I had an interesting trip a month ago, funded in part by the Australian Academy of Science. They require a report, and I was wondering if a blog post could double as a publicly-readable and transparent report. Given that the entire trip was for the purpose of open science, this seems appropriate. In the end, they have a proforma, so this is supporting information.
Sydney-Abu Dhabi-Geneva was followed (no pit stop) by a flight to Madrid to visit GlaxoSmithKline at Tres Cantos.
There is a group of chemists and biologists there who understand the idea that the free sharing of data brings about more collaborations and faster progress in science. This group really is extraordinary. Their deposition of antimalarial hits in the public domain in mid-2010 is worth thinking about for a moment.
Here is the sharing of bioactivity data and structures of thousands of compounds. These are not just commercial compounds bought in by GSK, but also synthetic compounds, made by GSK as part of various medicinal chemistry campaigns. Javier Gamo, the lead author of the GSK Nature paper, told me the story of how this deposition came to be, and described the extraordinary leap that occurred in releasing the data. We are all used to seeing talks by big pharma in which the structures of active compounds are not included or “R grouped out”. Indeed, it’s usual for the release of the structure of an active to be associated with a frustrating amount of paperwork. That’s even the case in academia, where I would be seriously advised against sharing data on unpatented bioactives. But here there are thousands of actives. Not only are some of the compounds highly potent, but they are whole-cell potent – an important new movement in the discovery of antimalarial compounds.
Now I remember when the release of data occurred. It was described as GSK going “open source” in many newspapers. That’s inaccurate. Data are just data, unless people work on them, in which case they become a project. Open source describes a process, an activity. To capitalize on the data released by GSK requires people to work on the data. This is such a rich field of hits that GSK don’t have the time to work through all the series alone. They are seeking partners – indeed that was the rationale for releasing the data in the first place. We’re now working with them, in an extreme form since everything we’re doing is freely available, rather than being part of a more traditional bilateral collaboration.
Our immediate work is to validate the hits. We’ve just finished their resynthesis, plus the synthesis of a few other compounds. These have now arrived at Tres Cantos who will do IC50s and progress the best compounds to rate-of-killing assays. Stuart Ralph and Vicky Avery are also looking at these compounds – we need to be sure of their worth if we’re going to look at them further. These are sensational in-kind contributions to the project.
The team at Tres Cantos have evaluated their “TCAMS” set and grouped the best compounds into sets. The arylpyrrole set that is the starting point for the open project is one of those identified. There are lots of others. When I was there the team (including Felix Calderon) told me they’d just published another study where a few of the sets had been evaluated. Interestingly the medicinal chemistry campaign was short and sweet – a small number of variations in different parts of the structure had led to shallow SAR (i.e. small changes in bioactivity from changes in various parts of the structure) and this is a negative for the start of a campaign. This series was abandoned because of this, coupled with there being plenty of other hits to go after. (We need to be following this model in what we’re doing in Sydney). What I found particularly interesting was one of the assays in place to determine the likelihood of resistance occurring to a given compound. This is possibly commonplace in medchem/parasitology, but it was new to me. Emergence of resistance to a known drug is apparently quite repeatable. One can effectively run an “evolution emulator” and re-develop resistance to known compounds. It is then possible to run a parallel experiment with a novel compound, and time the development of resistance/tolerance. A rapid onset of resistance essentially kills that series. This is what had happened with GSK’s evaluation of one of the latest TCAMS sets. Javier’s opinion is to include this assay earlier in the evaluation of hit series. It’s something we’re going to want to do with the arylpyrroles when we have a few more analogs ready to go.
From Madrid I flew back to Geneva. There is a strangely high concentration of important public health people around the airport. The following morning I had breakfast with Piero Olliaro from WHO/TDR to plan what to do next with our schistosomiasis project following the publication of our papers. We’re looking at some new ways of making the molecule. We’re also excited to be able to provide enantiopure PZQ to any groups needing some. Please contact us if you’re in need, though the procedure’s pretty easy and needs no fancy equipment. Over a coffee and a spectacular pastry we also discussed whether there was mileage in a consortium working on open source drug discovery for schisto. I think there is, provided the consortium is big, and is open. The “neglected” in neglected tropical diseases comes partly from there being a lack of interest in the development of new drugs. Interestingly one of the big drivers for new “schisto” drugs is the veterinary sector. Vet drug providers would be less keen on the openness, naturally. However, their customers might take a different view. Imagine a rotational dewormer identified by a robust, open research process, available at cost, funded in part by the increased productivity of the livestock/fisheries sectors.
Later that morning I went to the Medicines for Malaria Venture (MMV), another visionary organisation, strongly encouraging the sharing of data and a more collaborative approach to drug discovery. MMV are supporting our current project financially, and last week we just heard that we have secured 3 years’ funding for the project from the Australian Government. I gave a talk at MMV and spoke with our project champions Paul Willis, Jeremy Burrows and Tim Wells. They agreed that the key aims of our current project are 1) to conduct a kernel of activity of the project and 2) to leverage help from others, with particular value arising from practical input, e.g. the synthesis of compounds. It really is my vision for a project of this kind to involve other labs around the world in a coordinated effort, where all data are open and publication of milestones is rapid. If you want to come on board, you can.
