Discussion Notes:FinaliseL1Spec

OK

OK

It was agreed that The Source and Sink should be treated as syntactically different, but have the same Glyph. Also it was agreed that these nodes should only be attached to one arc (degree 1).

OK

OK

OK

OK

OK for Simple Chemical and Macromolecule AS proposed a Multimer Complex - this was not resolved.

AD proposed that we split this to a Perturbation and a Phenotype node.

Perturbation should have an anvil shaped symbol and only be associated with a modulating arc.

Phenotype would be a hexagon and only be the target of a modulation arc.

ED suggested we should try and deal with Generics in the specification. There was some debate as to whether this is a good idea. Agreed to look into this further and it may be something for the next version.

There was concern that making the subunits visible enables users to attach significance to the topology of the complex diagram. However SBGN does not support this. Options discussed were:

1) Not show subunits at all.

2) Show and provide some mechanism for describing topology.

3) Leave as is.

We decided to do 3.

There was debate about whether we should allow users to link consumption arcs to the subunits of a complex. The options were to:

1) Allow this but attach no meaning to it.

2) Not allow it (current spec).

3) Not allow it but investigate requirements from Huaiyu and report back on Wednesday.

We decided to adopt option 3.

Issue raised by NL about whether to add states to compartments.

No conclusion about this.

Raised issue about whether Compartments were nested. Concluded that they are not.

Compartment edges don't mean anything. Idea is that diagram is independent of compartment topology.

Much discussion about using tags for interface specification. To be discussed further.

OK

OK

OK

OK

OK

OK

OK

OK

OK

New arc need because gap in syntax. Provides input to logic gate.

OK

OK

Questioned whether needed. Suggested that could add as Unit of information to a process.

FB suggested changing to a square so it's a process.

SLM/HM suggested using annotation of process.

ED worried that this is a bad idea as a may need to be extended to handle other temporal information.

Defer it until next version.

We need to address this. Examples in current spec don't work. We should identify a good solution to this during the week.

Much discussion and what looked like different understanding of how tags work. This needs to be pursued.