Steve Oliver's Transport Group
Steve Oliver, who I met at the EuroFung Aspergillus annotation meeting, is interested to initiate a GO transport meeting to look at some aspects of function and process. He is now based in Cambridge and we have agreed that now is a good time for us to look at doing the work. I am in contact to try to figure out a good time to visit.
18th January 2008
I visited Steve Oliver at the Biochemistry Department (new building) on Tennis Court Road. I gave him a tour of the new high level terms and he was impressed with what we have done so far. We talked about how annotation should work where drugs are being transported by transporters that normally transport non-drug solutes and I explained about our plan to structure the graph along chemical structure lines, rather than have people annotate to 'xenobiotic transporter activity' or 'drug transporter activity'. He said that would work best for them too. He has a colleague called Paul Dobson who has recently completed a review of the transporters and Steve is wondering if Paul would be prepared to work with us to update the terms, as his knowledge is now completely up to date. He is going to ask on Monday 20th when he visits him in Manchester. There is another colleague called Andre Goffeau who might also help. I have searched for Paul Dobson's review but it doesn't show up in searches so I don't think it can be in print yet. It is in one of the nature review journals.
28th March 2008
Steve Oliver wrote to say that he has spoken to Paul Dobson (paul.dobson AT manchester.ac.uk), the post-doc in Manchester to discuss our ideas.
He is also interested in possibly holding the next EuroFung annotation jamboree at the EBI. I wrote back to say that all of this sounded excellent, and would he like some information on costings for the jamboree. I explained about all the infrastructure that we have set up to support this kind of thing.
10th April 2008
I am now in correspondance with Paul Dobson and am reading his review paper on drug transport. 'Carrier-mediated cellular uptake of pharmaceutical drugs: an exception or the rule?' Paul is a bioinformatician who has been reading a huge number of papers to write this review and so is very up to date. He is currently on a transition between grants, but is going to get back to me soon.
6th May 2008
Paul Dobson sent this e-mail. He has been looking at the GO graph and is exploring some of the transport terms.
1. Does GO:0005215 cover non-transporter-mediated movement (endocytosis being an example)? The synonyms and child terms lead me to believe not, but the definition seems to permit this. 2. I'm not clear on why there is a distinction between substrate roles in an ontology of molecular function (particularly things like odorant transporter). Does the wider biological role of the substrate matter in describing the molecular function? I suspect this is a failing of my understanding of ontologies. Thinking particularly about my drug stuff, it isn't the *application* of the drug that is important but the *pharmacophore* recognised by the transporter (for example, PEPT1 transports antibiotics, dipeptides, analgesics, antivirals, and so on, so a more useful description might be 'transports substrates containing a peptide bond', though this is too simplistic a description of the pharmacophore). I'm not at all sure how you might go about encoding the pharmacophore within the ontology though and a better solution might be to list everything that is a known substrate. 3. Why are drug- and xenobiotic-transporter activity siblings? Related to this is a consideration of the point at which a drug stops being a drug and becomes a toxin (a function of dose). 4. I'm not sure whether you might want to capture something about mechanisms of transport, as in passive/active, uniport/symport/antiport, etc. (already quite well described in the TCDB classification). 5. Electrochemical gradient is perhaps a more useful concept than concentration gradient in defining passive transport. 6. From the top-level definition transport activity can refer to movement into, out of, within or between cells and (unless I missed it) it might be useful to incorporate these concepts beneath the transporter activity root. I think my main concern is the one about using substrate roles as a way of describing the molecular function of transporters as many (but not all) are not intrinsically related to transport and could fulfil the same role even if they crossed membranes by an entirely different mechanism. I can see the value in describing things like 'siderophore transport' though.
We are going to follow these questions up by skype as there is a lot of material to cover.
29th May 2008
Participants: Paul Dobson, Jennifer Deegan.
On Skype text chat.
We discussed some questions that Paul has about the transport graph. He is rapidly getting to know the GO structure and we explored several issues, like the difference between function and process. He suggested the following paraphrased structure:
[i]secondary active transport process
---[hp]generation of ion gradient by atp expenditure function
the structure is currently like this:
[i]secondary active transport function
[i]generation of ion gradient by atp expenditure function
We also did some work getting Paul set up with OBO-Edit and by the end of the meeting he had the ontology file loaded and graphviz working and the better layout installed. He was also carrying out compound searches for terms.
3rd June 2008
(Contribution to wiki from Paul Dobson)
Here is one way of arranging the information (not an ontology as I don't know enough) that I find rather gruntling. Noting that my main work in this area concentrates on drug uptake by SLCs, I am sure that the wider transporter community will have further changes.
+Passive carriers/Facilitated Diffusion
++(consider that further descriptions might consider gating mechanisms?)
++(see question 2 below)
++Primary active transport
+++...(consider that further descriptions might consider drivers: ATP, light, etc.)
++Secondary active transport
1. Why is uniport sometimes said to be a secondary active process?
2. Are symport and antiport necessarily coupled to *active* transport? By *active* I mean that they are coupled to a gradient that is set up at some energetic cost. It leads to me wondering about how secondary active transport is best handled - is the source of the gradient all that relevant to the molecular function of the transporter?
Here are some sites that might be of use
3rd June, 2008
Participants: Paul Dobson, Jennifer Deegan
Paul explained his academic background. He has a biochemistry degree and a bioinformatics PhD, and works with Robert Stevens, so is quite experienced at discussing ontologies. He also works with Doug Kell who is in contact with ChEBI.
Set up a global render to show the transport terms that have been recently overhauled.
General Transporter ideas
Discussed a particular gene product that Paul suggests may not currently be covered in our system: http://en.wikipedia.org/wiki/NaKATPase
The diagram of the mechanism of this gene given on wikipedia suggests that it be primary active transport function with a coupled dependent symport function. We considered whether this should be separately annotated to two function terms or if we would need a single term to cover both functions. We did not reach a concrete conclusion.
Discussed the drug transport terms and the action that has been taken over these and the move to mirror the ChEBI structure under response to drug. We considered whether this would work for drug transport.
Paul suggests that we either need to do something like this or else classify drugs by their pharmacophore:
"The pharmacophore is the set of essential structural components required to be a substrate [for a transporter]. For example, PEPT1 (almost) requires the peptide bond."
Other characteristics of interest include "shape, volume, charge distribution"
Participants: Paul Dobson, Jennifer Deegan.
We discussed further the interactions with ChEBI and the GO. Then Paul showed the supplementary to his Nature paper, which contains essentially a large set of manual annotations to GO terms (which may or may not yet exist). The annotations are of gene products from human, mouse and rat to transport terms. All are involved in drug transport. I have passed the text on to Emily Dimmer of GOA so that she can give me feedback on how easy it would be to convert these to actual annotations.
This paper has been sent to Harold Drabkin, Victoria Petri and Emily Dimmer, cc'd to Paul Dobson so they can ask any questions.