Let’s hear it for interdisciplinary science. Everyone says what a good idea it is. The research councils strategic plans tend to laud it. And yet, and yet….Do they mean it?
Last week I attended an excellent conference in Oxford. Entitled Physics Meets Biology, it was organised by the Biological Physics Group (BPG) of the Institute of Physics. As its name suggests, and as the group organising its name also suggests, this is inherently interdisciplinary science. It’s exciting stuff (you can see why I think so on my first blog at the Guardian), the field is moving and expanding and more and more groups are establishing themselves in the UK in this area. The BPG was set up in 2007 – I was its first chair, but passed the baton on a couple of years ago to Andrew Turberfield – and this was the third in a series of conferences with the same name, run at 2 yearly intervals. During the 5 years of its existence the group has built up a thriving community, or perhaps more accurately acted as a focus to bring together disparate groups which were already extant. There’s lots of neat stuff going on in the UK and what really struck me at this year’s meeting was the number of different centres which were presenting talks and posters; there were also a good number of early career researchers. When there are a few key centres dominating an area it seems to me that there is the danger of heading for fossilisation. When the research base is broad and the average age relatively young, it means that new blood, new ideas are constantly entering the field. Many of the invited speakers were international, but the bulk of the attendees were UK based. The organising committee had ensured a good gender balance of excellent speakers, and there was no sign of women holding themselves back in the questions sessions either.
One of the invited speakers was Doug Kell, Chief Executive of the BBSRC (he has written briefly, but enthusiastically, about the meeting on his own blog). He gave an interesting talk largely derived from his early work, well before he assumed the mantle of running the Research Council, particularly looking at bioinformatics approaches rather than physics per se. (For any international readers, it is probably worth spelling out that the BBSRC stands for the Biotechnology and Biological Science Research Council, because one of the US invited speakers asked me afterwards what the acronym meant; it is too easy to take these things as read.) The few slides he did show from the BBSRC perspective disappointed me, though. He quoted some large number of millions of pounds that he alleges the BBSRC are spending on ‘biophysics’, and then elucidated that most of that was spent in Biochemistry Departments. Without wanting to denigrate what may be being funded under this programme, it is disappointing to realise just how little appreciation there is of the difference between the kind of research being discussed at this meeting and that traditionally done in Biochemistry departments. Biological physics, as the BPG mean it, is not the same as traditional biophysics of the nerve impulse/ion channel/measuring signalling pathways variety done in Biochemistry and allied departments, even if it is now coupled with cutting edge fluorescent microscopy. The latter gets funded quite substantially by the BBSRC as Doug pointed out; rather little of the former kind of biological physics does.
One difference, that may seem subtle but is actually quite fundamental, is that much of the biological physics that was presented at this meeting would probably appear in a grant proposal as ‘we want to see what happens when we do X’, or ‘we have a new tool that should cast light on Y’. It might be worded a little more precisely, but it often is not of the ‘my hypothesis is that…’ variety, and hypotheses are rather explicitly required (eg on referee forms) by the BBSRC. I have been annoyed by this ever since I sat on some previous manifestation of a BBSRC committee more than 10 years ago. Then, if the hypothesis was stated as ‘if gene A is down-regulated I predict there will be an effect’ it seemed to be regarded as adequate, even if what that effect was likely to be was never explained. If it was,’ I can now image at unprecedented resolution and with chemical specificity and I will examine what happens when I add molecule J to cell K’, the lack of a hypothesis meant no grant was likely to be awarded. Yet, in many cases I would hazard a guess that the second proposal was an awful lot more exciting than the first, which too often can look like nothing more than sausage machinery. It is deeply disappointing that a tribal hang-up about hypotheses still lingers round the BBSRC, even though the nature of so much of science has changed. Physics, indeed the EPSRC, do not have this same fixation.
I pushed Doug over dinner about the fact that the figures he quoted did not relate to ‘biological physics’ as this particular community meant it, and our work tended to sit uncomfortably at the join between BBSRC and EPSRC (the Engineering and Physical Sciences Research Council) . I wasn’t expecting to make much headway because it’s an(other) argument I‘ve had before with senior BBSRC staff. The answer I got was similar to the one I got when still Chair of their Committee C a couple of years back: ‘Why should our particular interdisciplinary field be regarded as in need of attention more than any other, or deserving of special treatment?’ I find this answer deeply disturbing. It is implicitly saying that they know that there is a problem, but because there are many junctions where there may be problems, they regard it as too difficult to do anything and so they’ll look the other way and stick with the easy mainstream topics. Other concrete examples where interdisciplinarity causes problems, to my certain knowledge, can be found in the way research around good health, as opposed to disease, is handled; this is neither within BBSRC or MRC remits and yet ought to be something that is viewed as crucially important. A second example is what happens for topics that sit between BBSRC and NERC (National Environmental Research Council). In other words, whatever the problems may be within a single research council and its constituent grant-giving committees, it is infinitely worse when an area sits at the interface with another Council. Yet, reading their strategy documents over many years, lip service is paid to the importance of interdisciplinarity. In practice there are far too many cracks down which excellent science can disappear.