That evening I flew to London for the weekend to see friends and family. It was bizarrely hot in London that weekend. Beautiful to see the new Shard going up rapidly, and I rediscovered the near-perfect Royal Oak pub in Borough round the corner from where we used to live.
On Sunday I flew to Barcelona, to the EU Congress on Tropical Medicine. I gave a talk on open source drug discovery which generated a lot of interest. It was great to meet up again with Jose Gomez-Marquez from MIT who’s taking a very interesting approach to DIY diagnostics, and whom I first met at SciFoo.
There were interesting talks at this meeting, but it had quite a broad focus, and there were a large number of policy sessions. I worry a little about the proliferation of groups, i.e. consortia, with names and acronyms, which meet and discuss and emit reports. The crucial feature of many of these organisations is that they are not open, i.e. they operate behind closed doors, and then broadcast. Frequently a funding agency and several universities/NGOs will get together to look into something, say the provision of new drugs for TB. But from my perspective there is little to be gained from my knowing that this organisation exists, because it’s unlikely I will ever be part of it, or influence its direction. Hence I don’t get very excited or involved, and I rather tune out, unfortunately. It did feel as though this meeting was particularly heavy on the “new consortia” announcements, none of whom seem to need anything from anyone. I think if we’ve learned anything in the past few years it’s that it’s important to allow people to input into processes and projects at any stage, and to be detailed about the research being conducted, rather than just summarising it at the end. If I were to advise these groups, it would be to release early and release often, and not polish the outputs too much.
There was a very interesting talk from Robert Jacobs from Scynexis about the development of a boron-containing drug for Human African Tryps (sleeping sickness). I hadn’t heard of a drug containing boron before, and tweeting on this subject led me, via John Overington, to a post on this very subject by Derek Lowe. I would love to know the point, in a med chem. campaign, when someone says “Hey, let’s try boron now”. It makes me think about the hits we’re looking at. Boron, anyone? Phosphorus? I experimented with Storify to put together the correspondence.
On the Wednesday I flew to Milan, and then took a train to Modena. This beautiful town was host to the EU COST Action meeting on New Drugs for Neglected Diseases. I had time to check into the hotel and walk to the conference venue before the first session began, and I was due to speak, again on open source drug discovery.
This meeting was more focussed than the Barcelona meeting, but again I had a lot of very supportive comments about the open nature of the research we were doing, and how openness removes many of the thorny problems of traditional research, such as duplication of effort, or roadblocks to progress because a team is not in touch with the external people needed at a given point. The question I received most, both in Modena and in Barcelona, was What about publications? How do you publish something that’s already out there? I was able to point to the two papers we’d just published the week before to say that this was not a problem, and that publication of open projects is extremely important for bringing people up to speed with where the project is at, as well as marking milestones.
Dihydrofolate reductase and pteridine reductase PTR1 were mentioned on multiple occasions as targets of interest for rational drug design, with the latter being particularly cool for doing two reactions in one active site – wow. And it also looked interesting from the point of view of maybe being able to catalyze the asymmetric reduction of dihydroisoquinolines, which is a separate project we’re doing in my lab.
There were a lot of nice talks at this meeting. My view of medicinal chemistry is now a little skewed, and I can’t listen to a talk from a closed group of students/academics about making a molecule against a certain target without thinking “Why aren’t you just sharing all your data and working with all the other groups looking at this area in real time, rather than slowly publishing and telling us about the choicest results at meetings?”
The COST meeting was extremely pleasant socially, too. You can tell you’re at a European conference when the entertainment is an opera recital in an old church.
And you know you’re in Modena when your pizza is drizzled in sticky vinegar that looks like crude oil. This one had lard as a topping.
On Friday I returned to Milan on the train, then flew back to Sydney. Always terrible losing a whole day on the return journey. The whole trip was tremendous, for many reasons. I met a lot of new people who were interested in helping out with our open projects. I had time to think about what we’re doing, and to receive advice and well-informed questions about the approach. Mind duly broadened. Thank you Aus Academy of Science and Leopold Flohe, organiser of the COST meeting.
There are so many things to do now on the open projects. Besides making new antimalarials, evaluating new catalysts and working to improve the electronic lab books we’re using, we need to recruit new labs to be part of the experimental effort and speed things up as much as we can. That’s our main emphasis in the coming months. We need students, undergrad lab directors, anyone interested in making compounds to join us and become part of a larger team.
In February next year I’ll be talking at the AAAS meeting in Vancouver on what we’ve been doing. We’re then going to be hosting a one-day open source drug discovery meeting on malaria on February 24th, and it’s great that Saman Habib from the malaria OSDD project in India will be attending. The whole thing will be streamed, for those who can’t make it. I need to get on with organising this, and if anyone has any experience in streaming and archiving a conference in which there is meant to be worldwide participation, I’d be delighted for some pointers because at the moment it looks … challenging.