Re: [nmrpipe] Error using "nlinLS" function to fit lineshape and extract intensities [1 Attachment]
This problem has been discussed in earlier posts.
Just get a new version of the software.
Uncharacteristically for NMRPipe, "nlinLS" has a hard-wired upper
limit in the number of parameters that can be modeled in the
evolution dimension. I didn't remove the limit, but I increased
it to 524.
Also, in case you haven't used it: these days, "nlinLS" is usually
operated indirectly via "autoFit.tcl".
Also also, current versions of NMRPipe have an interactive viewer
for displaying evolution curves (showEvolve.tcl) which includes
an option to send them to a fitting application (fitXY.tcl).
Examples are at the NMRPipe download site, files "relax.tar.Z" and
Quoting Sujoy Mukherjee <sujoyiit@...>:
> I'm having some problem with lineshape fitting usling the 'nlinLS'
> function. It gives some strange "memory allocation failure" error
> that I've not seen before. I've used this to fit lineshapes and
> extract intensities in pseudo3D datasets, consisting of multiple
> 1H-15N HSQC type experiments quite regularly for several datasets
> and worked fine. Its just that one particular dataset is having this
> problem. Each of my spectra has ~90 peaks with ~20 peaks that are
> overlapped to some extent - and has been denoted properly using
> CLUSTID. There are 25 points (Z_A0 to Z_A25) that vary with a
> relaxation time. All of my spectra have been processed by nmrPipe
> with ~10Hz of GM linebroadening to give a gaussian lineshape to the
> I'm attaching a screenshot of the window that shows up while
> executing the nlinLS script.
> If someone has seen this kind of error or can provide some
> suggestion as to what could cause this, I would highly appreciate it.
> Many thanks in advance,
> The INTERNET now has a personality. YOURS! See your Yahoo!
> Homepage. http://in.yahoo.com/