First, thank you so much for this vibrant community.
I have two quick question about LCModel.
I have some phantom 7T .RAW data with acetate and lactate. I have also created a .basis file with acetate and lactate (from simulations) and I am trying to fit the data in LCModel. I know it’s a very trivial example, but I want to test LCModel for other purposes. However, I am getting a MYBASI 7 error (“two few or none basis metabolites”). Has anybody had experience with that? Does LCModel break if it’s two few basis metabolites?
Do people insert lipids manually to the .basis set? Or does LCModel “detect” it automatically?
Thanks much for creating this community. It has already helped me a ton.
Thanks for your kind words. Let’s see if we can help you.
LCModel, by default, is looking for certain landmark peaks (usually Cr, Cho, NAA, Glu, mI) to do a preliminary analysis. These need to be included in the basis set, and if they aren’t, LCModel throws this error. See the manual (page 162) for an error definition of MYBASI 7. The good news is that you have acetate in the spectrum, which should be a strong enough singlet that you can use it as the reference signal. For that, you’ll need to edit the control file. This is described pretty well in section 11.6 of the manual - let me know if you find that unclear.
Lipids in LCModel are not part of the basis set, but are generated internally, i.e. simulated ‘on the fly’ with certain parameters that are specified in the control files. The defaults are linear combinations of some Gaussians with certain center frequencies and widths. These center frequencies and widths become free model parameters during the optimization, but also have expectation values specified in the job filem too, and there is a penalty term in the cost function for deviating from them. See chapter 11.7 of the manual for details.