that some specific I-related SNPs are connected to height ? I mean real studies done at the SNP-level and controlling that the correlation is not induced by other shared traits.
To clarify the reasons behind my request :
An observed correlation didn't imply the existence of a causal relation.
A well known exemple is the anti-correlation between Pirate number and Climate evolution.
Do you think that this correlation proves a causal relation ?
It's true. This extremely scientific graph proves it: Photo via http://bama.ua.edu/ You can see that as the number of pirates in the world has decreased over the past 130 years, global warming has gotten steadily worse. In fact, this makes it entirely clear that if you truly want to stop global...
www.forbes.com
Now regarding the claimed correlation about height and haplogroup I, yes it exists (but it depends how you look at it),
1) Is it that significant ?
If I look at populations with "low" (< 30%) I-fractions, there is no clear correlation in fact.
2) The correlation is carried by Scandinavia and north-western Balkans.
First, Balkans and Scandinavia exibit different kind of haplogroups, I2 and I1 respectively.
Haplogroup I is defined by some ~200 SNPs on YFULL, if you have an effect shared by the two branch I1 and I2, then it has to happen on one/some of these ~200 SNPs.
A more scientific analysis, particularly by 2019, would have been to not select defferent world population, but to select individuals from the same population carrying different haplogroups ! Is a German I-carrier significantly taller than a German R-P312 carrier from the same region & same social class ?
Why such test would be needed ?
The major reasons is that you can't consider that populations on the European continent are sharing all their DNA aspect but the Y-DNA. Therefore, when performing a correlation analysis, to evaluate if the correlation is really between the two variables you study you need to control all other variables (as much as possible). In the study I quoted, it is definitely not the case.
Therefore, from their data, it is impossible to know if the correlation is real, or if the correlation is in fact an artefact arrising from the existence of another genetic/environnemental parameter that is driving the height and that is also geographically locallised around north-western Balkans and Scandinavia.
To be convincing such analysis have to made a the level of SNP (with as much of the other SNPs being identical) on a large amount of individuals ... not with population averages ... The guys who did that have little to no knowledge about statistical analysis. If I was the referee, such paper would have been refused (after, I used to be known as a pain in the ass for the authors when they had the bad luck to fell on me as a referee).
3) What happen when we zoom on geographical regions ?
-For Scandinavia alone (Island, Denmark, Norway, Sweden) ... in fact, it shows a slightly anti-correlation of height and I-haplogroup.
-For Western-Balkans alone (Montenegro, Serbia, Croatia, Bosnia) it also shows an anti-correlation of I-fraction and height.
-When looking at ancient URSS (Russia, Bielorussia, Ukrainia, Latvia, Lithuania, Estonia) and adding Poland, Romania, Bulgaria for the fun ... there is a huge anti-correlation signal.
If the signal disapear (or can even be inverted) when you zoom on local regions it is a very a bad indication ... sounds like if you go at the scale of the individual carrying I-haplogroup, you will just see nothing.
4) Germany is scoring as high for height than Denmark or Norway despite having way less I-carriers.
Latvia and Estonia are also scoring very-high as all countries around the Baltic-sea, while they have relatively low I-fractions.
Therefore, it seems that nearby countries either sharing ancestry histories or environnemental conditions have a similar height average independently of the I-fraction.
To me this figure proves the exact opposite of the author claim.
Anyway, this kind of study is totally unable to prove their claim ... to do so, they would need to do it at the scale of SNPs.
An exemple of a way better work, while significantly older :
Strong signatures of positive selection at newly arising genetic variants are well-documented in humans[–] , but this form of selection may not be widespread in recent human evolution[] . Because many human traits are highly polygenic and partly ...
www.ncbi.nlm.nih.gov
I will conclude by that : a correlation is not the proof of a causal relation, particularly if you have many other variables out of control in your analysis.