FASTQC help

[arkal]

I'm not a java programmer so I need some help doing something I hope there's someone here who can assist!

According to the fastqc documentation, for calculating duplication levels it considers only the first 200,000 reads (to be memory efficient).... Memory is not an issue for me. I want a comprehensive output on my seq duplication level. Can anyone help me in modifying the module which does the above said task to accept and process all reads?

Reason: I want to compare the results to those from picard's ELC.

Warning: Is FASTQC open source? Can it be modified without permission? It's GPL v3 released so i think it can be modified openly

Thanks in advance!

[simonandrews]

I'm not a java programmer so I need some help doing something I hope there's someone here who can assist!

According to the fastqc documentation, for calculating duplication levels it considers only the first 200,000 reads (to be memory efficient).... Memory is not an issue for me. I want a comprehensive output on my seq duplication level. Can anyone help me in modifying the module which does the above said task to accept and process all reads?

If you want to change this limit the value you need to change would be the line:

private final int OBSERVATION_CUTOFF = 200000;

..in OverRepresentedSequences.java. Or you could just remove the check all together by deleting the later lines:

if (uniqueSequenceCount == OBSERVATION_CUTOFF) {
  frozen = true;
}

You'll probably also need to change the default memory allocation in the launcher since you'll probably run out of memory in the 250MB allocation which is provided by default (edit the fastqc wrapper and increase the 250 to a much larger number).

Reason: I want to compare the results to those from picard's ELC.

Even with these changes you probably won't get the same answer as ELC as they're not measuring the same thing. The headline value in FastQC is the percentage of sequences which are not unique, which isn't quite the same thing as the percentage of different sequences. I'd prefer to have the latter value in FastQC but inferring that from the sample which we take has proved to be tricky (though I think we now know how to do this for a future release).

Warning: Is FASTQC open source? Can it be modified without permission? It's GPL v3 released so i think it can be modified openly

Yes, FastQC is open source and you can modify it for your own uses with no restrictions at all. The only restriction from the GPL is that if you distribute a modified version of the program you need to release the source for that under GPLv3 as well.

Hope this helps

Simon.