Problems about the gff3_to_zff.pl

[Huangyizhong]

Hi there,
I want to use the SNAP to annotate the mammalian genome. Because there are no trained file about the mammalian, I want to train it by myself. How about the train the human genome for SNAP for us to use , I think it will use widely. By the way, no response about the script: gff3_to_zff.pl, how to construct the zff file correctly?
perl ../SNAP/gff3_to_zff.pl final_annotation.gff >genes.zff
Need help!
Thanks advance!

[iankorf]

Most mammalian genomes are so similar to each other that you can use the same parameter file for all of them. Also, there's so much experimental data that gene prediction probably isn't very useful. There is already a mammal parameter file in the distribution, but that was produced long ago so maybe it's a good idea to train a new one. The ZFF format is described in the README.