For (a) the vast majority of variation has low frequency, and the majority of traits are multi-locus (if they are the opposite: high frequency and single locus with effects - it will be obvious - things will run in families very clearly). Rule of thumb - you need 1,000 to 5,000+
-
-
Replying to @ewanbirney @Steve_Sailer and
For (b) using skin colour to genotype people is about 30-50 odd loci around the genome, out of a notional need for about 300,000. Looking at people is just plain madness as a genotyping strategy. Very best case: you can do association to some aggregate signal of this 30 odd loci
7 replies 2 retweets 7 likes -
Replying to @ewanbirney @Steve_Sailer and
In addition, using the extreme phenotype of "getting into the 100 meter final" is a bad way to measure the continuous variable you are after (I'm imputing alot here). For starters you need the other side of this >>
2 replies 1 retweet 3 likes -
Replying to @ewanbirney @Steve_Sailer and
ie, your "control". This could be "everyone who doesn't get into the 100m final" in which case I will encourage you to recruit Usain Bolt's cousins into your study and try to work things out from there.
2 replies 2 retweets 4 likes -
Replying to @ewanbirney @Steve_Sailer and
Finally, with (c) you have to go on blind hope your assumption is right that the genotypes you measure are random with the complex environment of "becoming a 100m Olympic athlete".
2 replies 1 retweet 3 likes -
Replying to @ewanbirney @Steve_Sailer and
(because you have decided to measure genotypes in this really cookie, weird way of "looking at someone" you dont even have the GWAS test of this assumption in the QQ plot or other techniques. You just have to hope).
3 replies 1 retweet 3 likes -
Replying to @ewanbirney @Steve_Sailer and
Now, let's say we wanted to assess 100m sprinting genetics. It's a valid trait. What would be a *good* study design?
7 replies 1 retweet 4 likes -
Replying to @ewanbirney @aylwyn_scally and
Here's a good study design for race of 100m sprinting ability: worldwide competitions in 100m dash. Those who run the global qualifying time get invited to Olympics. They run 2 or 3 preliminary rounds, then fastest 8 men race for Gold. Winner gets $30 mil per/yr in endorsements.
2 replies 4 retweets 43 likes -
Replying to @Steve_Sailer @ewanbirney and
Almost like a natural experiment
3 replies 1 retweet 18 likes -
Replying to @clairlemon @Steve_Sailer and
Except it's not: you're looking across different countries/regions with very different social histories, environments etc. Many (non-genetic) things not being controlled for. This is not a scientific approach at all. Do you think you understand this topic better than Ewan et al?
4 replies 0 retweets 3 likes
In that Steve understands it at all and Ewan has a negative understanding, yes, he clearly does
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.