[R] anyone know how to calculate chi square value from P value?

[R. A. Bilonick]

On Mon, 2009-05-11 at 12:36 -0700, Anyuan Guo wrote:
> Dear all,
>     I have P value of a list of markers. Now I need the chi square value 
> with degrees of freedom 2.
>     I noticed there are several Chisquare functions (dchisq, pchisq, 
> qchisq, and rchisq), but it seems all are not for my purpose.
>     In microsoft excel, there is a function CHINV to do this, such as 
> CHINV(0.184, 2) is 3.386, that means the chi square value for P value 
> 0.184, degree 2 is 3.386.
>     Does the R package has some function to do this?
> 
> Thanks
> 
> Anyuan

It never hurts to read the documentation:

?qchisq

for example. Here "q" stands for "quantile." You want the Chi-square
value that has 0.184 area to the left. To compute this:

> qchisq(1-0.184,2)
[1] 3.385639

You have to subtract 0.184 from 1, because qchisq finds the value with
the given area below it. To double check the result:

> 1-pchisq(3.385639,2)
[1] 0.184


Rick B.