The cost of validation

Many interesting speakers in this workshop in Dundee but most of them fall in the mathematics part of biomathematics. Among the few who do not is the biologist Vito Quaranta (Vanderbilt University). Although I have been told many times that things are changing for the better in that respect, the scarcity of life scientists and medical doctors in these type of conferences tells me that there is still a lot of work to do to convince them that computational and mathematical biology is not only relevant but necessary.

The talk from Vito Quaranta was not so much about science as about doing science at the interface between theory and experiments. He is lucky to count with the resources of the Vanderbilt Integrative Cancer Biology Center. Otherwise the problem of validating the mathematical and computational models with theoreticians come with would be next to impossible. This theoretical models make a number of assumptions about the properties of tumour cells, tissues and micro environments and predict outcomes that in many cases have to be contrasted with in vivo and in vitro experimental results. This experimental work is really challenging given the level of fragmentation of knowledge and expertise in biology and medicine. Different labs with different experimental techniques, machinery, cell lines and the necessary permissions to perform animal experiments and access human clinical data are required to validate one single theoretical model. That means that unless centres like the one in Vanderbilt become much more common most theoretical models will remain experimentally untested unless they proof to come out as the result of the consensus of the theoretical biology community.

Cooperation in a tumour and workshop in Scotland

I find myself in Dundee, in Scotland, attending a workshop entitled Mathematical modelling and analysis of cancer invasion of tissues. It promises to be an interesting event and some of the attendees are working on topics that are very close to mines so it is good to know what is their contribution to the state of the art.

When I arrived this morning I was expecting good stuff from people like Philip Maini (Oxford), Bob Gatenby (Arizona), Vito Quaranta (Vanderbilt) and Sandy Anderson. Still today’s most relevant talk for me was given by Anna Marciniak-Czochra (Heidelberg) who presented work based on the research presented very recently by Robert Axelrod (and reviewed in this blog here). Axelrod’s work is about how the collaboration between tumour cells could mean that cells do not have to acquire all the necessary capabilities (mentioned in Hanahan and Weinberg’s 2000 work) in order for the tumour to become agressive. This is a word model but in Marciniak-Czochra‘s presentation a mathematical description was shown in which the characteristics of the growth factors (eg. diffusion strength) can determine how useful this collaboration is. It looks like an interesting model and hope a paper will come out soon so I can take a look. Still it seems that a paper that covers Axelrod’s work more comprehensively is still work to be done.

The links between scientific disciplines

I have just found this amusing article about how some guys took more than a million and a half scientific papers from 776 different branches of science and came with the graph included in this post. You can find a larger version (5Mb) of it at this site. It shows how often papers from different disciples are cited by the same paper so it can give a measure of what fields are more likely to inspire interdisciplinary work.

Most of the research seems to be in Medicine and biochemistry (including all the -omics stuff). Math seems to be more unconnected to many other branches of science that I thought but to be honest I am not very sure about the methodology. More about it can be found in Mapofscience.com.

World Community Grid and Cancer

Since the guys at SETI came with the idea of using the CPU time of internet users when they are not using their computers, several other projects draw inspiration from this idea to obtain some sort of highly distributed high performance computing when the funding is not there.

One of these projects is called the world community grid which involves many research centres and universities and tries to tackle several problems that should be of general concern. One of the projects is about Cancer. One of the ways to go about cancer research is by using tissue microarrays in which samples of tumour cells are treated differently and the results of the different treatments can be obtained and compared in a comparatively efficient way. I am writing this from Columbus airport but when I get the chance of getting back to Dresden I should install this client on my Linux workstation. They do have versions for Linux, Mac and Windows.

Of course one thought is that if I know that I will not use the computer in a while the right thing to do (assuming one cares about the world) is to switch the computer off but I guess that those times in which the screen saver kicks in I would be happier thinking that my computer is doing something interesting instead of just displaying pointless and CPU intensive openGL pictures.

Columbus workshop and interactions with life scientists

Being a workshop on mathematical biology one of the issues we all face here is how to work with life scientists and thus one of the panel session yesterdays was precisely about that.

It seems that there are different kind of problems theoreticians might find when dealing with clinicians and experimentalists depending on a number of factors:

• What kind of people are they? Are they ‘math-skeptic‘? do they have affinity towards theory?
• Do you want them to share their expertise with you or do you want to influence the experiments they perform so they can be used in your theoretical model? The latter is significantly more difficult.
• Do you work with biologists or with physicians? There is a real difference between the average PhD and the average MD that does some research on the side when it comes to understand the usefulness of theory.

Some tips where also offered by some participants on how to make finding and establishing collaborations. Mainly it helps to attend seminars from the life sciences departments, get yourself familiar with their stuff and get your face known to them so you don’t come as a complete stranger.

