Dear all
I am currently interpreting relaxation data and hetnoe using ccpnmr.
My experiments were recorded in an interleaved manner so I would expect that the spectral noise should be the same in all matrices within the same experiment.
But, when I have a look into the spectrum window, noise levels seem to be correlated with peak intensity.
For instance, in my hetnoe experiment, due to the dynamic nature of roughly half of my protein, ca 50% of the peaks in the saturated spectrum are negative. This seems to result in a noise level that is half as big as the noise level in the reference experiment, or, both experiments have been acquired at the same time.
Same is true for the R1 series in which the noise decreases with the increase of the relaxation delay.
What I found on this subject:
ccpnmr.analysis.core.ExperimentBasic.getNoiseEstimate(dataSource, nsamples=1000, nsubsets=10, fraction=0.10000000000000001)
Estimate the noise level for a spectrum by choosing a random nsamples points and finding subsets with the lowest standars devation/
Wouldn't it be better to calculate noise as for instance in nmrview by defining a region without peaks before performing this operation ?
Cheers
Beate
|