The objective of metabolomics is to characterize in the most complete and comprehensive way the pool of small molecules which are the end product of the metabolism.
The pool of these molecules is known as metabolome.
Metabolomics aims at measuring the metabolites in a quantitative way and to characterize their relations and associations.
The molecules included in the metabolome can be extremely diverse in particular if plants/microorganisms are concerned
The chemical diversity results in different properties which would require different methods of analysis
Some of the metabolite are present in high concentration, while other highly relevant compounds are present only in small traces. Our analytical method should be able to see them at the same time
The chemical diversity makes extremely difficult to have only one analytical method able to analyze everything
The approach is so sensitive that every un reproducible pre-analytical issue will affect the data
We need an analytic technique:
In all the techniques presented so far, each metabolite give rise to a large set of different signals
All the variables coming from the same molecule are highly correlated
Even with the perfect instrument the “numbers” will change. Try measuring your temperature 10 times in a row ;-)
When chromatographic separation is present, elution times will slightly change from run to run. For GC the situation is better, but far from being perfect
In metabolomics we are loading on the instruments a complex mixture of chemicals. This complex chemical environment will affect the response of the molecules in an often unpredictable way. The importance of this effect is so strong in MS that this technique is almost always coupled with chromatography
With chromatography the different molecules reach the ionization source at different times, reducing the matrix effect.
The approach is so powerful that all aspect related to sample handling/storage/extraction will affect the final result
We use the analytical method of choice to analyze a selected set of metabolites. The list can go from around 10 to something around 500.
The measured variables are metabolites
The numbers we collect is the concentration of the individual metabolites
We “throw” a raw extract in our instrument, trying to analyze as much signals as possible. How many things we are able to see is unknown.
The measured variables are not metabolites
The numbers we get are the peak areas of the individual features
Aka … I have a refrigerator of samples … shall we give a look?
Remember that instruments are sensitive, so errors will unavoidably show-up in the final data (garbage in, garbage out)
As you have already understood, in untargeted metabolomics:
Most likely, chemical annotation is the current bottleneck in untargeted metabolomics
e.g. how to sample in a representative way in GC? How to quench the metabolism of microorganisms? how to store biofluids?
Two stage design:
Are my results fitting inside the established (domain specific) body of knowledge?
e.g. Are my markers reasonable under the light of the human metabolism?