N15 Relaxation Experiments


Lewis Kay’s Relaxation Experiments - T1 T2rho and NOEFile Names

Choose names carefully so the modelfree scripts can work
Suggesed names are
  • T1_ for T1 experiments
  • T1rho_ for T1rho experiments
  • NOE_0 and NOE_1 for noe off/on experiments
  • NOE_2 and NOE_3 for duplicate noe off/on

N15 T1 Relaxation (setup)

Load A__templates/N15T1_lek_pfg_sel_enh to get parameters and sequence.
Set tof, pw, pwN to calibrated values as usual.
May need to fine tune gzlvl8 (in increments of 50) to maximize signal.
If C13 labeled then set c_flg='y'.
Do a quick 1D array to estimate T1 by setting ncyc=0,40,80,120,160,200,240,280 and starting with "go" (NOT au). Set th on peak of 1st spectrum, then set th=th/3 and see which spectrum has that peak height. Set ncyc back to zero.
run VNMR macro rmT2pts(T1 estimate, # points) to calculate nicely spaced vales for T1
enter these ncyc value, rounded to be even numbers, into array Rtimes
Rtimes=0,12,24,34,...
To display current values of Rtimes, run maco "pr_Rtimes"
The relaxation delay is given by ncyc*5 msec, so ncyc=20 gives a delay of 100 msec. ncyc must be even.
Before running series, run a 1D scan and carefully adjust phase.
To set up a series of 2D experiments, make sure the following are set correctly:
Rtimes array (see above)
wexp = 'format(r2,0,0):n2 n3=n1+n2 svf(n3) r2=r2+1 if r2<=15 then ncyc=Rtimes[r2] au('wait','next') endif'
(NOTE: to enter ' in middle of string escape it with \)
n1='/export/home/Kay/vnmrsys/data/mydir/myT1data_')
r2=1 and ncyc= first ncyc value in array Rtimes(usually 0).
Make sure that the directory designated in n1 exists and does not contain files with the same name.
Start experiment series with au('wait','next')

Notes taken during Ranjith's visit:

Current version does not use pw_sl. Set tpwrsl = -16
pw_shpss should be a 180° pulse width
spss_pwr is the power for the 180° pulse (e.g. 45)
To obtain a quick look at how fast the T1 decay is going to be, set nt=1, phase=1, and array ncyc

N15 T1rho Relaxation (setup)

Load A__templates/N15T1rho_dk_enh to get parameters and sequenc.
Set tof, pw, pwN to calibrated values as usual.
If C13 labeled then set c_flg='y'.
set HD_flg = 'u'
HD_flg='u' and ncyc=2.
Delay is given by time_relax value instead on ncyc.
time_relax_max (= 0.1) is maximum permissable value for time_relax delay.
Do a quick 1D array to estimate T1rho by setting time_relax=0,.,02,.04,.06,.08,.1 and starting with "go" (NOT au). Set th on peak of 1st spectrum, then set th=th/3 and see which spectrum has that peak height. Set time_relax back to starting value (usually 0).
enter relaxation delay time values you want into array Rtimes
Rtimes=0,.01,.02,.03,.04,.05,.065,.08,.1
To display current values of Rtimes, run maco "pr_Rtimes"
Or set HD_flg = 's'
Delay is given by ncyc*10 msec
enter desired ncyc values into array Rtimes
Rtimes = 0,1,2,3,4,5,6,8,10,12,1,4,8
Before running series, run a 1D scan and carefully adjust phase.
To run a series of T1rho experiments do the follogin
Set Rtimes to desired delay times (see above)
set r2 = 1
do one of the following depending on HD_flg
set relax_time to first time point in Rtimes array (HD_flg='u')
wexp='format(r2,0,0):n2 n3=n1+n2 svf(n3) r2=r2+1 if r2<=size('Rtimes') then time_relax=Rtimes[r2] au('wait','next') endif'
(NOTE: to enter ' in middle of string escape it with \)
set ncyc to 1st value in Rtimes array (HD_flg='s')
wexp='format(r2,0,0):n2 n3=n1+n2 svf(n3) r2=r2+1 if r2<=size('Rtimes') then ncyc=Rtimes[r2] au('wait','next') endif'
(NOTE: to enter ' in middle of string escape it with \)
n1='/export/home/Kay/vnmrsys/data/mydir/myT1rhodata_')
Make sure that the directory designated in n1 exists and does not contain files with the same name.
Start series with au('wait')

Convert T1rho to T2

You must first prepare tables containing the R1 and R1rho data. The format of these tables is that output by the curvefit2table module of the modelfree programs. See modelfree documentation.
You must export a crosspeaks file from felix that contains the chemical shift information. Only the first assignment name is used and should match that used in the R1 and T1rho tables.
Execute the following program to construct a R2 output table
t1rho2t2 R1file R1rhofile xpeakfile outputR2file carrierFrq spinlock RFfreq
carrier frequency is the center of the spectrum in ppm (e.g. 4.7)
spinlock is the spin lock field strength in Hz used in the T1rho experiment
SL = 1 / (4*pwn_sl)
For example, SL = 1 / (4*.000157) = 1592.4
RFfreq is the N15 spectrometer frequency in MHz (e.g., 50.6621 on the 500)
The output file will contain the R2 values in the same format as the input tables.

