Find-a-Drug: Distributed computing

Next month, we are planning to resurrect the distributed In Silico screening using our THINK software that first become well known in the Oxford University Screen Saver Life Saver cancer research project.

A couple of decades later THINK 2.01 is somewhat more sophisticated. The de novo derivative generator gives many more interesting analogues as a result of an extensive range of ring exchange and ring substitution transformations. The conformational generator now includes functionality to permute unknown chiral centres and ring conformations.

We envisage seeding the initial jobs with molecules which we screened last year (and submitted to PostEra Moonshot) which had known binding fragments derived from combinatorial chemistry libraries and searching commercial available compound sets. We will also look at repurposing existing drugs and their derivatives. At this stage we wont be using the de novo functionality to grow molecules from small fragments just to create analogues. Last year we used a variety of targets for covalent and non-covalent binding.

Our approach also provides in silico P450 toxicity screening and selectivity screening against human proteases. We anticipate the AI approaches used to select molecules for analogues, P450 and selectivity screening to become more sophisticated as data is generated.

Further information will be available at Cure Covid