Any cost to adding glycine as metabolite for hercules data?

When i add glycine as an included metabolite instead of using the default metabolites for Hercules, I seem to get better fits around 3.5 and a more flat residual for the sum fits. Is there a cost or downside to adding glycine? I am curious why it was not added as a default, and I should not be adding it?

Hi @sdw,

For healthy brain spectra, glycine is mainly overlapped by myo-inositol and is not detected reliably. However, for certain tumor types, mI decreases while glycine substantially increases 10.1093/neuonc/noaa034. So, if you are looking at tumors, it should definitely be included; for healthy tissue, not so much.

The reason that you see flatter residuals is that you increase the degrees of freedom for the model to account for a sub-optimally simulated mI spin system.