Vitamin B2 (riboflavin) metabolism (R-HSA-196843)

[ukemi]

• [x] Make mouse models: gomodel:6446bfcb00001336 and gomodel:645d887900000006
• [ ] What about in other cellular compartments: PMID:23946482, I think mitochondria too? And see PubMed:16165386 and Reactome unreleased reaction "SLC25A32 transports FAD from cytosol to mitochondrial matrix"
• [x] We are missing an upstream reaction link between the transporter and the kinase: R-HSA-3165230 and R-HSA-196964. Added
• [x] The annotation for the transporter is correct, but it doesn't look right in the pathway diagram wrt location. Fixed location in diagram
• [ ] I think we can deepen the mapping of R-HSA-196955 to FAD diphosphatase activity (GO:0047884). But now that I look at the ref, I'm not sure it supports FAD as a substrate. Interesting annotation question. According to Hrubsa et al. 2022 PMID:35276844, the enzyme responsible for the reverse process is not known but there is evidence to the contrary, although not consistent, summarized in the Google doc. Zimmermann et al. 2012 PMID:22555564 repeatedly mentionsFAD as an Enpp1 substrate in passing but never reviews specific biochemical evidence (and see ENPP1 notes in the Google doc).
• [ ] It also looks like the catalytic domain of Enpp1 is extracellular (again see ENPP1 notes in Google doc). Do we need to dissect apart uptake and transport processes from metabolism, as for cobalamin?
• [ ] Should R-HSA-196950 be mapped to FMN phosphatase activity, GO:0103027. The paper cited (PMID: 2775236) shows activity on a lot of different substrates, but not FMN. It is always difficult for me to decide when to extrapolate to similar molecules.
• [x] Fix charge states and add Rhea reaction names in Reactome (as in ticket #241). Charge states OK, missing protons added to reactions, Rhea reaction names added as synonyms

Edits / additions to Reactome pathway are logged in this Google doc

[ukemi]

• I think we should look at other cellular compartments. I also kept running across the mitochonrial metabolism in my literature searches.
• The transport question brings up an interesting point. In some models, I include it in the core metabolism model but in the case of the cobalamin we split it out. My general rule of thumb is that if it is a single transporter activity that gets a metabolite from one place to another, then I put it in the metabolism pathway. If it is more complex, like in the case of the cobalamin, I split it out. We can certainly discuss this. Would it make sense to generally consider transport to be a part of some type of metabolic process as well as a class in itself?
• There is a way we can capture the lack of direct evidence for a substrate. If we map the enzyme activity to a parent term that encompasses a class of molecules (nucleotide) and put evidence on that and then add a substrate as an input with no evidence or some different evidence,we could tease out what we know from what we hypothesize/predict. The one drawback of this is that in order to use this kind of information really well, we would have to strictly adhere to what the evidence on an edge really means, that it is specifically evidence for that edge. The general community has already strayed away from that quite a bit. I've always been bothered by evidence representation in GO-CAM.

[deustp01]

• We should look at other cellular compartments. I also kept running across the mitochondrial metabolism in my literature searches.

Three finished but not externally reviewed reactions start us in this direction. If they look good to you, we can release them (and think about linking them to organelle-specific FAD-requiring metabolism). Mitochondria SLC25A32 transports FAD from cytosol to mitochondrial matrix Peroxisomes (the two directions of a reversible event are represented as separate reactions in Reactome, so they can be linked into different pathways, e.g., the reversible steps of glycolysis / gluconeogenesis) SLC25A17 exchanges PAP, FMN, AMP for FAD, CoA-SH, NAD+ SLC25A17 exchanges FAD, CoA-SH, NAD+ for PAP, FMN, AMP

[deustp01]

• The transport question

This looks like a pathway-boundaries issue that may be straightforward enough so that we could actually come up with some useful general guidelines - good for a weeds discussion with @vanaukenk .

[deustp01]

• what the evidence on an edge really means

Reactome has always taken a Marxist approach to evidence as in Chico and Groucho reviewing a contract: "No like that bit" / "Then take it out". Here again weeds discussion with @vanaukenk would probably be really useful for trying to come up with / restore useful community guidelines.

[ukemi]

Shall we focus on this pathway on a weeds call? I think it might be worth it. I will modify my mouse models to align with what we come up with for Reactome. The models I have are seeded to align with Reactome at the moment, but if we change Reactome I should change my models accordingly.

[deustp01]

Shall we focus on this pathway on a weeds call?

Sounds good. Next week is Steven. Maybe start then if there's time and continue. B2 is a good use case for both the boundary and evidence issues.