[phenixbb] autosol MIR with more than one native?

Terwilliger, Thomas C terwilliger at lanl.gov
Sun Feb 3 06:50:51 PST 2013


Hi Ed,

Here the output is a little confusing...

The "truncate_consensus_model" parameter tells hyss to trim down the heavy-atom solution to the expected number of sites that hyss has been given.

Separately autosol has guessed the number of sites and given that guess to hiss. For MIR data, this guess is not necessarily very accurate (autosol doesn't know the number of sites any better than you do, for sure).  It has evidently guessed 10 sites based on the size of the structure.  

The heuristics for guessing numbers of sites in autosol are in $PHENIX/phenix/phenix/autosol/utils.py (function: get_number_of_sites):

Se/S:  NCS copies * number of MET
Br/I: NCS copies * (1 per 20 residues up to 10 per chain)
General heavy atom: NCS copies * (1 per 100 residues up to 2 per chain)

The parameter truncate_ha_sites_in_resolve=True is quite different. That one is for truncating density at heavy atom sites in the first stages of density modification


All the best,
Tom T
________________________________________
From: phenixbb-bounces at phenix-online.org [phenixbb-bounces at phenix-online.org] on behalf of Edward A. Berry [BerryE at upstate.edu]
Sent: Saturday, February 02, 2013 11:22 PM
To: PHENIX user mailing list
Subject: Re: [phenixbb] autosol MIR with more than one native?

Thanks!

***forgot to say- I have no idea how many sites there are, 14 is quite possible.

Another question-
in AutoSol_run_12_/TEMP0/ANISO_xtal_PHX.sca_iso_1.sca_hyss.log:

Search parameter: truncate_consensus_model: True
 . .
Score for consensus model (14 sites, before truncation): 0.525 (cc)
Truncating consensus model to expected number of sites.
Correlation coefficient for consensus model (10 sites, after truncation): 0.324

Am I shooting myself in the foot by truncating the model?
"truncate_consensus_model" is not set in autsol.eff.

There is: truncate_ha_sites_in_resolve = Auto *True False

is that related?


Terwilliger, Thomas C wrote:
> Hi Ed,
> Unfortunately, autosol only has a way to deal with one native, so you'd need to average them together or run 2 runs, one with each.
> All the best,
> Tom T
>
> ________________________________________
> From: phenixbb-bounces at phenix-online.org [phenixbb-bounces at phenix-online.org] on behalf of Edward A. Berry [BerryE at upstate.edu]
> Sent: Saturday, February 02, 2013 10:39 AM
> To: PHENIX user mailing list
> Subject: [phenixbb] autosol MIR with more than one native?
>
> Is it possible to run phenix.autosol MIR problem with two "native" datasets?
> And does this make sense, or would it be better to scale the two natives
> together to make a single dataset before running autosol?
> eab
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb
>


_______________________________________________
phenixbb mailing list
phenixbb at phenix-online.org
http://phenix-online.org/mailman/listinfo/phenixbb


More information about the phenixbb mailing list