How can i find the same species(and compartment) in the SBML models(in biomodels database)?

[cruise h]

I use the name in the species to find the same species, but i wonder if i should use the URL in the model to find it.

Evenworse, i find the compartments name in the models are sometimes very annoying, someone just use 'compartment' to name a compartment, i use the URL to find the same compartment , but i doubt that the URL maybe out of date or the different URL in the compartment maybe the same compartment.

I would be very grateful if some experts like you can answer my question.

Thank you!

[Robert Phair]

I am not an SBML expert. But I am a user of SBML as a standard for model exchange between software tools. I understand your frustration with models that effectively ignore the SBML concept of compartment. About 10 years ago I did a rough survey of the number of compartments in SBML models in the Biomodels archive. The average number of compartments was less than 2. You are probably familiar with the KEGG database of pathways in which there are few, if any, explicit compartments. You can look at the cholesterol biosynthesis pathway and never know that some of those processes are happening in endoplasmic reticulum, some in cytosol, and some in mitochondria. When you see a model that does not have more than one compartment or the one compartment is called "compartment," then you can reasonably assume that the idea of compartments was not seen as useful to the creator of that model. For example, if you are just doing enzyme kinetics in a spectrophotometer cuvette, and all the substrates, products and enzymes of interest are in solution, it's hard to work up much interest in the name of the compartment where the reactions are taking place. Models like this are essentially biochemistry-focused. They resemble the (very) old biochemical pathway paper wall charts from decades ago. That's the model of KEGG and human-cyc and others. Such models have their place, and they have proved their worth. Biochemistry, of course, is just the first step toward model building in human biology. As soon as you enter the world of cell biology, compartments become essential. Some molecules are membrane bound, and the varieties of organellar membranes and the intra-organellar spaces must become part of your model. Many molecules are found in more than one cellular compartment (I know you know this). As soon as a model includes transport, for example from extracellular space to cytosol, or from cytosol to mitochondrial matrix, or from ER lumen to Golgi lumen, you may have a single molecule that corresponds to multiple species (in the SBML-sense of "species"). If you want to build models of molecular cell biology, compartments are essential. Then you might decide to jump to the level of physiology. Now you have somewhere around 300 different cell types, each of which has its own complement of organelles, plus all the physiological compartments like blood plasma, interstitial space, cerebrospinal fluid, synapses, and more. Sometimes your project requires one level of detail, sometimes another. If you care about specific compartments, you should know that there are many, many SBML users who also care about them and use them every day. We encourage you to use them too. They make your model more specific. Just remember that when you name your compartment "cytosol," you are still leaving a lot unspecified. Is it the cytosol of a red blood cell, or a cardiac myocyte, or an activated T cell?

With your help, maybe we can get the average number of compartments per SBML model up to 2 or even 3!

There is, as you know from frustrating experience, no requirement for annotations that tell you more about the biological context of molecules or compartments. SBML can be used to encode a simple system of ODEs, with nothing other than numerical IDs to differentiate among the state variables, but it can also take advantage of SBO terms, and databases like ChEBI to make your model rich and detailed.

This has become a sort of blog post about compartments. I apologize for its length.

A model without compartments can still teach you something about the modeled pathway that you did not know. Sometimes you will search long and hard for the primary reference because the model creator did not have time or interest in deep documentation. More and more, however, professional curators in groups like Reactome, Wiki-pathways, and the European Disease Map community are doing a fantastic job of using SBML notations to make models richer and deeper resources.

Gook luck with your work!

Robert

--
You received this message because you are subscribed to the Google Groups "sbml-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sbml-discuss...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sbml-discuss/04b35c28-cc2c-4d1d-b46a-ad963fb9e3f4n%40googlegroups.com.

[cruse hu]

Thank you so much.

Robert Phair <rph...@integrativebioinformatics.com> 于2021年3月2日周二 下午1:49写道：

To view this discussion on the web visit https://groups.google.com/d/msgid/sbml-discuss/CA%2BP6KoxvCuR0vYKEEZD0RSNBgQ2x6-fkM94O8pZH2xeXpCwr0g%40mail.gmail.com.