Workshop in Ohio

I will not probably have much time for posts next week since I plan to attend a workshop in Columbus, Ohio. The workshop is organised by the Mathematical Biosciences Institute of the Ohio State University and is entitled Workshop for young researchers in mathematical biology (remember what the meaning of young researcher is from my previous post from my visit to Barcelona :().

At any rate there will be some interesting people both in the category of keynote speakers and “young” researchers. Some of them doing bio mathematics of cancer so expect a report on that when I come back.

cancer genes

The information can be found tailored for all types of users. For those who want an easy take here is the BBC version. Nature has a nice overview and the article proper.

Here is my take: a (fairly large) group of researchers mainly at the Sanger, in UK have studied hundreds of genes that are mutated in about 200 types of cancers. The trick here is to find what genes DO drive cancer as opposed to ‘just happen to be mutated’ in a cancer. At the end of the day your average tumour cell in an advanced stage tumour is likely to contain several mutations and many of them will probably be hitchhikers not necessarily contributing to the overall fitness of the cell. Unfortunately the result of the research is that the number of genes mutated in many cancers is higher than expected and telling apart driving genes from others will be a challenging task. One thing of working with so many types of cancers (200) is that genes that might not play any significant role in one type of cancer might turn to be important in the next.

Telomeres, cancer and aging

One quite fascinating thing in animal biology is the question of immortality or, to be more precise, the lack of it. While most unicellular organisms can divide for as long as they have the luck to find resources and space to do so, human cells can divide only a limited amount of times (approximately around 50 times, although this does not apply to stem cells that can divide an unlimited amount of times). In principle the limitation in divisions for most human cells is due to a mechanism that has been evolved and is not an intrinsic limitation. The cancer hypothesis is that the limitation makes the appearance of cancer more unlikely. If a cell is limited to just a few divisions, if it acquires a mutation that mutation is unlikely to spread to far.

The reason for this limitation are the telomeres, situated at the end of the chromosomes, that get shorter each time the cell divides. Once these telomores reach a critical size and become to small the cell will enter a state called senescence by which they will not divide again.

This is an interesting link in which they talk about this and how in the next few decades we might know enough about the effects of limited cell replication in human life expectancy, how to increase it (maybe for ever) and how to do that avoiding nasty side effects (like increased probability of dying from cancer). The website in which this is hosted is covering all sorts of news, many of them of dubious scientific interest, but the information in the link looks sound.

On the other hand in a more reliable source (PNAS) there is a nice study on how telomere dysfunction can cause genetic instability. They work on a disease known as Werner syndrome but it is quite useful stuff for cancer research. This Werner syndrome results in people aging prematurely and researchers at the Salk institute have found how extra short telomeres can be the source of the problem.

Bioinformatics and google

First of all, I am not a bioinformatician. I even thought I had an idea of what is the aim of bioinformatics but after a visit to the group of Nuria Lopez Bigas at the Biomedical Research Parc in Barcelona I realised that the scope of what I thought the field is about was too narrow (they do some nice machine learning and statistical work on disease related genes).

I have seen this video on Google video (only one of these sites that allows the user to download the video for offline use) some time this weekend. It is just a talk by David Vise, the author of a book about Google, to staff at Google Inc. Google is one company that fascinates (and in a sense, worries) me tremendously. As a side note, this blog is posted in one of their servers.

The thing is that at the end of the talk, David mentions the interests of one of Google founders (I think it is Larry Page) on biology. Then I thought (and I am sure that I will be the last of a long list of people) that, wow, that is really a good match: google and bioinformatics. Biology until know was mostly a science in which practitioners collected facts. There is loads of data and little idea of how to make sense of it, asides from evolution and a few other very general ideas. One of the aims of google is precisely to find patterns in the zitabytes of information stored in their servers. I have the feeling that we will see more of Google in that field.

Old Dawkins video on cooperation

I have found this video on google video. It is an old Richard Dawkins BBC TV programme in which he goes through some topics of which I am quite fond and which I personally think have some yet-to-be-explored relevance to cancer research.

The documentary was produced a few years after Dawkins wrote The selfish gene and not long after Robert Axelrod had written The evolution of cooperation. It takes from Axelrod’s research on cooperation (whose own take on how this can be observed in cancer has been mentioned in this blog before) to illustrate how cooperation might evolve in an place in which agents (say humans, bacteria or buffaloes) are selfish. The topics are the usual ones in game theory such as prisoners dilemma (how playing if for an undetermined number of times changes what is the best strategy), tragedy of the commons (if everybody is overusing a resource why shouldn’t you, and if few people overuse, why shouldn’t you since it will make no difference whatsoever?).

The tragedy of the commons seems to me a suitable game to model global warming (why should your country cut on carbon emissions if nobody else does or why should you if everybody does? this seems to apply to most countries but the likes of U.S. and China whose weight is to big to be considered just another player) or cancer (at the end, tumour cells do kill the host and thus themselves).