|
I am interesting in any personal tips or iternet links to detailed tutorials and documentations how exactly run wise2(genewise) and analyse its results. Could you share your own experience in wise/genewise? Or could anybody know somebody in Sthlm bioinformatcs area who possibly had such a practice? My simplest question to date is what was wrong in my first try? There is WISE2 gene prediction packet installed on UPPMAX (> module load wise2). I tried to run it but failed to get any output, niether results nor error messages. My input data are as follows. Genome.fasta = multiple fasta seq-s fie, dna, ~200Mbp, our model organism, insect; Proteins.fa = multiple fasta seq-s file, proteins, a set of Drosophila seq-s. I run genewise including several options for output formats and got no output files in working directory. I run program as a SBATCH job. >genewise Fly_proteins.fa genome.fasta -both -gff -genesf -sum & Thanks in advance |
|
As already have been mentioned, Genewise isn't really a gene finding program. Genewise is used to splice-align protein sequences to a nucleotide sequence (usually a genome) to get an indication where there might be genes and in particular to get support for introns. You can then take this information and combine it with gene models from gene finding programs such as Genemark with Evidencemodeler (EVM), to hopefully get the best of both worlds. If you still wish to use Genewise, if you want to run it on a multifasta-file you must use the variant program genewisedb, included in the gene wise package. The genewise binary will only search the first fast entry in your fasta-file. I would, however, like Dag above, recommend Exonerate that is easier to parallelize and does a similar job. That one is not installed on Uppmax so you would need to supply your own binary. If you provide more information about your project and what you intend to do with the gene wise results, we should be able to aid you further. |
|
Hey, I haven't actually used this program recently, but this software package was developed at the EBI by Ewan Birney and documentation, links, email etc can be found at http://www.ebi.ac.uk/Tools/Wise2/doc_wise2.html. GeneWise is the integral part of Ensembl gene prediction pipeline so should be updated regularly. However, the code may need some tweaking to run on UPPMAX cluster. Have you tried to install the package and run it on a desktop first? This would differentiate between potential problems specific to UPPMAX installation, and problems with may have to do with syntax, or current release bugs. with best regards, Lukasz PS. A BILS colleague suggested that GeneMark is currently popular among many users (http://exon.gatech.edu/).
(12 Sep '12, 17:29)
lucash ♦
Hi, I agree with Lucash if you are after a gene prediction program. I thought that you wanted to match protein sequences to a genome using GeneWise, which is a different use of genewise. Regarding gene finding, I have used GeneMark-ES (one of the GeneMark programs that is specialised for eukaryotic genomes) and it worked very well for fungi. It does not need a well annotated training set which can be a major problem to get/generate. There are a boatload of gene prediction software out there so more information about the project would be valuable before I can suggest alternative predictors... good luck with your project! Dag
(13 Sep '12, 12:44)
DagAhren ♦
|
|
Hi ale! Have you considered using Exonerate instead? It does basically the same thing as Genewise but much faster and is developed by the same group (Birneys lab at EBI, UK. Genewise is getting a bit old so unless you have a particular reason for running Genewise, I would suggest that you try exonerate. Check out this page: http://www.ebi.ac.uk/~guy/exonerate/ Hope that this helps! Dag |
