Processing the spectra in TopSpin

Only two special issues are important when processing dynamics data with TopSpin:

1) Do all the xfb transformations using the NC_proc argument with the same parameter. For example, you can do always "xfb NC_proc 0". This keeps the program from applying its own rescaling to the data, so you don't lose information about relative intensity among spectra. If you don't do this, you get crazy NOEs and unfitable T1 and T2 data.

2) Be sure that you apply the same phase and baseline parameters to all the T1 experiments. The same applies to T2 and NOE.

NOTE ON PROCESSING THE NOE EXPERIMENT:
The Heteronuclear NOE experiment is acquired in the interleaved mode, this means that the irradiated and non-irradiated spectra are acquired as a single experiment, each scan alternating the irradiation on and off. In order to process this spectrum you need to use an AU program called "splitinvnoe".


Fitting the data with Sparky

Fitting the data with Sparky is very easy. Just open all the (already processed) 2rr files for 1 experiment (I suggest you change the name of each 2rr file to a more suitable one, for example, you can call the experiments "2rr_T1_10ms", "2rr_T1_30ms", etc). You may save this files in Sparky format just in case it crashes (!!!), and you may even save a project for T1, another for T2, yet another for NOE.

Fitting T1 and T2 data

Take a spectrum for which you see all the signals very well (i.e. its delay is very short). If you don't have peaks on it, load a peaklist ("rp" command, you can get an excel file for creating Sparky peaks list here) and move all the peaks to their corresponding best positions using the "pc" and "pi" commands.

When you're done, copy all the peaks from the spectrum where you tuned the positions (command "oc") and paste them to all other sectra ("op").

Now go back to the spectrum where you can see nice peaks, select all the peaks, and type "rh". This opens the "Relaxation Heights" window. Select "Heights at each position in each spectrum", enter "Setup" and in the setup window chose which spectra you want to include in the analysis (typically all) and enter their corresponding time delays in ms. Accept this window, also accept the "Relaxation Heights" window, and type "rh" again: the "Relaxation Heights" window will appear, this time with a table of time constants, their standard deviations, and the heights taken from each spectrum. This table can be saved in text format so as to be imported in a spreadsheet. Besides, if you click on a row you can see the decay profile for the corresponding aminoacid.

Calculating the Heteronuclear NOE

Just load the two (processed) 2rr files (actually you may call them "2rr_irradiated" and "2rr_non-irradiated") in Sparky, load a peak list on the non-irradiated one, center the peaks using "pc" and "pi", and copy these peaks and paste into the irradiated spectrum.

For each spectrum, do "lt" and save the peak list including height information. Import both files in a spreadsheet, and calculate the NOE as the ratio between the intensities of the irradiated and non-irradiated spectra.


Calculating the total correlation time from T1 and T2 data with the program tmest

First you have to select which aminoacids can be used in the analysis, according to deviations of R2/R1 and NOE values. After you've selected all the aminoacids, you create a file suitable for the tmest program and run it (this program is from Palmer's group)

Prepare a spreadsheet with this columns:

A= AA number | B= T1(s) | C= T2(s) | D= (T1-T1av)/T1av | E= (T2-T2av)/T2av | F = D-E

Where T1av and T2av are the average T1 and T2 values repectively (that is, average of columns B and C)

Now create a column G whose value will be:

- "OK" if the absolute value of F is smaller than 1.5 times the standard deviation of column F

- "no!!!" otherwise

In Excel this could be done like this (example for row 3, spanish version):

=SI(ABS(F3)>DESVEST(F)*1.5,"","OK")

All those aminoacids for which column F is "OK" can be used for analysis of its R2/R1 ratio if the NOE criterium is also "OK" (see below).

Now put in column H all the NOE values, and create column I, which will be:

- "OK" if the NOE is bigger or equal than 0.6

- "no" otherwise

In Excel (example for row 3, spanish version):

=SI(H3<0.6,"OK","no!!!")

Now calculate, in column J, the R2/R1 ratio (=T1/T2) for those aminoacids who have "OK" in columns F and H. Calculate the average value and standard deviation of this ratio, and write down these numbers (they will be used in the tmest program)

Run the tmest program and answer its questions:
Enter "15n" to the "input nucleous" question
"600.13" for the "proton spectrometer field"
The average and standard deviation in the R2/R1 ratio, separated by a comma
An initial guess for the tau-m (in nanoseconds)


Calculating the Spectral Density Function at frecuencies 0, wH and wN for each residue

Warning: the provided formulas apply only to data recorded at 600 MHz

Prepare a spreadsheet with this columns (you can use this template):

A= AA number | B= R1(1/s) | C= error in R1 | D= R2(1/s) | E= error in R2 | F = NOE | G= error in NOE

Column I will be J(wH) and J its error:

I = (F-1)*(B/5)*(-0.101)*4/72100^2

J = RAIZ((B*G)^2+(C*(F-1))^2)*(0.101*4*0.2/72100^2)

Column K will be J(wN) and L its error:

K = 4 /(3*72100^2+4*35400^2)*B*(1-(7/5)*(-0.101)*(F-1))

L = 4 /(3*72100^2+4*35400^2)*RAIZ(((1-(7/5)*(F-1)*(-0.101))*C)^2+((-7/5)*(-0.101)*G)^2)

Column M will be J(0) and N its error:

M = 1 /(3*72100^2+4*35400^2)*(6*D-3*B-(18/5)*B*(F-1)*(-0.101))

N = 1 /(3*72100^2+4*35400^2)*RAIZ((6*E)^2+(-3*C+(18/5*0.101)*(C*(F-1)))^2+((18/5*0.101)*(B*F))^2)

1