Spring 2013




UK-QSAR Spring 2013

The UK-QSAR and ChemoInformatics Group

Welcome to the Spring 2013 UK-QSAR Newsletter!

In this edition we have a focus on open innovation since this new paradigm in drug discovery is something which could have some impact on many of us, and we are very grateful to Martino Picardo of the Stevenage Biocatalyst Centre and to Geoff Lawton of INMedD for their pertinent and stimulating contributions below.  Please note that any opinions expressed by individual contributors are their own personal viewpoints and do not reflect any position adopted by the UK QSAR & Chemoinformatics Society.

Our ADMET-themed Spring Meeting is also coming up very soon.  Details are below.

As ever, please send any feedback or suggestions you have for future newsletters to Susan Boyd at  newsletter@ukqsar.org.   This is your newsletter – please help us to make it how you’d like it!


Spring Meeting 2013

The Spring Meeting will be held at Unilever, Colworth Park, Bedford, on 23rd April 2013.   The theme of the meeting will be predictive ADMET and the programme features talks on hERG & p450 modelling amongst other topics.  Registration is now open so please sign up to secure your place as numbers are limited and meetings are usually oversubscribed.

At previous meetings we have noted a number of “no-shows” on the day of the event.  Whilst unforeseen emergencies sometimes do crop up, if you do register but later find that you can no longer attend please let us know so that we can open up your place to someone on the waiting list who would otherwise have been disappointed.

If you plan to travel by rail to the event, please note that Bedford station is much closer to the venue than is Milton Keynes.

The Provisional Agenda is now at http://www.ukqsar.org/2012/10/04/spring-meeting-2013/.


So Just What is Open Innovation?

Geoff Lawton, a Founder of INMedD, a drug discovery social enterprise

I confess I dislike the term ‘open innovation’ and usually try to avoid its usage. Part of it, of course, is the typical ‘chemist’ scepticism for language hi-jacked by others to form buzz words. On reflection, if chemists had been more embracing when ‘biochemistry’ and ‘molecular biology’ were coined, we may not have needed to invent ‘chemical biology’ in order to try and regain lost ground in the public perception.

Another issue is the wide range of situations in which the term ‘open innovation’ is used. These vary from the truly open Structural Genomics Consortium whose aim (http://www.thesgc.org/openaccess/about/details) is to put all of its findings in the public domain, through to a pharma company that wants to increase its access to external innovations with the intention to exploit them to generate returns on their shareholders’ investment.  Don’t get me wrong. I am a big fan of collaborations in new medicines discovery. In fact this is likely to be the best route to success in today’s troubled world. Anything that facilitates these collaborations (see for example Martino’s accompanying piece and the work of the Stevenage Bioscience Catalyst) is a good thing.

Open implies trust and transparency and trust is often the bigger problem.  Trust in the competency of one’s partners is critical. For the collaboration to be truly open there also has to be trust of partners’ motives. This is best achieved when motives are aligned and this implies that when the inevitable conflicts arise in the partnership they can be solved to mutual satisfaction. Very often today ‘open innovation’ is applied to partnerships which rely on the legal agreement between the partners to resolve conflicts. This tends to satisfy only the lawyers.

The rapidly developing drug discovery social enterprises around the world (http://drugdevelopmentalliance.com/index.php ) can provide a focus for more trusting partnerships. These organisations are under the control of social stakeholders and may give confidence in a more truly “open” motive and allow ‘honest-broker’ behaviour to provide lubrication for the partnerships. They may also be able to effectively manage the elusive boundaries between pre-competitive and competitive technology.


Drug discovery CROs are also a key component of the collaboration landscape and are currently providing much of the employment prospects in the UK.   Another important development is the IMI Lead Factory (http://www.imi.europa.eu/content/european-lead-factory). I look forward with great anticipation to this blossoming and becoming an important focus for collaborations.


Finally I would like to disagree with Andrew Leach’s response to the question  “How can we inspire new students to take up the discipline?” in the previous newsletter.  His answer: ‘I fear that I am not sure that it is currently fair on students to try and do this’.  There is indeed a current hiatus in the practice of drug discovery as pharmacos come to terms with their unsuccessful strategies of the past decade and act to reduce costs by switching off research, and thereby ‘eating their seedcorn’.  Unless this changes, they will die!


But the burgeoning collaboration landscape provides good reason for optimism.   Outside the traditional view of pharmacos, the underlying situation is very positive for employment (including computational chemists).  The market for medicines continues to grow.  There will be 1 billion more people >65 years by 2050 and many diseases increase in incidence after age 65.  Increased information flow drives global access to new medicines. Governments globally invest $200bio pa in academic bioscience (in addition to pharma and biotech spend) and this generates a rich output of ideas and opportunities. Increased understanding of disease genome and molecular pathology plus vastly improved drug discovery processes demand exploitation for the benefit of patients. There is every reason for bright young folk to pursue science studies and to apply their intellect and knowledge to solving the needs of future patients. This societal need will force the invention of sustainable business structures and these will provide employment.  What could be more attractive as a career prospect?

