rkwalters
commented
7 years ago Hello, Just like with genetic correlation the LD score estimate of heritability isn’t bounded. So if the true heritability is low, that value plus some sampling error can lead to a negative estimate (especially if the sample size is small and thus there’s lots of sampling variation). Unless there’s something unusual about your analysis that makes a negative estimate likely to be a technical artifact there isn’t much you can do to “fix” this. It doesn’t really work to talk about genetic correlation when you don’t have evidence for heritability anyway.
As far as the definition from the methods, I’m guessing you are referring to the definition in terms of best linear predictors? Note that definition is in terms of population parameters, so directly computing that value will be largely attenuated by sampling error. It also depends on a joint model, where the effect of each SNP is conditional on all SNPs rather than the standard marginal effect estimated in a GWAS, which is non-trivial to fit jointly for all SNPs.
Cheers, Raymond
On Aug 29, 2017, at 1:55 AM, chenyan53535 notifications@github.com wrote:
Hi, I am trying to use your software to calculate the genetic correlation between phenotype1 and phenotype2, but the estimated heritability of phenotype1 is negative, thus I cannot get the genetic correlation between phenotype1 and phenotype2. why the estimated heritality could be negative? and how to deal with this problem? Thank you. Another question is whether we can calculate the genetic correlation just according to the defination of genetic correlation in your online methods, it is more simple and does not occur the negative heritability? Need your help.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/bulik/ldsc/issues/88, or mute the thread https://github.com/notifications/unsubscribe-auth/AILEvTJrW_tmOgfWmpSCZF1lIP2L6SRCks5sc6e5gaJpZM4PFaaE.
Hi, I am trying to use your software to calculate the genetic correlation between phenotype1 and phenotype2, but the estimated heritability of phenotype1 is negative, thus I cannot get the genetic correlation between phenotype1 and phenotype2. why the estimated heritality could be negative? and how to deal with this problem? Thank you. Another question is whether we can calculate the genetic correlation just according to the defination of genetic correlation in your online methods, it is more simple and does not occur the negative heritability? Need your help.