Review of Field Based Chemistry Europe 2013

Following on from the successful Field Based Chemistry North America last month, the scientific program of Field Based Chemistry Europe took place at Madingley Hall, Cambridge on 20th June 2013. A mixed program of speakers from academia and industrial organizations presented on a range of topics related to field based computational chemistry.

This blog gives a short overview of each speaker’s talk. Links will be added to the presentations as they become available.

In ‘Field points and structure based drug design’, Dr Gianni Chessari, Director at Astex Pharmaceuticals, presented the Astex approach to field points.

Dr Chessari described a case study of an X-linked inhibitor of an apoptosis protein in which field points were used to understand SAR and the ligand-protein interaction.

He concluded that a good understanding of the electrostatic interactions between ligands and proteins leads to efficient fragment optimisation and he noted that small molecule Cresset field points compare well with field points derived from ab initio methods.

Dr Mark Mackey, Cresset’s CSO, described ‘Upcoming features and future horizons’ for Cresset. He described the following new features:

• The new XED FF3 released in Cresset products last year,
• Activity cliffs and disparity analysis, available soon as a beta version,
• The recently released BlazeGPU that gives a 40 fold speed increase to Blaze.

He also described the work that Cresset is doing on:

• Calculating protein similarity
• Calculating off-target effects using 3D descriptors and features.

Dr Mackay also touched on topics that Cresset are looking at for the future:

• More work on analyzing protein fields.
• Extending disparity analysis to qualitative models for SAR.
• Extend off-target model work.
• Building field based ADMET models.

In ‘Application of Forge and Spark to develop new drug candidates from lead compounds derived from network pharmacology models’ Dr Ben Allen of e-Therapeutics described how network pharmacology can help to identify drugs when nothing is known about their binding. His presentation described how they use networks to identify the protein footprint of compounds. They use the Cresset tools Spark and Forge for structural ligand-based bioisostere identification.

Dr Tim Cheeseright presented Cresset’s new Activity Miner module, which will soon be available as a plug-in for Torch or Forge. Activity Miner measures 2D and 3D similarity and compares it to activity to create a disparity score. The results can be displayed using four different views:

• A colour coded disparity matrix that shows the disparity values for all pairs of molecules. This makes it easy to visually identify areas of high SAR.
• A disparity table that shows the pairs with the highest disparity value. The table also includes physical activity and structure.
• A cluster view that shows the molecules clustered by field similarity, but color coded by activity. This can be used to identify changes of activity within clusters.
• An activity view that includes a focus compound. Other compounds are displayed, ordered by their similarity to the focus compound. These compounds are colour coded by their disparity, making it easy to identify the field changes that have resulted in the change of activity.

‘New drugs from old compounds’ described Re-Pharm, a new offshoot of Cresset. Re-Pharm uses Cresset software to find candidate compounds for reprofiling in order to find a new clinical use for an existing treatment.

After lunch, Dr Daniel Kuhn, of Global Computational Chemistry, Merck Serono presented, ‘How to convince your medicinal chemist – from 3D QSAR to scaffold hopping’. He described a 3D QSAR project using Forge and a scaffold hopping project using Spark.

Spark provided solutions intuitive to medicinal chemists which have been picked up for synthesis. He pointed out that good communication is essential to successful work with medicinal chemists. In particular, one of the strengths of Spark is being able to copy and paste solutions straight into chemical drawing tools such as ISIS Draw.

Dr Ian Wall from GlaxoSmithKline presented ‘Computational Approaches in fragment based drug discovery’ in which he described several in-house computational tools they have developed to help medicinal chemists get access to the maximum amount of data to inform their compound selection.

He outlined their approach to fragment based drug design and emphasised the importance of optimising fragments to maximise their activity before growing them.

In ‘The trend to outsource computational chemistry’ Dr Martin Slater, Cresset’s Director of Consulting, outlined the increasing fragmentation of the pharmaceutical industry and the decrease in R&D spend that have contributed to the increase in outsourcing of all aspects of drug discovery, including computational chemistry. He described a recent consulting project for Mission Therapeutics to produce a small compound library. Further reading: The trend to outsource computational chemistry; Cresset services.

María José Ojeda Montes of Universitat Rovira i Virgili presented ‘Conformational sampling tools’. She described a methodology they have been developing to explore the conformational space around leads. They found that the acceptable reliability for finding the bioactive pose decreased as the number of rotatable bonds increased above seven.

They hope in future to generate a database of pre-generated conformations of small molecules with less than seven rotatable bonds, which includes 70% of the small molecules of the Zinc database.

Dr Peter Hunt, Senior Research Investigator at Novartis, presented ‘Crystals, Co-formers and Cresset – adventures on the high C’s’. Co-crystals are a crystalline material consisting of 2 or more electrically neutral molecular species which are solids at ambient temperature. He gave the example of carbamazepine, which is very prone to forming co-crystals. However, the solvate furfural is toxic. The challenge is to take the toxic solute and to replace it with an acceptable co-former.

He noted that Cresset software generates ideas as to what might work and what might work better. They believe that co-crystal screening will grow in importance and the ability to rank acceptable co-formers will become a need.

The final presentation of the day was by Dr Andy Vinter, the founder and Chairman of Cresset. His talk, entitled ‘Kinetics versus thermodynamics – insight into protein function’ described the difference between the active β2AR/”G” vs the inactive β2AR. He described how studying the difference between the very similar active agonist and the inactive antagonist will reveal a lot of information and will be a useful test of the XED force field.

Dr Tim Cheeseright of Cresset closed the meeting by thanking the speakers.