Real-world complications that we can help you overcome
The following are all questions that Tessella’s consultants have addressed in work on drug discovery sites of global pharmcos:
- What should the predictive method cutoff be? How does this depend on the screening strategy you are using?
- How should ‘cost’ be estimated – time taken; variable costs, use of capacity, lost opportunities, or a mix of these?
- What financial value should you place on time saving?
- How does the value to be placed on a single compound depend on the variety and redundancy of options available at any stage of discovery?
- Should we think about prediction reliability and filtering impacts compound-by-compound, or in terms of quality of decisions on entire chemical series?
- How can we support people to use the admittedly imperfect information that they have on costs, potential value, and sources of risk, at early stages of R&D?
- What if the errors in methods are not independent, or the compound measures interact (e.g. solubility and bioavailability)?
- Where planning a cascade of tests (typically in final stages of candidate choice), how should the team best sequence work to assess the different failure causes and success factors – on the basis of the test cost; test time; or also considering relative risks of failure, and the predictive performance of screening methods?
- Given that many selection decisions are multivariate, made on the basis of multiple parallel measurements, what is the best way of choosing the measures and setting progression criteria?
We also have informal alliances with partners who provide specialist chemoinformatics and bioinformatics and statistical expertise, or we can work as business advisors and technical consultants with your own specialists. This widens the services we can offer to include innovative thinking about chemical diversity and polypharmacology:
- As medicinal chemistry is a question not only of filtering, but also exploration into new chemical space, and is also iterative, how far can these principles and visualisation approaches help choose the right new chemistry?
- How can we model and enhance lead optimisation performance?
- When should we subset a library for screening?
- Given that many drugs act through multiple mechanisms, what should be our approach to filtering using multi-target interaction profiles?
- Given that the organisation may be blind to false alerts, if ‘failed’ compounds are never followed up, what should be the balance between ‘exploration’ and ‘exploitation’?
