Glimmer: Multifasta analysis

[Daniel Fernandez]

Hi everyone.
I am new user at glimmer. We have a bacterial genome in multifasta, contigs. We can predict the ORF, if we could a complete genome in one fasta, we could, but we with Multifasta, we can make a a lot of things and I'm pretty sure that we're wrong.

First, we separate the Genome.fas in 128 fasta files (beacuse it has 128 sequences), for make a *.train file we create a lot of *.train for any contig and put them all together in one file run.train

long-orfs -n -t 1.15 Contig1 run1.longorfs && extract -t Contig1 run1.longorfs > run.train
...
long-orfs -n -t 1.15 Contig128 run128.longorfs && extract -t Contig128 run128.longorfs >> run.train

then we create a *.icm model

build-icm -r model.icm < run.train

but I compared it with other analysis of my friend, for he, all *logorfs are the same name, so he replace in any line this file. But my friend and my, we can the SAME NUMBER OF ORF!!!.

long-orfs -n -t 1.15 Contig1 [B]run1.longorfs[/B] && extract -t Contig1 [B]run1.longorfs[/B] > run.train
...
long-orfs -n -t 1.15 Contig128 [B]run1.longorfs[/B] && extract -t Contig128 [B]run1.longorfs[/B] >> run.train

build-icm -r model.icm < run.train

So, my friend and my, we can the SAME NUMBER OF ORF!!! although we did diferent things. I'm pretty sure that we're wrong.

So. How do you analyse MultiFasta with Glimmer?
And how do yu extract the fasta from *.predict file in mulifasta?

Please, healp.