mdifilippo
commented
2 years ago Hi! The short answer to your question is that it depends on the data you have, and I will clarify here below what I mean:
- in the first scenario, if you only have metabolomics data and nothing else (I mean you have neither extracellular constraints nor constraints on the transcriptional activity of genes involved in your network), nothing prevents you from using INTEGRATE. However, practically speaking, applying the sampling to a completely unconstrained network means getting all and nothing.
- Therefore, to be closer to the biological reality you want to simulate, it is necessary to at least integrate extracellular constraints in your model. In this scenario, you will skip all the initial part of the pipeline where RAS scores are computed (because we assume that you haven't transcriptional data for your samples), and after that you will firstly perform the sampling on the models constrained with just extracellular data, and then you will compute the concordance between FFD and RAS (skipping the computing of concordance between RPS and RAS again for the missing of transcriptional data). With just the concordance between FFD and RAS, basically you can obtain a partial information because you can only say for each reaction if flux variations between bacterium 1 and bacterium 2 go hand in hand with RPS variations. So we can just say if there is a metabolic control of the reaction without however knowing if a combined metabolic and transcriptional control may occur.
I hope I was clear. Marzia
hites77
Can this be used to integrate only metabolomics data in bacteria metabolic model ?