Subject: Re: [nmr_sparky] 2d noesy peaks integartion error linefit faileld to converge
From: Eiso AB
Date: Feb 7, 2012

Previous: 811

Hi Srinivas

cnu4us wrote:

Hello sparky users ,

I am using 2D noesy spectrum (H,H) to do structure calculation , while doing peak integration it is showing errors as follows .

Linefit fails to converge

peaks are not integrating and some other regions it can integrating very few peaks ( around 10% ).

I am using these following parameters to integrating .

integration method : Gaussian fit

this is normally ok, unless you use an exponential window.

allow peak motion : YES
Adjust line width : yes
fit baseline :no

substraCt off fit peaks :no

this should be YES in almost all cases I can think of.

if it fails to converge if there are usually too many
peaks inside the lowest contour. increase the (lowest)
contour level so there are less peaks inside and
integrate smaller groups of peaks again.

all peak parameters within a lowest contour will
be fit together so when you integrate too many
peaks within a contour the optimisation space
will be too large.

use data above lowest contour :YES

use data with in rectangle :yes
group peaks in contour boundary :yes
group close peaks : yes

not sure if the 3 above here are right, try other settings
if you still have problem (and check the manual)

max motion : 0.01 0.01

line width min : 2.0 2.0
line width max :80.0 80.0

you could make these closer to the actual linewidth

dont expect to be able to use the same parameters
for the whole spectrum, integrate smaller groups if it
doesnt work

And a last remark, in my experience just using peak
heights is much more robust than using fitted integrals
as NOE intensities. The errors in the intensities you can
get from a mis-fitted peak are much larger than the
error from e.g. overlap in the height of the peak.
And if your linewidths are fairly homogeneous the height
should be proportional to the integrated intensity.

goodluck, Eiso

group range : 30.0 30.0
maximum minimization steps : 10000
minimum tolerance : 0.1

any one make some helpful comments !

Regards
Srinivas p
penumutchu.srinivas@... mailto:penumutchu.srinivas%40gmail.com