Notes taken during Ranjith's visit:

Use most recent option with Hd_flg = 'u'
Set ncyc = 1 (must be greater than 0 in order to run; but the parameter is not used otherwise).
Set time_relax to desired value, set time_relax_max to 0.1 (100 msec)
Trim and nrpul, tofml are not used in the 'u' option.
Set dpwr2sl = 44, Set pwn_sl = 160
pwhd should be around 20 usec – set tpwrhd to achieve this (Ranjith actually used tpwrhd = 48 with pwhd = 13.3 usec since so few pulses are actually used now in the u option)
Estimate T1's by arraying time_relax with ni =1, phase = 1 to get a series of 1D spectra. The T1 is approximately the time to achieve 1/3 the initial intensity. Set time_relax = 0.005 for initial value rather than 0).
Setting time_relax time points –
Set up 8-9 time_relax values. The maximum time_relax value should not exceed the T1 (or T2 value) since cross-correlation effects become more important at later times and affect the result.
The time_relax values should be set to achieve 8-9 equally spaced values on the y-axis. Run rmT2pts to calculate the best time_relax values. Enter estimate of T1 or T2 in msec and number of desired points (e.g. rmT2pts(70,9). Keep minimum delay to around 0.005.
Analysis of data requires the chemical shift of each N15 nucleus and the field strength.

N15 NOE experiments (setup)

Load A__templates/N15NOE_off_lek_pfg.fid to load the sequence and starting parameters.
Set the usual tof, pw @tpwr, pwn @dhpwr2 to calibrated values.
Check calibration for pw_noe @tpwrnoe=45 on the 500, and at power of 50 on 800.
May need to fine tune gzlvl3 for maximal signal and minimal water.
To do this, use the T1 exp and array gzlvl8 (T1's gzlvl8 is equivalent to NOE's gzlvl3) in a 1D exp with gzlvl8 in increments of 50.
Set d1 to total relaxation time (5 sec or more; 10 is best).
May want to use nt=32 for good S/N.
If C13 label then set c_flg='y'
Before running 2D, run a 1D scan and carefully adjust phase.
To set up a run with NOE turned on, copy parameters to another exp (mp 3,4) and then change the following parameters:
ncyc=1000 for a 5 sec NOE buildup
d1=5 (total delay = 5+5 = 10 sec; same as NOE off)

From Notes taken during Ranjith's visit:

This experiment uses very strong gradients (e.g. 30000 for gzlvl8). Some people have cut them in half.
Set pw_noe to 18 and tpwrnoe = 45. With a tpwrnoe of 44, pw_noe was 20.25 usec. Note that the time for a pw 90 is set and the code changes it to be 120°.
Set d1 = 10,5
Set ncyc = 0, 1000 (total time delay will be 10 sec with this combination of d1 and ncyc.)
On the 800, if there is a water problem, use a 90° pulse in the 1H channel after the d1 delay followed by a gradient to destroy the water.
The 1H's are saturated with a series of 120° pulses.

VNMR Processing

You can now use the Tcl process window in VNMR instead of grad_sort for many of the 2D experiments. For this to work you must have the proper dg templates installed in your directory. User Kay is already set up for this.
Transform the 1st increment and phase using RP.
Adjust the window functions, etc and press the F1xF2 button on the Process Tcl/Tk window pane. You must use this button or the ft coefficients will not be correct and the spectrum will not look right, unless you run grad_sort_nd first.
If phasing in D1 gets out of wack, FT 1st increment and set RP=RP-90.
Alternatively, after adjusting the phase you can enter in the command line "processT1" (see below).
Note: The Tcl/Tk interface for processing L. Kay's sequences require the appropriate files in ~/vnmrsys/templates/dg. Currently supported experiments are T1, T1rho, and N15noe.

Automatic Processing of Relaxation Series

First make sure the T1 and T1rho series are started with phasing properly adjusted. Properly phase the NOE experiments before saving them. Save all data sets to one subdirectory. Then you can easily process all the data in that subdirectory with one command.
First delete ~/felix directory to make room for the proessed phasefiles.
Examine the file ~vnmrsys/maclib/processT1 and make sure it does what you want.
Then navigate to your data subdirectory using the VNMR menus buttons.
Now enter "ftdir" in the VNMR command line.
All the data in that directory will be processed and the phasefiles written to the ~/felix directory ready to read in with FELIX
Notes: the macro ~/vnmrsys/maclib/processT1 does the processing. You can edit it to enter the RP value, the window function command you want (gaussian, gausshift, pi3ssbsq, or sqcosine), and any other processing parameter. Note that the "dc2d" command will remove stripes that would otherwise appear due to long relaxation delays. If you set up the first experiment carefully, you should not have to add anything to this macro. You can also run the processT1 directly from the command line to process an individual spectrum. You should be able to use the same macro on T1, T1rho, and NOE data sets. There is a backup copy (~/vnmrsys/maclib/processT1_back) in case someone has changed the macro.

Grad_sort (no longer needed)

This is not needed for processing with nmrPipe, or for VNMR if the dg templates are set up properly. See Below.
Multidimensional data are collected using sensitivity enhancement and fids are packed differently than Varian VNMR files. You must unpack or sort them using the grad_sort_nd program before processing with FELIX. Type grad_sort_nd
for syntax.
For a 2D data set with ni=128 and np=1024 type in the data directory:
grad_sort_nd fid fid.sorted 128 0 0 1024
fid is the original data, and fid.sorted is the new data which can be transformed with FELIX

ModelFree