Feedback on Mapping for "SBML core"<->"SBGN PD", "SBML layout"<->"SBGN PD", and "SBML qual"<->"SBGN AF"

[Haoran Yu]

Hi everyone,

I'm currently working on creating mapping diagrams between SBML and SBGN, I have 4 mapping diagrams in mind:
(1) "SBML core"<->"SBGN PD"
(2) "SBML layout"<->"SBGN PD"
(3) "SBML qual"<->"SBGN AF"
(4) "SBML core"<->"SBGN AF"

So far I created the first 3, here are the diagrams:
(1) "SBML core"<->"SBGN PD":

(2) "SBML layout"<->"SBGN PD":

(3) "SBML qual"<->"SBGN AF":

Mapping (3) still needs some work, the values of Input and Output still needs to be mapped somehow, although I am not sure how to proceed.

Could anyone give me some feedback to the 3 mappings please? Are there anything unclear, are there mistakes?

I am working on mapping (4), but I am stuck because it is unclear how I should represent SBGN AF glyphs using SBML core alone.

I have 2 options for "SBGN AF" -> "SBML core":

[1] use SBML AssignmentRule to map SBGN "Modulation arcs"+"Logical operators"

[2] use SBML KineticLaw to map SBGN "Modulation arcs"+"Logical operators"

In the case of [1], SBGN "Activity nodes" map to Species, and Species initialAmount take on value of a Parameter, which is associated with an InitialAssignment.

The resulting value of an Activity Flow would be calculated using an AssignmentRule, where the Math calculates values provided in the Species, and produce a boolean. This boolean value would determine whether to change the value of the resulting "Biological activity" or "Phenotype".

In the case of [2], the approach would be similar, but I feel that by using a KineticLaw, I am forced to create a Reaction between the Input and Output Species, which makes the mapping look like "SBGN PD" -> "SBML core". I think "SBGN AF" should not map to any Reactions, because we cannot assume a Reaction exists for an Activity Flow model. AF is supposed to represent activities without knowing the details of the intermediate steps.

Also, I do not know whether mapping of "SBML core" -> "SBGN AF" would be possible, because I assume the "SBML core" model would be quantitative. Converting from a quantitative model to a qualitative SBGN AF could be incompatible. Under what circumstances would a "SBML core" -> "SBGN AF" be feasible?

Thank you for your help!

Haoran

[Haoran Yu]

I'm also going to attach the mapping diagram for converter (4) "SBML core"<->"SBGN AF":

https://github.com/NRNB-GSoC2017-SBML2SBGNML-Converters/SBML2SBGNML/blob/master/documentation/SBML%20core%20_-_%20SBGN%20AF.png

Thanks,

Haoran