[Seqinr-forum] zscore() function in SeqinR

[Coghlan, Avril]

Dear all,

 

I am wondering would it be nice to make a small addition to the zscore()
function in SeqinR, so that it computes a P-value for the Z-score?

 

As far as I understand, this would involve adding a line to calculate
the P-value based on the Z-score (where zscore is the variable holding
the Z-score):

1 - pnorm(zscore, mean=1,sd=1)

Is this right?

 

Would you consider adding this to the zscore() function? I think it
would be nice if it gave a P-value.

 

Also, I am wondering would it be nice to extend the rho() and zscore()
functions so that they work for word sizes of greater than 2 ?

I think this would be easy to do for the rho() function. I can provide
some suggestion for code if you are interested.

For zscore(), I think the easiest way would be to calculate a P-value
would be to use simulated sequences to calculate the expected
distribution of Z-scores according to some null model (eg. a multinomial
model for DNA).

What do you think? Would you like me to send my suggestion for some
code? I have written part of it already. 

 

Regards,

Avril

 

Avril Coghlan

University College Cork, Ireland

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.r-forge.r-project.org/pipermail/seqinr-forum/attachments/20100119/340a638d/attachment.htm