C.2 Computer Aided Molecular Design (CAMD)

  • Gani, Rafiqul (Project Manager)
  • Harper, Peter Mathias (Project Participant)

Project Details


Computer aided molecular design (CAMD) provides a means for determining molecules having a desirable set of physico-chemical properties. As the physico-chemical properties are directly or indirectly related to the structure of the molecule(s), methodologies for CAMD are typically based on "exploiting" these relationships.
A multi-level approach is currently under development. In this approach each level has it's own generation and testing steps. The results from each level "trickle down" to the next level. This ensures that the size of the combinatorial problem is continuously held at a level where it can be handled effectively without any so-called "combinatorial explosion". Furthermore, by having the most time consuming operations at the higher levels where the remaining candidates are the most promising, efficiency with respect to execution time is naturally obtained. Since the actual calculation/estimation of the properties is done using purely predictive methods, the accuracy, speed and applicability (with respect to level of detail needed in the molecule description) are aspects which have to be taken into considerations when selecting how and where to apply the various methods. Note that there may be several different available methods.
It is possible to design molecules with a significant level of detail (down to a 3-dimensional model on atomic level) without suffering from combinatorial explosion and without sacrifice of computational accuracy. Once a 3-dimensional model is obtained results can be exported directly into molecular modelling programs for further analysis. Current work is also investigating the developing of an interactive CAMD feature. The objective here is to design the molecule with desirable properties through interaction with the user.
Effective start/end date01/05/1997 → …


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.