Of course, the BBSRC is not alone in failing to match action to its words on this front. When Dave Delpy took over as Chief Executive at EPSRC I heard him say that he appreciated there was a problem with the refereeing of interdisciplinary proposals, and that maybe a specific group of people who were capable of taking an overarching view of things – as opposed to some narrow proprietary interest – should be identified to be charged with refereeing such applications. It sounded plausible, but there has been no movement on this front that I am aware of. All I hear is that, since it is the community that is doing the refereeing, the problem must lie with the community. I am not convinced. It is EPSRC staff who choose the referees (even if applicants nominate some, they aren’t necessarily used and others will be added into the mix by the staff).If EPSRC personnel play ‘safe’ and ask for people from the core disciplines (or even if they do this because they aren’t familiar enough with the science to know any better), and if they don’t keep track of the fate of those applications that don’t fit squarely in obvious categories to see if they are being systematically disadvantaged by the referees chosen, no progress will ever be made.
It is interesting that the EPSRC has recently identified the Physics of Life as one of its so-called Grand Challenges. At the moment this exists as nothing more than a network, funded initially at a level that will simply facilitate meetings to bring the community together to allow them flesh out what key questions should be tackled, coupled with a bit of inter-lab travel for feasibility experiments and discussions. I trust that this means that, as the network does identify key questions and routes to their solution, appropriate funding, possibly ringfenced, to allow the research to be carried out will be put on the table. If not, this effort is doomed to failure. Maybe my pessimism over what has happened in the past in this field now belongs to the past; it would be nice to think so. I will certainly be interested to see what transpires. But if, as has happened too often previously to me and many of my colleagues, disputes arise as to whether proposed research fully fits into the EPSRC remit or has too much biology in it to find a home there or, because it mentions words like cancer or other diseases, it is really MRC work, then real innovative science to solve the Grand Challenge will be blocked. That is where ‘falling down the cracks’ is so particularly frustrating.
These arguments, in case there is any doubt, have nothing to do with the long-running sore about Impact. In fact I think the field of biological physics is rich in examples of how there may be long-term impact from the research being carried out, but that is neither here nor there for the points I am trying to make. A large part of my frustration with both research councils is that they seem to think an argument that says all proposals will be found a home in one or another is a sufficient response to the problem of the discontinuity between them – it isn’t. Finding a home means that applicants end up having to distort their science (and therefore not necessarily apply to do the best science they could; I am of course referring simply to responsive mode proposals here) to fit one remit or the other. Otherwise they are told it is ‘too biological’ for EPSRC, but ‘not biological enough’ for BBSRC. But it is also the case that the refereeing and panel process are not suited to deal with interdisciplinary projects. The style of writing may differ from mainstream topics, but of itself that tells the reader nothing about the quality of science. It is merely interpreted as ‘failing’ because single subject science is seen as the correct default position. That is like saying women make ‘bad men’ because they’re different. In social sciences this is known as the deficit model, and is as bad an argument for science as for gender.
I am told not infrequently that the research councils are doing the best job they can; that they try hard to satisfy everyone. Those statements may be true, but that doesn’t make it a level playing field for people moving into new areas which aren’t properly represented on referee lists, or on panels actually giving grants out. I am afraid this is just a manifestation of ‘regression towards the mean’, where the mean defines mainstream science. It is a sure-fire route to stifle novel science and simply maintain the status quo.
Although this post is written largely around one specific scientific angle, I think the problems are endemic for any science at the interface between disciplines or research councils. It is a generic problem. We should have a seamless funding landscape and we do not. (The US is now doing much better on this front either than we are, or than they used to do, at least in the area of what I think they call the Physics of Living Systems.) Watching the way some chemists, synthetic chemists in particular, have reacted to any suggestion that funding in their (one might say traditional) field might be slightly cut, indicates just how much conservatism is endemic. They have acted as a lobby group, though not necessarily very successfully. Maybe it is time those of us who do not fit squarely into any box to stand up and be (even more) vociferous too. Not to use special pleading for our field to be particularly protected, but to call for practices to be as fair to us who work at the margins of any given discipline and who cross boundaries, as for those who sit squarely in the middle of some well-established subject. As things stand that is not the case, it hasn’t been the case in the recent past, and the research councils show no obvious sign that things are likely to change radically to back up those warm statements about the importance of interdisciplinarity.
