Re: [modeller_usage] Selection criteria for building sequence database - pdball.pir

[Modeller Caretaker <modeller-care@ucsf.edu>]

Praveen Raj wrote:
> I'm trying to rebuild the pdball.pir file from a set of pdb files, using 
> the model.make_chain() as mentioned in modeller website and I noticed 
> certain filter criteria as parameters to make_chain() such as
>
> minimal_chain_length=30
> minimal_stdres=30
> chop_nonstd_terminii=True
> max_nonstdres=10
> minimal_resolution=99.0
> structure_types='structureN structureX'
>
> And while processing I found that certain pdb files are skipped due to 
> this filter criteria. My query is
> 1) Are these the standard criteria for building pir file?

They're what we use internally to build the database, yes.

> 2) Or adjusting the values to 0 (except min. res) will alter the purpose ?

You can certainly change these values if you like. The filtering is 
there simply to remove "bad" structures, such as those that are not 
proteins, are very short chains, or are at very low resolution.

	Ben Webb, Modeller Caretaker
--
modeller-care@ucsf.edu             http://www.salilab.org/modeller/
Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage