AllPaths-LG

[scalabrin]

I am trying to force AllPaths-LG to use my paired-end data with "standard" insert size of about 300bp.

As per documentation it does fail. But is there a way to fool AllPaths and pass it fragment and jump libraries even though I do not have them?

An idea I had was to run a quick assembly with any assembler (e.g. ABySS) and them map all my reads on it and grab those aligning in a short range and put them in a fragment library file and those aligning at long distance and put them in a jumping library. But I am pretty sure it will fail. Before I try this complicated way, is there any alternative and simpler way to do it?

[scalabrin]

Anybody able to help me?

Is there any way to cheat AllPaths-LG to make it work with short insert (300-800bp) libraries?

[krobison]

Out of curiosity, why are you trying so hard to get Allpaths-LG to work outside its design envelope, given the number of other assemblers out there?

[scalabrin]

Simply because AllPaths turned out to be one of the best assemblers in Assemblathon1 and in GAGE!

But I also had some experience with other assemblers and I am not that exicted either in terms of usability or in terms of results.

[Wallysb01]

It was only one of the best assemblers because it was functioning how the creators intended. "Fooling" Allpaths to run would probably lead to less than ideal results. Something like SOAP or ABySS, which are more flexible, will probably work better than a coerced Allpaths.

And if you don't like the other's results, are you sure you the problem isn't with your coverage or library types? I don't mean to sound like a jerk, but it is not as if Allpaths is that much better than any of the other assemblers out there.