How to write a faster seqrcomplement function

Hi,

I noticed that whenever I use the function seqrcomplement, the line that calls this function takes about 80-90% of the total running time.

If I use

SRC = int2nt(5 - nt2int(fliplr(S)));

instead of

SRC = seqrcomplement(S);

then this reduces the time to half, but it is still slow if I need to process millions of sequences. Is there a much faster alternative to seqrcomplement ?

Thanks!

Answer 1

[Roberto]

I really don't know if there's a faster alternative, but if you have millions of sequences try parallel computing...

% parallel computing

matlabpool(8) % number of workers! Setup first
parfor i = 1 : 1000000
    % your code here!
end
matlabpool close ; % end workers

there are a few considerations that you might take, I'm not an expert but you can try this...