Difference between revisions of "Therapeutic Applications of Computational Biology and Chemistry 2007"
|Line 52:||Line 52:|
The Company is based in Cambridge.<br>
The Company is based in Cambridge.<br>
Latest revision as of 06:12, 15 April 2019
11 to 13 March 2007 at the Wellcome Trust Conference Centre, Hinxton, Cambs, UK.
Attended by Jennifer Clark with poster: 'Does your Species Show in GO?' and for some parts by Midori Harris.
This conference was split between discussion of really detailed chemistry and very interesting discussion of how bioinformatics can affect the cutting edge of medical treatment. Take home messages:
1) Drug discovery urgently needs bioinformatics support. They are now very aware of the fact and are looking for help with this.
2) There is a shocking difference between the quality of proprietry tools and the free equivalents provided by academia. This is seen to be holding back medical advances.
The conference was quite small and mostly attended by very senior people in big pharmaceutical firms and some sales people with stands.
NCRI Call for Grant Proposals
The most interesting thing that I found out is that there is a call for funding proposals for any bioinformatics infrastucture system that can help bring cancer information together:
An academic called Richard Begent
was very interested to pursue this avenue in collaboration with the GO. He says that there will be 19 new centres opening up to tackle cancer issues in the middle of April, as they have just got funding. He is keen on GO but feels that we cannot hope to annotate to completion with the current system. However he suggests that having one trained annotator per institute to talk to the biologists and make the annotations might solve this problem. He knows one case of this setup already. I suggested to him that we might write a grant proposal to improve the signaling terms in the GO and to train an annotator in each of his 19 centres. He said he thought that was a very interesting proposal that we should discuss it more. I think it would be helpful to talk more about exactly what we can do for this community before going ahead as he seems to understand the field very well and might be able to give us better pointers if we talk more about what's happening on each side. Richard Begent is based in London so is easy travelling distance from the EBI and would also be very easy for Ruth Lovering to include in her immunology work. I mentioned her plan to him and he said he would be interested to be involved.
One talk showed that the prognosis in a cancer case can be very accurately judged by looking at 6 specific pathways and several of the pathways were signaling. It seems as if getting our signaling terms worked out could help with cancer research.
I also met a man called Mark Forster who is quite senior in Syngenta at Bracknell. He says they looked at GO but decided not to use it because it did not have the terms that they needed. He says they need the process terms to cover the processes that would be affected by herbicides. I asked if he would be interested to have someone visit to add the these terms to the GO on the fly and he said 'yes probably'. I will contact him to see if he is still interested.
Peter van der Spek
There was a good presentation by Peter van der Spek (Erasmus Medical Center, Rotterdam, The Netherlands) "From molecular phenotyping to clinical benefit"
He says that in his hospital in the Netherlands they can use microarrays and ontologies to figure out which 7 in every 100 children with leukemia can be cured with a B vitamin suppliment instead of a bone marrow transplant. However he said that that is only possible because he brought very expensive tools with him from his industry work and that these tools are not normally available. Apparently the american military has an awful lot of data and some excellent tools for this kind of thing and that he has access to them because of his previous work. It was clear from the conference that getting such data gathering software into the NHS and publically available is one of the really important challenges for the pharmaceutical industry as it would allow them to capture data that they don't currently have access to.
There was another company called Ingenuity Systems demonstrating a tool for research. The tool incorporated a biological ontology which mirrored GO but was massively extended in-house. Apparently they have had 160 people developing the ontology for 8 years and now it is basically finished and they are down to only 80 people mainting the ontology. It contains 600,000 terms. They have 35 core molecular biology journals fully annotated as far back as electronic records go and they annotate another 100+ as required to cover particular processes. The tool has an annual academic licence fee of £4000.
The annotators are hired as short term consultants as they are required. The tool interface did not show the ontology structure at all but in response to a search just showed list of the term names broken down OBOL-style like this:
regulation of cyc, dich, rad, tcl7, hacg.
inactivation of cyc, dich, rad, hacg.
binding to cyc, dich, rad, tcl7, hacg, sonic hedghog, cheap date.
So the genus of the term was in a column on the left and the differentia that fitted with the search criteria were listed on the right. The differentiatia were hyperlinked. If you clicked on a gene name then you got to the list of annotations which really looked very much like the list of annotations in the current AmiGO.
There was a company there advertising a tool that uses GO, though no specific feedback on GO was available from the representative.
The Company is based in Cambridge.