Hi and welcome to the forum.
Try to break your problem down into simpler sub-tasks. E.g: you want the top 10 counts, so it would make sense to sort your input files first:

Code:

sort -n -k2,2 -r file1

will do a numeric (-n) sort descending (reverse -r) on second field (-k2,2). Now to find the top 10, you just need to look at the first 10 lines.
So I'd approach this with feeding the sorted files into awk:

Code:

awk '#do the hard work' <(sort -nrk2,2 file1) <(sort -nrk2,2 file2)  <(sort -nrk2,2 file3)

Now to pull the maximum of the top 10 from each input, you could do something like:

Code:

awk 'FNR<=10{  #I only care about the first ten lines in each file
  if($2>cnt[$1])    #get the global max among the files
    cnt[$1]=$2
}
END{
  for(i in cnt)
    print i "  " cnt[i]
}' <(sort -nrk2,2 file1) <(sort -nrk2,2 file2)  <(sort -nrk2,2 file3) >output.txt

output.txt should now contain something like:

Code:

chr_1_1_50  100
chr3_101_150  65
chr1_1_50  120

(in random order, since 'for(i in cnt)' doesn't sort anything).

I don't quite understand what do you mean by
or what is your desired output. But if you take it one small step at a time, you're gonna eventually get there.
E.g. you could read the lines from output.txt and grep for the name in the input files to get the other values:

Code:

while read name count ; do 
  grep $name file1 >> globalTop10inFile1.txt
done < output.txt

etc.

Approaching the problem in this step-by-step fashion, it's much easier to debug -- you can verify the intermediate results easily.
Give it a shot and let us know how it goes!

mirni