Re: Ponderosa Analyzer Questions

[Woonghee Lee]

Dear Jacob,

Those numbers are all arbitrary. You can some number to another runs with same option and protein but their are no exact criteria because all proteins behave differently and the pseudo energy will be different depending on your inputs. Clearly the number will rise if you provide more information and restraints and it does not mean that the resulting structure is worse. You can only brief assume that lower target function means less violations and as long as you see your structure is well-folded with sufficient restraints, it is good sign. On the other hand, lower number does not mean good when your structure is being unfolded. 

I always advise only fix only a part of big violating restraints first and rerun again and do this iterative because violation is based on your CURRENT structure which is not yet fully determined. That means, you can sort out good constraints if you fix everything all at once including disabling constraints. When fixing constraints, my criteria for disabling vs expanding is, if I see clear corresponding peak that makes sense in 3D structure and other supporting constraints also in contact map, that becomes 1. keep as it is for another run or 2. expand in case approximation was done a little aggressive. Because a peak height can be increased by convolution, too. If a constraint never makes sense like no clear corresponding peak either i-j and j-i and structure does not make sense at all in PyMOL, then it is time to disable. However, I would do it just a few constraints first and rerun because for the reason I mentioned above. Your standard right now to assess violations is only an intermediate structure.

Best,

Woonghee

On Wed, Jan 8, 2020 at 3:49 PM Lewellen, Jacob <jal...@ostatemail.okstate.edu> wrote:

Dear Dr. Lee,

I am currently using the distance constraint validator of Ponderosa Analyzer to remove violations for a 144 residue globular protein. I have connected Pymol and Sparky to the validator and have disabled a small number of constraints with large violation distances and ambiguous assignments. I had a couple questions about the software that I could not find the answer to in your group's papers and Youtube videos:

1. What does the TARGETFUNCTION, TARGETFUNC_SCORE, and OVERALL_SCORE measure and what values are considered acceptable, or how can one determine this?

2. On what justification should one considering disabling a constraint versus setting that constraint to a limit to remove violations?   

I was wondering if you knew of any good resources for answering these questions, or if you had any tips for improving our structure. I appreciate your time.

Sincerely,

Jacob Lewellen

 

--

================================================

Woonghee Lee, M.S., Ph.D.

Assistant Scientist

        and

Principal Investigator

National Magnetic Resonance Facility at Madison (NMRFAM)

and Biochemistry Department

University of Wisconsin-Madison

DeLuca Biochemistry Laboratories

433 Babcock Drive Rm B160E

Madison, WI USA 53706-1544

wh...@nmrfam.wisc.edu, wle...@wisc.edu

Office: +1-608-263-1722

https://sites.google.com/view/wlee-group

www.nmrfam.wisc.edu