search

stephenslab / susieR

R package for "sum of single effects" regression.
https://stephenslab.github.io/susieR
Other
169 stars 42 forks source link

is it acceptable to have pip=1 but posterior beta=0? #200

Open WeiCSong opened 9 months ago

WeiCSong commented 9 months ago

Hi, thanks for the great tool! In my analysis I found some variants that have PIP>0.99, but when using susie_get_posterior_mean() i got beta=0. I'm not sure whether this is reasonable or there is any issues with my data, do you have any suggestion? Thanks for your help!

gaow commented 9 months ago

@WeiCSong are you sure you did not misaligned the variants? What you said does not sound likely ...

BTW, if you want to get the regression coefficient, you should try coef(fitted) where fitted is output of susie() or susie_rss(). That will also give you intercept.

WeiCSong commented 9 months ago

Thanks for the suggestion! I'll check my data for this issue.

@WeiCSong are you sure you did not misaligned the variants? What you said does not sound likely ...

BTW, if you want to get the regression coefficient, you should try coef(fitted) where fitted is output of susie() or susie_rss(). That will also give you intercept.

WeiCSong commented 9 months ago

@WeiCSong are you sure you did not misaligned the variants? What you said does not sound likely ...

BTW, if you want to get the regression coefficient, you should try coef(fitted) where fitted is output of susie() or susie_rss(). That will also give you intercept.

Another small question, could coef() or other functions also extract the sd for regression coefficients, as get_posterior_sd? thanks!

gaow commented 9 months ago

I think get_posterior_sd is the only way for that purpose