Exonerate protein2dna:bestfit problem

[woa]

Hello I'm having a problem to run the "protein2dna:bestfit" model in my 64 bit computer and under both Linux Cluster and Cygwin.
For the Linux Cluster I've tried both the readymade executable and compiled code.

The program just quits without producing any output.

However it runs fine with my old 32 bit Ubuntu.

Did anybody face the same problem and any suggestions to solve it?

I'm approaching a deadline fast, so any help is much appreciated

Thanks in advance

[SES]

You may not be using the appropriate settings for your data. Did you get an alignments when you use the basic protein2dna model (without the 'bestfit')?

[woa]

Hello SES, Thanks for your answer. Yes, I'm getting perfect alignment when using just 'protein2dna' without the bestfit option.

Some details on the version, the command I used and input/output:

exonerate from exonerate version 2.2.0
Using glib version 2.22.5
Built on Aug 26 2013

The simple command I used :

./exonerate -m p2d:b -q test_protein.fasta -t test_dna.fasta --exhaustive yes

The program just quits without producing any output as follows:

** (process:1922): WARNING **: Exhaustively generating suboptimal alignments will be VERY SLOW
-- completed exonerate analysis

The Pastebin single Protein and DNA file links are here:

http://pastebin.com/ex1BhJHF

http://pastebin.com/zWLYKaw3

Thanks

[woa]

Also note that I used both the readymade executable version and also compiled it on my own machine. I've used two sources for compilation. From the Ensemble CVS and also the tarball available from EBI.

https://github.com/nathanweeks/exonerate

[SES]

In that case, I would just use the results you have with the basic protein2dna model. Is there a reason you think the bestfit model needs to be used? Note that with bestfit the entire protein must match (IIRC), and it could be that this is not possible with your data, so you get no alignments (when specifying the bestfit model).

[woa]

I've used the same sequences in my 32 Bit Ubuntu to get proten2dna:bestfit alignment. The bestfit alignment is longer than the non-bestfit(as expected, as it aligned the whole length) of protein. I would like to get a global alignment to the protein, thats why I'm using it. If I could tweak some parameters for non-bestfit, so that the alignment is global-ish and close to bestfit then thats fine too. But i'm not sure which parameters( like gap-penalties may be) to tweak and to what extent.

[SES]

There is also the affine:global model, for doing global alignments. Though, I'm not sure you can do protein-dna alignments with that model. It sounds like you got the protein2dna:bestfit working now? If not, I think you can just specify that you want exonerate to do exhaustive alignments and it will find the best possible.

[woa]

No, protein2dna:bestfit is still not working