Simulating 454 reads with ART

[mdk308]

Jenzo - the author's response is that yes, you can use indel_error_profile as is, and also that a new version will be coming out soon. I believe that this new version will contain scripts to generate the other three files. I pass on the download link on this post as soon as I find out that it is available.

[Jenzo]

Yes, you're right, three files can be created without reference. Perhaps taking indel_error_profile as it is will suffice your needs..
Thanks for contacting the authors.

Perhaps you can have a look at 454sim, you can tweak their profiles, too.

[mdk308]

Well for three of the files (length_dist, qual_1st_profile, and qual_mc_profile), you don't need a reference as these are based purely on quality scores and have nothing to do with the correctness of the base call (unless i am understanding this wrong). The other file (indel_error_profile) sounds like the only one that would require a known reference but I would actually forego this if I could get the other three to work by themselves.

Jenzo, have you tried any other simulators? I am looknig for something that will include a realistic quality score profile, so I figured this program would be perfect since it could mimic the exact error profile of my own sequencer. Let me know if you are aware of any other sim programs that might be able to do this.

I will try emailing the contact listed on their web page and let you know if I figure anything out.

[Jenzo]

Hi mdk308,
same here, I would love to try ART in a pipeline we're developing at the moment.. I suggest, that you have to sequence something known (perhaps an artificial piece of DNA) and then analyse the reads with respect to their reference. A pity that there is nothing more about how the authors created these files, since they needed data as basis, too?

[mdk308]

Hi, I am trying to use ART to simulate 454 reads. I chose this program because it looks like you can use an experimentally obtained read set to create a quality profile, which will then be used to generate the simulated reads. The end result is that you have reads which mimic the error profile observed on your own sequencer.

The issue is that I am not sure how to generate these files, and I don't see any documentation as to how to go about doing this. (There is a sample profile provided but I'd have no idea how to generate these files manually).

If anyone is familiar with this process, I'd appreciate any advice you might have.