Introducing Discover Assist; an integral part of SBC Services

Martino Picardo, CEO, Stevenage BioCatalyst Centre


Since the opening of the Stevenage Bioscience Catalyst (SBC; www.stevenagecatalyst.com ) in February 2012 the number of big developmental initiatives for the growing biomedical campus has increased.  The open innovation incubator, co-located on the GSK R&D site in Stevenage, held its inaugural open innovation event on November 15th   2012 The Open Innovation Summit hosted some 160 delegates across industry, academia and the investment community. Discover Assist played an important part in the success of the event. More recently a number of different types of start-ups have taken tenancy, including Aptiv Solutions, Magdi, Cambridge University and Labstract (others listed at http://www.stevenagecatalyst.com/about/our_tenants/ )

The SBC mantra is to foster collaboration (partly through open innovation), to add value to early stage “small molecule” and “biologicals” type development projects. We can assist with moving through the project from research to development to market stage This is achieved by providing access to key technical and commercial expertise at the right time and in a cost effective manner. One of the ways SBC can do this is through an advisory panel of experts, called Discover Assist. The group comprising Prof. Bill Dawson, Prof. Stan Roberts , Duncan Judd and Martino Picardo, are available to assist (on a first hour free basis) with projects at the interface between biology and chemistry. In addition projects requiring advice on funding as well as gaining access to key Pharma and Biotech expertise, are also on the horizon for Discover Assist services.  Discover assist team members will be conducting road-shows in a (University) location near you and will look forward to meeting you there. For further details, please contact martino.picardo@stevenagecatalyst.com.



News – Committee Changes

We are delighted to welcome Nora Aptula of Unilever and David Livermore of Takeda to the organising committee.  We would also like to extend our thanks to David and to Takeda Cambridge for their very kind hospitality at the Autumn 2012 meeting in Cambridge and to congratulate them on their efficient organisation of the event.  Nora will be the Unilever contact for the upcoming Spring 2013 meeting and we would like to thank her for the enthusiasm and energy she has already invested in this role.


Applications Developer, Royal Society of Chemistry, Cambridge, UK


ChemSpider Sales Executive, Royal Society of Chemistry, Cambridge, UK


Informatics Manager, RedX Pharma, Wakefield, UK


Bioinformatician, Target Discovery, Biofocus, Leiden, The Netherlands


Application Scientist, Chemical Computing Group, Cambridge, UK or Cologne, Germany


Research Scientist for Methods/Software Development, Roche, Basel, Switzerland


Application Scientist, Dotmatics Limited, Bishops Stortford, UK


Java Developer, Dotmatics Limited, Bishops Stortford, UK


Porter’s Papers – A review of recent literature from Rod Porter

(Full text of Rod’s bi-monthly newsletters, which include med chem literature reviews in addition to the more general/comp chem papers included here, can be found on his website at http://rodporterconsultancy.com/newsletter/).


Virtual ADMET modelling

A couple of papers discuss the calculated and physicochemical properties relating to promiscuity. First up1 is a paper looking at 3D descriptors and promiscuity as measured by successful clinical progression of compounds. The authors suggest that shape-based 3D descriptors such as the radius of gyration and shadow indices discriminate off-target promiscuity of a set nicotinic ligands better than do fraction of sp3 carbon (Fsp3) and the number of stereo centres which others2  have proposed as key indicators of clinical success. The authors also extended this analysis to show these values were predictors of progress through preclinical and Ph1 development and were also found to be good indicators of solubility. Shadow Index or Fsp3 critieria were generally met (84%) by marketed drugs but were not met by withdrawn or discontinued products. It appears that spherical compounds with few aromatic rings have a better chance of making drugs – does this help explain some of the upsides of macrocycles? I must admit I am not quite convinced about the robust of the arguments over selectivity here.

As a bit of a contrast Keseru3 emphasises the importance of LogP and basicity in determining selectivity which rather reinforces my feelings against the routine introduction of basic centres to make a compound soluble – there are other ways of doing this – see above – apart from anything else.

A warning comes from Kenny and Montanari4  that one does have to be careful with data analysis as some strategies can exaggerate the significance of relationships –they introduce a term “correlation inflation”. They particularly highlight the risk of analysing binned data and averaging groups of data points before analysis

With respect to promiscuity there is a lot to be said for some targeted screening to look for particular nasties such as hERG or 5-HT2B receptor affinity an approach extolled by a group of authors5 sharing in vitro screening data from four major pharma.

Following the in silico or physicochemical parameters and attrition theme is a paper from Wenlock6 reviewing the relevance to ADMET behaviour of a range of physicochemical properties including, perhaps what may be considered the usual suspects of, ionization lipophilicity, H-bonding and solubility amongst others – both measured and predicted.

