Computational Methods in Translational Research

Susan Boyd


We know all too well the changes which have been happening in the pharmaceutical industry in the last few years.  Since Derek Lowe’s post on the topic last year (http://pipeline.corante.com/archives/2011/04/14/total_pharma_job_cuts.php), the number of casualties has continued to swell with job losses at AstraZeneca, the announced closure of Cellzome in Cambridge (UK), cuts at Merck Serono across Switzerland and likely cuts at Roche and Novartis in the USA.  So what does the future hold for the pharma industry?  Chris Swain has published a very interesting article on the topic recently (see http://www.samedanltd.com/magazine/11/issue/175/article/327).  There seems to be general consensus, however, that one growth area in computational chemistry in recent years has been in translational research.  I set out to explore exactly how in silico methods are helping this sector by speaking with some of the key players involved.


Due in no small part to the increasing drive towards open innovation in pharmaceutical R&D, many academic groups are now actively involved in drug discovery research. Computational chemistry and computational biology can be valuable tools in helping to convert discoveries and ideas into new therapies with real application to disease and illness. At MRCT, for example, Dr Andy Merritt’s computational chemistry group focus much of their effort on support of internal medicinal chemistry projects through use of structure-based design methods and SAR-rationalisation, but they also contribute significantly to design of the MRCT proprietary screening collection, the design of 3D fragments, and to supporting early stage academic structural biology projects.


Naturally, much early discovery work is centred around discovery and validation of potentially interesting new targets, which are often identified by cell-based screening approaches. Computational chemistry & biology approaches can be used to help identify a range of possible therapeutic applications of such targets, and can assess the likely druggability/ligandability of the targets. An example of this is the concept of senescence scoring from the CRUK-funded Beatson Laboratories (http://www.biomedcentral.com/1471-2164/11/532), whereby the concept of “senescence profiling” was explored to examine the levels of senescence signals in various tumours. The senescence scores were based on expression profiles of senescence biomarkers from phenotypic data. The correlation of senescence scores with growth inhibition in response to around 1500 compounds which had been screened for toxicity was explored using a binary decision tree algorithm which predicted likely activities of the 1500 compounds. From the study it appears that senescence scores may predict cellular therapeutic sensitivities, as differing subsets of the dataset (as defined by differing senescence signatures) appear to be enriched in different pharmacophoric elements predictive of gene family propensity.


Similar or related approaches can also be used to predict off-target effects of compounds, to explore drug repositioning opportunities or simply to assess target tractability for small molecule approaches.  Construct design is a key component of protein crystallography programmes, and computational methods can be of great benefit to design by identifying similar structures/sequences and using these to predict likely insert/deletion regions and to identify domain boundaries.


Of course, more traditional computational methods can also be applied to support medicinal chemistry programmes, including hit finding, scaffold hopping and lead optimisation projects.

Translational research is often funded by grant awarding bodies/charities, requiring researchers to devote considerable energy and time to develop grant applications.


At Cancer Research Technology computational chemistry is used primarily where protein structural information is available, but increasingly they have seen benefit in the application of computational approaches to assess new projects for target tractability.  Dr Tony Raynham, Head of Chemistry says “At CRT we’ve found that comp chem has accelerated research in several of our programmes, and we are increasingly using comp chem to help validate targets early in the discovery process.  We see in silico approaches as an integral part of our discovery efforts.”


So there you have it.  Comp chem is already proving pretty handy to have around in translational research, but if we can expand our armoury to devise even better ways to tackle the target tractability/druggability conundrum, there could be great opportunities for us in the future