0
$\begingroup$

I have a Genomic data that shows the interaction between genomic regions that I would like to understand which interactions are significant statistically.

Dataset look likes:

chr  start1   end1   start2   end2   normalized count
 1     500    1000   2000     3000       1.5
 1     500    1000   4500     5000       3.2
 1     2500   3500   1000     2000       4

So, I selected a random number of data (as background) and fitted the normalized frequency into the Weibull distribution using fitdistrplus R packages and estimated some parameters like scale and shape for those sets of data (fit.weibull= fitdist(data$normalized count,'weibull')).

Now I would like to calculate the probability of each observation (like a p-value for each data point) under the fitted Weibull distribution?

for example:

if I have 100 data point with normalized count like:

 xs <- seq(10, 65, len=100)

How can I calculate the probability of those set?

  • Can I use one of the below functions? and which one is better?

       pweibull(xs , fit.weibull[shape], fit.weibull[scale], lower.tail = TRUE)
    
       qweibull(xs , fit.weibull[shape], fit.weibull[scale], lower.tail = TRUE)
    
  • or I should calculate CDF for each value (from those 100 point) then estimate p-value=1-CDF(x)?

  • is pweibull () and 1- CDF() are the same?

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.