Comments from Adrien Rougny
I have three main comments:
- In all the following we use macromolecules in the rules. In the converter, does the nature of the entities matter? Or is it sufficient to know they are active/inactive, regardless of their nature? If their nature does not matter, could we replace the macromolecules with unspecified entities, and explain that in the following an unspecidied entity stands for any EPN?
- Also I would suggest not using state variables to indicate whether the EPNs are active/inactive, because this suggests the EPNs actually carry those state variables in the input PD map. We could maybe use a color code to indicate which EPNs are active (and leave inactive EPNs uncolored, as it is the default). Finally if we keep state variables, we should maybe change them to stadiums (PD L1V2.0).
- In each rule we use some kind of "bricks", that is patterns that are valid full SBGN (sub)-maps (e.g. modulated processes with reactants and products). We could rather use "half" bricks (e.g. modulated processes with only products for example). That would allow avoiding the combinatorics due to considering full bricks, significanly reducing the number of rules, while keeping the exact same result at the end.
In the next section I added the rules for finding active EPNs. I used unspecified entities and a color code for active EPNs (green for active), as an example of my previous comments.
Rules for finding active EPNs
Following are the rules that define which EPNs are active.
0.1 act
0.2 mod
0.3 stim
0.4 cat
0.5 inh
0.6 nstim
Here is the specification for the transformation rules. We start from most elementary patterns and follow with increasingly complex ones. All these patterns are also test cases for regression testing of the Converter program. With each case we ensure that converter code translates left-hand diagram exactly to the right-hand diagram, while both being considered as standalone schemes.
Posttranslational modification
Associated terms: SBO:0000182 conversion, GO:0036211 protein modification process, GO:0140096 catalytic activity, acting on a protein
What is described here is not really posttranslational modifications per se, but rather some kind of activation/deactivation processes (not sure for the rules of the form i-X-i though). Could we rename the section by “Activation and deactivation”, for example?
1.1 i-sti-i
1.2 i-cat-i
1.3 i-nsti-i
1.4 i-mod-i
1.5 i-inh-i
1.6 i-sti-a
1.7 i-cat-a
1.8 i-nsti-a
1.9 i-mod-a
1.10 i-inh-a
1.11 a-sti-i
1.12 a-cat-i
11.13 a-nsti-i
1.14 a-mod-i
1.15 a-inh-i
1.16 a-sti-a
1.17 a-cat-a
1.18 a-nsti-a
1.19 a-mod-a
1.20 a-inh-a
Associated terms: GO:0008152 metabolic process
Why are 1.27, 1.28 and 1.29 in this section? Shouldn’t they be in the previous one? Also there is a problem with the drawings of the reversible processes. We should redraw them.
1.21 m-sti-m
1.22 m-cat-m
1.23 m-nsti-m
1.24 m-mod-m
1.25 m-inh-m
1.27 a-catr-a
1.28 a-catr-i
1.29 i-inhr-a
1.30 m-catr-m
1.31 m-inhr-m
Source and sink
Could we rename this section to “Production and degradation”?
Associated terms: SBO:0000179 degradation
2.1 ss-sti-i
2.2 i-sti-ss
2.3 a-cat-ss
2.4 a-inh-ss
Complex association and dissocitation
New section to regroup “Oligomerisation” and “Complexes”.
Associated terms: SBO:0000177 non-covalent binding, SBO:0000180 dissociation, GO:0005488 binding
Heterocomplex association and dissociation
We may use the term “heterocomplex” to differentiate from oligomers
5.1 Heterocomplex association (named complex)
5.2 Heterocomplex association (unnamed complex)
5.3 Regulated heterocomplex association
Generally, shouldn’t associations be regulated via stimulations rather than catalyses?
5.4 Heterocomplex dissociation
Oligomerisation
In the same section we have multimerisation so couldn’t we simply rename the section by “Multimerisation”?
All the following rules could be grouped into a unique entry for the sake of succintness: we could rename the section to “Homo-multimerisation”, write the label of the unit of information on the multimer as “N:X”, the label of the target biological activity as “A-Y”, and add a table like the following one, for example:
X |
Y |
2 |
dimer |
3 |
trimer |
4 |
tetramer |
5 |
pentamer |
6 |
hexamer |
7 |
heptamer |
8 |
octamer |
9 |
enneamer |
10 |
decamer |
>11 |
multimer |
3.1 Dimerisation x2
3.2 Oligomerisation x3
3.3 Oligomerisation x4
3.4 Oligomerisation x5
3.5 Oligomerisation x6
3.6 Oligomerisation x7
3.7 Oligomerisation x8
3.8 Oligomerisation x9
3.9 Oligomerisation x10
3.10 Multimerization 1
3.11 Multimerization 2
3.12 Oligomerisation x6 r
Translocation and transport
Renamed to “Translocation and transport”, because “regulated translocation” is a transport following the SBO.
Associated terms: SBO:0000185 tranlocation reaction, SBO:0000655 transport reaction, GO:0006810 transport
4.1 Translocation 1
4.2 Transport
Renamed to transport.
Hidden inhibition
Could not find any associated terms for this category.
6.1 Hidden inhibition 1
6.2 Hidden inhibition 2
I think this one can be deleted since it is a combination of 6.1 and 1.7.
Simple schemes
I don’t think this is a rule, not sure it should be in this list.
8.1 raf-mek-erk
New schemes
9.1 Currency metabolites 1
I think this is rule should be moved to the “Posttranslational modifications”.
9.2 Noncurrency metabolites 1
I think this is rule should be moved to the “Posttranslational modifications”.
9.3 Regulation combining 1
Is this a unique rule, implemented in the converter? It could be a rule among others that allow transforming logical operators (for example this rule could also work with an “AND” operator).