It is important to be extremely clear on the real meaning of the words: bioinformatics is only a term to intend the use of numerical calculus in biology, nothing more nothing less. Computational tools are routinely applied to cancer related themes, by the way it is probably the most studied area in systems biology and the number of microarray studies in cancer is huge, moreover P53 is the by far most studied protein (mainly by different kind of computational tools) and Protein-Protein Interaction network linked to cancer are widely studied, the same is true for more 'innovative' and 'theoretically new' approaches like fractal analysis of the shape of cancer cells or tissue organization field theory that have a very large computational flavour.

The basic point in my opinion is: 'Will computational approaches give an hand to move cancer research from the endless (and mainly unfruitful) mainstream research line of 'looking for cancer related genes' to introduce a really systemic, tissue based approach or we will only continue with the brute force of sequencing myriads of SNPs without any relevant synthesis ?' Rephrased 'Is biology sufficiently mature to accept quantitative studies in the core of hypothesis formation or they are only a posteriori 'ornaments' of already established paradygms ?'

At the risk of sounding blunt; if you think that there are no Cancer Bioinformatics specialists, then you have not been reading the literature. All of the cancer microarray and NGS-based studies have some sort of Cancer Bioinformatics specialist at the helm. And there are no shortage of those manuscripts in Nature, Science or any of the top journals.

Dear Paul,

Thank you for your answer, and its freshness.

I am actually very keen to know how many people share your opinion.

My subversive question is indeed aiming to raise the very specificity of Cancer which is too often treated as "more or less the same" than the rest.

While it is most definitely not!

Too often classic tools are used on Cancer data where general assumption of homogeneity, diploidy, purity, etc. do not hold. Some output comes out of these "analyses" but their interpretation is very commonly flawed.

Alas there seems to be too little recognition for the specificity of the Cancer Bioinformatics field. Most of the time, you find a "Cancer specialist" in a Bioinformatics group rather than a Cancer Bioinformatics group.

Doesn't Cancer Bioinformatics deserve more recognition ?

Dear Michael,

Thank you for your "blunt" answer.

I guess I should have declared that I firmly believe Cancer Bioinformatics is worth discussing as this has been my field of predilection for a decade!

So I am of course aware that there are many bioinformaticians dealing with Cancer out there and the very prominent Cancer literature that relies on them. However these bioinformaticians are seldom the main authors of these studies.

I have reformulated my question to emphasise my point on the limited number of Cancer Bioinformatics GROUPS rather than SPECIALISTS.

Still, it is hard to have real Cancer Bioinformatics Specialists without a Cancer Bioinformatics Group. As mentioned above the general (not always) trend is to have a "Cancer specialist" within a Bioinformatics group, or a "Bioinformatics Specialist" within a Cancer group. If you consider the breadth of Bioinformatics and Cancer fields you would have to agree that it is a lot of expertise for them to carry.

I believe there are too few Cancer Bioinforrmatics groups in the world, where only the Very large Cancer Centres do afford one. This limits the potential specific developments in the field to a select group.

Dear John,

I guess that I have been a little bit spoiled in the cancer centers that I have worked with. The Dana-Farber Cancer Institute (which is relatively small), the Broad Institute (considerably larger), and Stanford (larger still) all have dedicated Cancer Bioinformatics experts. These people have tended to branch out and do other things, but their primary role is analyzing large cancer datasets. Here at Stanford, our group operates under the "Center for Cancer Systems Biology (CCSB)" guise. This is part of a large NIH initiative that supports a number of similar centers around the US that specialize in cancer bioinformatics as part of the Integrative Cancer Biology Program (http://icbp.nci.nih.gov/). I am surprised that there is not a similar set-up at Oxford, or a similar collaborative in Europe. How do you operate there?

Michael

It is important to be extremely clear on the real meaning of the words: bioinformatics is only a term to intend the use of numerical calculus in biology, nothing more nothing less. Computational tools are routinely applied to cancer related themes, by the way it is probably the most studied area in systems biology and the number of microarray studies in cancer is huge, moreover P53 is the by far most studied protein (mainly by different kind of computational tools) and Protein-Protein Interaction network linked to cancer are widely studied, the same is true for more 'innovative' and 'theoretically new' approaches like fractal analysis of the shape of cancer cells or tissue organization field theory that have a very large computational flavour.

The basic point in my opinion is: 'Will computational approaches give an hand to move cancer research from the endless (and mainly unfruitful) mainstream research line of 'looking for cancer related genes' to introduce a really systemic, tissue based approach or we will only continue with the brute force of sequencing myriads of SNPs without any relevant synthesis ?' Rephrased 'Is biology sufficiently mature to accept quantitative studies in the core of hypothesis formation or they are only a posteriori 'ornaments' of already established paradygms ?'

