try reparametrisation to optomize scaffold diversity

Open UnixJunkie opened this issue 10 years ago • 4 comments

use Paretto ranking, BM clustering and play with a. A change of the kernel function might also bee needed but this would have a huge impact on performance.

Oct 19 '15 09:10 UnixJunkie

related to https://github.com/UnixJunkie/ACPC/issues/6

Mar 08 '16 16:03 UnixJunkie

We should play with a (the dx parameter in fact), weights of each space (steric, hydrophobic and electrostatic). Let's not change the kernel function or it will become too slow. Issue https://github.com/UnixJunkie/ACPC/issues/6 is more important since it can improve perfs in several chemical spaces.

Mar 08 '16 16:03 UnixJunkie

the effect of the cut rot. bounds option should be studied too

Jun 14 '16 08:06 UnixJunkie

go in the p4 space for this. Also, go out of the molecule like one H2O shell away.

Aug 05 '16 08:08 UnixJunkie