Hello Yikan--
>
> I am doing the ensemble simulation against SAXS data, and found that
> when I set different values of xray_scale, the results are very
> differnt.
[good (<1) with large force constant, bad (>>1) with small force
constant.]
>
>
> 1) And I got three different chi^2. But I do not know which one is better?
> or none of them values are not so good?
Smaller is better, but see below.
>
> 2) what is the meaning of “dev” in Energy(dev)?
deviation. It is a measure of energy soread between the different
structures. You need to understand the nature of the violations in the
angle, bond and improper terms. Are these from regions which are held
fixed? If so, they may be ok.
> Did the value of the “xray energy” term
> above(3819.14,3552.58,656.99 ) had been scaled by the “xray_scale”
> in the
> script? and how?
Yes indeed, xray_scale is included in your xray.setScale lines.
>
> 3) did "nonbonded and torsionDB in this case” means VDW and RAMA? to
> get the reasonable or good results, how can i set the value of
> “xray_sacle”(the xray energy term) to match “nonbonded and
> torsionDB” terms?
Hmm. You *should* be using torsionDB and repel, by replacing the RAMA
and VDW set lines with
#
# Setup parameters for atom-atom repulsive term (van der Waals-like term).
#
from repelPotTools import create_RepelPot,initRepel
repel = create_RepelPot('repel')
potList.append(repel)
rampedParams.append( StaticRamp("initRepel(repel,use14=False)") )
rampedParams.append( MultRamp(.004, 4, "repel.setScale( VALUE)") )
# nonbonded interaction only between C1' atoms
highTempParams.append( StaticRamp("""initRepel(repel,
use14=True,
scale=0.004,
repel=1.2,
moveTol=45,
interactingAtoms="name C1\'"
)""") )
#
# Set up statistical torsion angle potential (torsionDB).
#
from torsionDBPotTools import create_TorsionDBPot
torsiondb = create_TorsionDBPot(name='torsiondb',system='rna')
potList.append(torsiondb)
rampedParams.append(MultRamp(0.5, 4, "torsiondb.setScale(VALUE)"))
# Selected 1-4 interactions.
# ---
# 1-4 interactions are ommited from the above term (repel) because they are
# mostly affected by torsionDB (torsiondb). However, torsionDB doesn't affect
# torsions of terminal, protonated groups (e.g., methyl rotation). Thus, 1-4
# interactions involved in such torsions are set up here. (Note: a more elegant
# solution might be implemented in the near future.)
#
from torsionDBPotTools import create_Terminal14Pot
repel14 = create_Terminal14Pot('repel14')
potList.append(repel14)
highTempParams.append(StaticRamp("repel14.setScale(0)"))
rampedParams.append(MultRamp(0.004, 4, "repel14.setScale(VALUE)"))
>
> 4) Last but most important: when i checked the output pdbs, I found
> the output does not change much compared to my initial
> structure. But the chi^2 of the xplor-nih is very small in the
> output file.(but when i caculate the chi^2 by Crysol the chi^2 is
> very big). I do not know why.
This is quite odd. You might check the SAXS results by running the
helper program calcSAXS:
calcSAXS -py \
-omp 8 -fit \
-expt YP1_ddb12.dat \
-numQ 100 file.pdb > file.saxs
[The -omp option simply speeds things up, and here assumes that the
computer as at least 8 cores.]
>
> 5) I have attached my scripts bellow. My SAXS data has been
> extrapolate and qmax=
> 0.3.
>
You should not extrapolate your SAXS curve - and please check the
curve at small q values for signs of aggregation. Both times you call
create_SolnXRayPot, include the option numPoints=50, and then your
curve will be evenly sampled over the range for which you have data.
Let's fix things up, and figure out what's going on.
best regards--
Charles
_______________________________________________
Xplor-nih mailing list
[email protected]
https://dcb.cit.nih.gov/mailman/listinfo/xplor-nih
