Hello,

I have a piece of code that is shown below. The code is a nested for loop where the inner loop is dependent on the outer loop. The outer most loop "i" value loops for every chromosome. The "j" value loops for every cytoband in a chromosome. The "k" value loops for every sample.

for (i = 1: NumberOfChromosomes) {

    for (j = 1: NumberOfCytobandsInEachChromosome) {

      for (k = 1: TotalNumberOfSamples) {

         z_1 = NULL  #reset this value before the calculation for every k 

         # do something with z_1 to get z_2

         # do something with z_2 to get z_3

         x[k, j] = z_3 #store the output value into a matrix

      } # end of k loop

      y [[i]] = x

    } # end of j loop

} # end of i loop

 I would like to make this code faster and more efficient. Could anyone suggest a way to use one of the apply functions  on this ? I have used apply (lapply and mapply) before, but never on nested for loops, so not sure how to do this. 

Any help would be great. Thank you.

Firstly, your question is about general R rather than Bioconductor packages, so this site isn't the right forum for it.

Secondly, the code within the braces of each for looks wrong; you should replace the = with in.

Thirdly, check out Circle (Chapter) 4 of "The R Inferno". Briefly put, apply is just syntactic sugar for the for loop. The function still runs a for loop under the hood. So really, you won't get any speed increase from switching to apply. Depending on what your skill level is, and what the tasks are inside the nested loop, you might consider rewriting the loop in C/C++ with Rcpp. Of course, if the tasks to be performed are compute-intensive, then any speed loss due to looping is probably negligible in the wider scheme of things.

Thank you for the feedback. 

The reason I asked the question was because I have found significant improvement in speed when I switch from for loop to apply. And the reason I posted the question here because the application is bioinformatics and deals with chromosomes and cytobands.

I will try the pre-allocation with y = list(), and also look into implement this code with bplapply . 

Thank you.