I guess Bioinformatics has been widely used in Cancer and many more diseases. The sophistication of computational approaches has made it possible to analyse the huge data from sequencing,remember that data generation is not a problem the problem is how to validate it.The justification of using Bioinformatics approach in Cancer studies is very widely accepted because through this the lab data could be validated easily...so it's worth of justification...according to my perspective.

Dear Michael,

Thank you for this clarification. You are indeed spoiled for choice as you pretty much described most of the groups I was referring to! However those group appear to hide under a different name than Cancer Bioinformatics.

As for the way we are organised in Europe, I am afraid giving you a list of groups is likely to upset whoever I might miss...

In the UK Cancer Bioinformatics groups are usually linked to the major Cancer funders. In Oxford we are supported by specialised departments and top Researchers.

Dear Alessandro,

Thank you very much for your contribution.

It is good to remind us that Bioinformatics is an older and broader field than the flavor of the moment, or the author favorite angle.

You appear to agree with my point that both Cancer and Bioinformatics are, have been, and will be, heavily used together. However this link, a specialist link, is rarely visible because of the various guises "Bioinformatics" can take.

The fact the specialist link is rarely visible is a good thing not bad, I hope specialisms will soon or later fade away, science fragmentation is one of the worse plagues of our work...

As a bioinformatics generalist, I have some limited experience of work with cancer-derived data sets. I can see that the issue of copy number variation is particularly important in expression in cancers, and this issue requires extensions to analyses of expression data. This might be called 'cancer bioinformatics'. However, I can also imagine that techniques for inferring copy number variants from expression data sets would also be useful in looking at other contexts, such as bacteria with varying numbers of plasmids, or varying sized arrays of tandem repeats of genes in plant genomes.

There has historically been a lot of investment in bioinformatics motivated by the desire to understand cancer. As perhaps the leading recipient of public bioscience funding, cancer research has also been one of the most productive engines for bioinformatics developments. But, the techniques developed to understand cancer usually tend to be more broadly applicable. The Broad and Dana-Farber were mentioned earlier, and as bioinformaticians, many of us have benefited from their work, despite working on other organisms and contexts.

If bioinformaticians wish to personally devote their skills to understanding cancer, and meet and collaborate under this umbrella, then I think that makes a lot of sense for them. We generalists will all be watching so we can borrow, bend and adapt the techniques they come up with. Maybe we might ask that they also keep half an eye on the wider horizons, and remember the public good that can come from them considering the wider context. If their tools help in the cure for cancer, that's great, but some of us are working on other problems, such as food security and antibiotic resistance, that are arguably equally important, but much less well-funded, and we need all the high-quality tools we can get our hands on.

Dear Jonathan,

Thank you for your contribution to the debate.

It is good to know as well that Cancer Bioinformaticians can be of use to more general field.

However I would like to avoid building an iron curtain between the Cancer specialists and Generalists. I see the Cancer as a speciality WITHIN general Bioinformatics, just like Data Mining or Computational Biology. Bioinformatician should be free to move back and forth this speciality, and keep on communicating around. This is as long as the specificity of Cancer is recognized, just like Data mining or Computational Biology.

As mentioned by Alessandro one should not fragment the field too much, but not assume one can analyse Cancer data without understanding (some of) its biology.

Dear A. Groban

Thank you for this very interesting reference.

It seems that Cancer is more often associated with Systems Biology than Bioinformatics, recognising de facto its specificity.

I agree to Alessandro Giulliani and more gentle position of Jean-Baptiste Cazier that (as I understood) too ardient elementaristic Bioinformatics can rather hide than reveal ( it depends on the user, of course) the general biological rules which have to be disclosed first of all at the level of Systems and therefore initially by rather General methods. The failure to combat metastatic cancer inspite of huge efforts (including Cancer Genome project) means that we still see mostly trees, not the whole forest - we have not formulated the really causial theory of cancer. For an example of such holistic approach, see our hypothesis: Cancer: a matter of life cycle? (Erenpreisa J and Cragg MS, Cell Biol Int 2007, 31: 1507-1510), attached at my ResGate homepage. It is getting some support now (Lagadec C, Vlashi E, Della Donna L, Dekmezian C, Pajonk F: Radiation-induced reprogramming of breast cancer cells. Stem Cells 2012, 30(5):833-844; Zhang S, Mercado-Uribe I, Xing Z, Sun B, Kuang J, Liu J: Generation of cancer stem-like cells through the formation of polyploid giant cancer cells. Oncogene, prepublished March 25, 2013).

I think it is extremely useful to discuss bioinformatics methods that are most suitable for cancer data but I'm biased of course as this is my field of application

:-)

Just to add two other references in the area:

"Cancer systems biology, bioinformatics and medicine" Alfredo. Cesario, Frederick B. Marcus (http://link.springer.com/book/10.1007/978-94-007-1567-7/page/1)

and a recent workshop (http://videolectures.net/cancerbioinformatics2010_cambridge/)