  1. D. C. Kombo et al, J. Chem. Inf. Model., Article ASAP DOI: 10.1021/ci300445e Publication Date (Web): January 18, 2013
  2. F. Lovering et al J. Med. Chem., 2009, 52, 6752
  3. A Tarcsay and G. M. Keserű J. Med. Chem., Article ASAP DOI: 10.1021/jm301514n Publication Date (Web): January 28, 2013
  4. P. W. Kenny, C. A. Montanari J. of Computer-Aided Mol. Design Published online 10 January 2013
  5. J. Bowes et al, Nature Reviews Drug Discovery, 2012, 11, 909-922 doi:10.1038/nrd3845
  6. M. C. Wenlock and P. Barton Mol. Pharmaceutics, Article ASAP DOI: 10.1021/mp300537k Publication Date (Web): January 24, 2013


Predicting PK & Safety

A valuable review of the application of in silico, in vitro and in vivo PK for the prediction of human PK from AZ is just out1. It appears to be targeted to DMPK scientists but it seems pretty handy as a survey of the field to me as a medicinal chemist.


Another hazard for screening is concern over metal impurities in samples. An article from Roche2 discusses the problem of zinc contamination that they have had problems with. The recommended solution by the authors is to run a counter screen in the presence of metal chelator PTEN.


1.  K. H. Grime, P. Barton and D. F. McGinnity Mol. Pharmaceutics, Article ASAP DOI: 10.1021/mp300476z Publication Date (Web): January 29, 2013

2. J. C. Hermann, et al ACS Med. Chem. Lett., Article ASAP, DOI:10.1021/ml3003296 Publication Date (Web): December 20, 2012


Halogen bonding

A comprehensive review on halogen bonding in medicinal chemistry makes a convincing case that we should forget regarding halogens as simply lipophilic blobs. The key point is the recognition of the anisotropic distribution of electrons in ArHal – excepting fluorine with its extreme electronegativity – in which a positively charged surface, the sigma hole, on the z-axis is available for interaction with a Lewis base. Covered in the review are a consideration of the strength of interactions and their manipulation by additional electron withdrawing substituents on an aromatic ring; interaction geometries and energy boundaries with a comparison of theory and practise – based on crystallography studies; a survey of successful application of halogen bonding and finally a defense of halogen. Halogens have often been regarded as a liability – too lipophilic and adding a lot of molecular weight for very little specific gain in target interaction – “its just a lipophilic effect”. The authors argue that affinity gains can be substantial claiming upto 100 fold for ArH – ArI based on an overt halogen bond and that while they are dense their actual contribution to molecular size is much below that of their molecular weight

R Wilcken et al J. Med. Chem., Article ASAP Publication Date (Web): January 03, 2013


Peptides = drugs?

Peptides as drugs can be an emotive subject for drug discovery scientists. A recent review1  highlights recent findings from a variety of fields that are converging on a new understanding of how conformation controls peptide bioactivity and bioavailability. The authors review some examples of peptides with significant oral bioactivity including, amongst others, α-amanitin a relatively rigid bicyclic peptide and the ubiquitous, more flexible, monocycle cyclosporine. Reviewing the case histories the authors conclude that cyclization is the key first step the trick of course being to find the relevant conformation. Refining the structure then follows with natural and unnatural amino macid substitutions and alkylations to optimise target affinity and physcicochemcial properties. However rational approachesa at this stage are limited. One thing that struck me was that often at least bicyclic structures are required – a bit more of a challenge synthetically.

A disappointment is that the authors do quote data on cell permeability in some instances with dye labelled analogues of a peptide and propose this is evidence of permeability of the peptide.

One of the big hopes for peptide drug discovery has been stapled peptides as alpha helix mimetics2. However recent data3 looking at stapled stabilized BimBH3 peptides suggest that this strategy is far from a universal panacea with loss of both target and cell activity on stapling acyclic peptides.

Snake/conotoxin peptides have been recognised for a long time as having some spectacular biological activities therefore a paper on mambalgins4, analgesic three finger acid sensing ion channels inhibitors isolated from Black Mamba toxin, is timely. Note these mambalgins are highly constrained with four disulfide bridges.


1. J. E. Bock et al ACS Chem. Biol. 2013 DOI: 10.1021/cb300515u

2.G. L. Verdine and G. J. Hilinski Methods Enzymol. 2012;503:3-33. doi: 10.1016/B978-0-12-396962-0.00001-X.

3. T Okamoto et al ACS Chem Biol., Article ASAP DOI: 10.1021/cb3005403 Publication Date (Web): November 14, 2012

4. S. Diochot et al Nature 2012, 490, 552 and A Fleming Nature Reviews Drug Discovery 2012, 11, 906


Interactive chemical space visualisation

A mapplet for the interactive visualisation of chemical space defined by principal component planes of 42 dimensional Molecular Quantum Numbers has just been introduced. Databases used for this purpoose are Drugbank, GDB-13 (near 1 billion molecules), GDB-11, Pubchem and Chembl. The graphic below illustrates the whole Chembl database, any molecule can be identified along with nearest neighbours. Zooming in allows selection of each molecule (not in GDB-13) otherwise each pixel may represent multiple molecules for which an average structure is displayed in the average molecule box. More information is available on the Reymond group website and in the reference.

1.  M. Awale et al J. Chem. Inf. Model., Article ASAP DOI: 10.1021/ci300513m Publication Date (Web): January 22, 2013 Copyright © 2013