On a completely different topic, readers who enjoy my thoughts on gender issues may be interested in my recent guest blog over at the BMJ website: ‘Science it’s a Girl Thing’ is not a cure.
The same problem applies to much of psychology. The dept I’m in has had people funded (in the past) by MRC, ESRC, BBSRC and EPSRC but as psychology is not central to any of these councils, there is a risk in every grant of being rejected as outside the remit. And research that spans psychology and another field can have an even worse time.
Another example of this generic issue is the field of biotechnology. Medical biotechnology has largely failed to become established as a field of academic research in the UK, despite BBSRC apparently having been designed to fund such research. The application of sciences, such as biophysics, to health has been one of the great successes of science, spawning an entire new industry, but after strong interest in the 1980s little research is being funded today because of fragmentation of funding between EPSRC, BBSRC, MRC and TSB. As stated above, biotechnology is not top priority for any of the funders. In the case of biotechnology, interdisciplinary research including medicine, biological and physical sciences and engineering is needed. This is routine in industry but very difficult to fund in academia.
The problem with being hypothesis driven is that you seek to confirm you hypothesis rather than testing all of the other equally likely hypotheses fairly. To rip off an example from Derren Brown:
You have a set of double sides cards which (as far as you know) may have numbers or letters on either or both sides. You can see 5 of these cards with only one face visible.
You can see: A 3 5 D 2.
What is the minimum number of these cards you need to turn over to test the hypothesis that odd numbers always have a letter on the other side?
So you immediately think of turning over the 3 and the 5.
Maybe you then think about turning over the A and the D….but actually they can’t either confirm or deny your hypothesis so you don’t bother.
You likely miss the fact that the card showing 2 should also be turned as, although it cannot confirm your hypothesis, it can certainly contradict it.
If you compare this process to one in which you say ‘here are 5 cards, try and deduce any rules linking the two sides of each card’, it becomes obvious that the introduction of a hypothesis has biased the process of ‘research’ and is likely only to find the self-confirming information available over the possibly contradictory information.
True science is OBSERVATION driven. In this case, once you have looked at both sides of all available cards, THEN you can start to try and deduce the connections.
One issue may well be that we choose to have separate research councils for different areas (and possibly this is what you were implying in your post). I imagine that this is because we fear that small areas would suffer if we had fewer, broader research councils. The US, on the other hand, seems to survive quite well with the NSF being one of the main funders of research. What does help is that you also have NASA funding space science, the Department of Energy funding some particle and nuclear physics, the Department of Defense funding some areas (high performance computing was one I came across, but there may be others), and there are probably others that I don’t know about. My gut feeling is that something like this would be preferable, but most are concerned that they would lose out if such as a system were introduced here and so feel safer with more research councils, with each covering reasonably specific areas. My view is that there is some merit in this concern and that the real problems is that we’re not (as a country) investing enough in research. Maybe a solution is to aim to convince the government that we need more investment in research but to aim for new money to be focused in interdisciplinary areas, rather than simply topping up the existing system.
Pingback: Systems medicine, polyomics and funding interdisciplinary science | Professor Douglas Kell's blog
I am a physicist by training and came to Cambridge to do a PhD in Physics but then changed to the Engineering Department and wrote my dissertation about the appliability of engineering design to the design of integrated health and social care services for people with intellectual disabilities. I did the latter as part of an interdisciplinary project between the local Cambridgeshire mental health trust, the Judge Business School, the Institute of Public Health, the Department of Psychiarty and the Engineering Department. So pretty interdisciplinary. I ended up doing a postdoc in the Department of Psychiatry for 6 months – until I left academia and became a management consultant. My main frustration with interdisciplinary science is that there are grants for individual projects but once the project is over, you end up not belonging anywhere. People kept telling me that my work was interesting but there was no career path out there – both in terms of places to publish and places to apply to. I was not medical enough for a psychiarty department/ journal, too medical for an engineering department and did not have the necessary background for a public health or business department.
This is an excellent and extremely important article. On the physics and biology issue this absolutely critical to the future of disease treatment in our view and your revelations are very disappointing but not surprising.
Perhaps we can supply some science to support your point. The genome has seen a bit of an obsession with hypothesis based on genetic differences that co-occur with the disease state or some other biological function. The elephant in the room is that life is driven by proteomics, genetics are 1 step removed. To quote drug baron blog;
‘Proteins are the stuff of life – responsible for every key process in every living cell. Nucleic acids, in all their diversity from DNA in chromosomes to regulatory microRNAs, exits only to create the right proteins in the right place at the right time.’
