Error while running autoAug.pl

[katharina]

Originally posted by tnguyen in the old forum on 09.10.2012 - 10:03

Dear all,
I am new to AUGUSTUS, and would be grateful if you could help.
I have a genome of a non-model organism and would like to predict genes from the genome.
1. I have downloaded EST database from the same species and closely related one, and created the 'hints' file hints.est.gff after cleaning using seqclean.
2. I have downloaded protein database from closely related species: protein.fasta
3. I run autoAug.pl using the following command:
perl autoAug.pl --species=mysepcies --genome=scaffold.fa.masked --hints=hints.est.gff --trainingset=molluscs_protein.fasta --estali=est.f.psl
and this error occured:
Program aborted. Possibly "scipio" is not installed or not in your PATH at autoAugTrain.pl line 780.
How can I overcome this problem? I did install 'scipio'.
Your help will be much appreciated.
TN

[katharina]

Originally posted by katharina in the old forum on 09.10.2012 - 20:18

Did you add Scipio to your path variable? Please add the path to the scipio executable to your PATH, e.g. in the .bashrc or in your .profile file. If you keep on working on the same shell after adding it, you need to reload the .bashrc or profile, e.g. with "source .bashrc". If you open a new shell, you don't need to do that.
Did this solve your problem?
Katharina

[katharina]

Originally posted by tnguyen in the old forum on 10.10.2012 - 00:08

Thank you Kathrina,

I did add the PATH as follows:
PATH="/group/aquaculture/mussels/sequencing/programs/Scipio/scipio-1.4:$PATH"
export PATH

and it is not working still.

[katharina]

Originally posted by katharina in the old forum on 10.10.2012 - 21:32

In this case, I recommend that you submit your data to our web service at http://bioinf.uni-greifswald.de/webaugustus
Submit the genome and protein file. After the jobs finished, you'll be able to download a parameter archive that you can use to execute augustus on your target genome using your already generated hints file.
Alternatively, you could submit protein, genome and EST file. The webservice will then automatically re-generate your hints and run predictions with hints from EST data.
Best,
Katharina

[katharina]

Originally posted by tnguyen in the old forum on 15.10.2012 - 14:48

Thank you Katharina.
T

[katharina]

Originally posted by adnguyen in the old forum on 10.01.2014 - 15:20

Hi Katharina,
Could you please give me some advice how to train my data using autoAug.pl script (I prefer not to submit my data to your web service). I have the following data of a genome:
- ESTs
- RACE
- proteins
- RNA-Seq and Trinity transcripts
The autoAug.pl of the 3.0 version still uses the previous version of PASA, therefore I plan to run PASA separately, and then feed the output of PASA in gff-formatted file into the autoAug.pl.
Thanks for your help.

[katharina]

AutoAug.pl will not be able to use all that evidence, directly.

You could feed ESTs (or proteins) for training and convert all other evidence to hints, manually. You might want to modify the extrinsic.cfg file to your needs.