If you don't mind writing some extra code, we've had good success with a
Monte Carlo implementation of a maximin diversity picker called BigPicker,
described in Blomberg et al, JCAMD, 23, 513-525 (2009). With this
implementation, you only need to keep the subset distance matrix in memory.
At each
Thanks Greg,
That's exactly what I needed.
I now have a question though, using modulo arithmetic means each environment
only sets one bit in the FP (checking the code this looks true). Is there a
reason why we only set a single bit with each environment? I mean we in the
greater sense, has
Hi all,
I made some small changes to the development version of the RDKit conda
recipes, so that building in a python3 environment will currently use the
code from the python3-support branch on GitHub. Building from sources may
this way just require the following commands:
$ git clone
Hi Greg,
Thanks! My Python is really rusty at the moment, so I am always unsure if
I am just not going through the steps in the most efficient manner, or if
the path that I am following is far from ideal. Granted I would prefer to
write this in Java, but I should probably move back towards C++
Hi Dave,
That's interesting, and I'll look into it. As I wrote Greg, I am an
aficionado of the Gobbi Lee method described here (since we are sharing
our favourite methods):
http://pubs.acs.org/doi/abs/10.1021/ci025554v
While the set I am looking at the moment contains only 26K molecules, I
5 matches
Mail list logo