Dear experts,
I would appreciate some advice on metabolite selection in Osprey, specifically the use of opts.fit.includeMetabs = {‘default’} versus including all metabolites from a custom basis set.
I initially ran my analysis using a custom basis set with the default option, which excludes several metabolites present in the basis set, including glucose (Glc), as I was not aware at the time that a custom includeMetabs specification could be defined. After switching to a custom includeMetabs specification (motivated by prior work showing that Glc+Tau can be detected with this acquisition), I included these metabolites explicitly and was then able to detect a Glc+Tau signal (according to my predefined criteria). This change also appeared to reduce the reliability of myo-inositol estimates in my data.
My understanding, based on Helge’s explanation in this post, is that the default set is a pragmatic choice to improve model stability by excluding weak, strongly overlapping, or pathology-specific metabolites.
In my data, the two approaches yield slightly different results, although the spectra and overall fit quality appear comparable. I am therefore considering reporting both versions (default vs expanded metabolite set), but I am unsure whether this is methodologically appropriate.
My questions are:
- Is there a general recommendation for when to use opts.fit.includeMetabs = {‘default’} versus including additional metabolites from a custom basis set?
- In a case like this, where the basis set and sequence are designed to support these metabolites, should I report both model specifications (default and expanded), or is it preferable to select and report only one, and based on what criteria?
Any guidance or references would be helpful.