The complicated engine of the proteome is a complex system which cannot be reconstructed in terms of how it works by prodding at the taps that feed it (genes). Genetics can provide clues but you are going nowhere unless you address the issue square on – how does proteomics produce function in an nuts and bolts engineering sense. Sadly the preferred pathway approach has an extremely low probability of working as per the laws of complex systems science. All of this is validated by the failure in the clinic time and again of drugs that looked promising at the pathway hypothesis stage as validated in simple models like cell lines and some animals. Biology and evolution look sadly much more complex than this simplistic linear approach.
Therefore you need tools that interpret proteomes – complex systems science; tools that can properly interpret how compounds interact with biology – see Professor Andrew Hopkins in Dundee as an example and tools that provide proteomic data of the sort that really counts. The latter needs advances in the physics tools for biology, to be able to supply interaction and protein information at the detailed and systems level. Conformation, structure, dynamic interaction, type of interaction, mutation identification. This is where the UK’s historic excellence in these disciplines could gain a huge head start and is a key pillar of the future of these enormous industries. If the government is serious about economic growth in the long-term, look no further. The fact this is playing 2nd fiddle to genetics pulling at hidden threads is not good news at all.
We also believe that many of the challenges we face e.g. energy security, climate change, food shortages, disease, water supply, cultural clashes can be aided by advances in science. Nearly all of the step changes in science needed will believe are likely to require interdisciplinary solutions (complex systems science involved many disciplines as an example). The emphasis on applied science with a clear route to commerce in a 1 fits all research funding proposition (if this is what is occurring) with disadvantages for interdisciplinary science and non hypothesis science will have grave consequences. There must be a dedicated space for blue sky science research if the UK is to have a part in the major prizes of the future.
We wish you every success and hope the BBRC can see the light, the science and the industry direction rather than just appease the traffic.
I tried posting this on the Science it’s a Girl Thing’ is not a cure article’s webpage, not expecting its approval by the pending moderator’s filters because its contentiousness. That was two days ago so I’ve posted it here instead.
Since my teens, I had never liked angular high heels and could not empathize with the general sentiment expressed by my peers that it in any way enhanced the sexual beauty of a woman. It in fact detracted from that beauty if something. As I got older my sentiments remained the same even expanding into the possibility of making money out of the purchase and resale of the website http://www.unisexhigheels.com which I owned for a few years. I was and am a confirmed democratic capitalist and the notion of the proportional efficiency of having different ideologies within a democracy working together, from which an individual could profit from, had intrigued me for quite a long time. I could affably mock the world and make money at the same time.
I relinquished ownership of said website a few years ago. Relatively recently, having seen a very gifted BBC presenter who is very tiny in stature needfully wear high heels as she straddled beside a taller working colleague, I saw that there might be a practical need for high heels. I concluded that if the wearing of high heels is an issue of it being a practical thing to do, they ought to be horizontally platform based. Believe you me, still the democratic capitalist, that I have very clear pictures of plausible patents in my head. Flat platform high heel shoes avoid the inevitable outcome given by the prolonged excessive usage of angular high heels in the deformation of the foot, so redolent of the Japanese foot binding ritual. If you look at photographs of the feet of women that were foot bound from an early age, their feet are almost exact replicas of a modern women’s foot in a high heel (See photo link below) as if in some deranged jest. It is quite shocking to look at and in spite of feet binding having been banned for some time, they cause doubt that women who consensually wear high heels are perpetuating their own subjugation in a nowadays ubiquitous, still male dominant culture. Do wear high heels if you are short. Technology is synonymous with being human and there is no shame in that, but please do not let an atavistic like fashion out do you out of having healthy feet. Wear flat platform high heel shoes instead.
The lack of objectivity in the video given is indicative of the deficiency in understanding the challenge in encouraging women to take up science as a career. A purely objective intellect is asexual in that women can soak up the same intellectual information as men and equally understand it. Whomever contracted the filming crew had no idea what they were doing and obviously depended on an outdated subjective view of the world they so erroneously portrayed. To visually sexualize scientists and sell this to the general public is to encourage the same unequal opportunities for employment prospects in the work place that have plagued us since women got the right to vote and before that. Please, call a scientist a human rather than label as a man or woman with the corresponding enforced and stereotypical superannuated role.
To see an example of the effects of foot binding click on http://wahoha.com/static/images/ci/47104-220×220.gif