Discussion Optimizing drug design

1
DiscussionOptimizing drug design

• Summary of color-cards of first day

2
Overzicht

• Where are we now?
• What does work? What does not work?
• Where do we want to be?
• What are the approaches?

3
Where are we now?

• We have 'pretty' good models (one log unit
  accuracy)
• Promising approaches for optimizing to those
  models experimental validation is in its
  beginnings as idea-generators they are used
• Multiobjective Optimization strategies are
  available from other disciplines and can be used
• Data Mining is 'pretty' advanced
• Public Databases are still to be extended
• Computer power is available but not fully
  exploited

4
What is desirable for the future? and how can
it be achieved?

• The acceptance of the tools by the chemists
  should increase
• A good communication language will help
• A human centric decision process (the chemists
  wants to participate in the decision and
  discovery process)
• Synthetic feasibility approaches
• Tools should explain rationale
• Good accuracy of the models is required
• More Data, better descriptors could quantum based
• Display model accuracy in decision process
• A good understanding of what the objectives
• What are models to predict human outcomes (2020
  virtual human will help)
• Objectives based on systems biology insights
• More financial funding from industry and public
  research foundations

5
What is desirable for the future? and how can
it be achieved?

• Exploration of a larger and more relevant part of
  chemical space
• Concise definition of chemical space maybe
  pharmacophore oriented toxicophores and
  conformation?
• Combine global and local searches' uncertainty
  management (confidence bounds)
• More constraint checks use metamodels for rapid
  search
• Dynamic knowledge use closed loop learning
• Use interactive learning