Caveat Emptor

From DISI
Jump to navigation Jump to search

We make software tools and databases available to the public, generally free of charge. When you use tools in the docking.org domain, you must do so at your own risk. This article describes some of those specific risks.

missing in smallworld

  • when using smallworld, be aware that only topologies that are indexed can be found. Thus in the current iteration of Enamine Private 3Q22 29B,

there are 29 billion individual smiles, but 3B of these are not indexed. That is to say that the anonymous graph ("molecular topology") has never been seen before, and thus there is no graph for this molecules in the map. Thus this molecule cannot be found using the search tool. It has not been indexed.

Thus there are 3 billion molecules that will never be found, UNTIL a new anonymous map is available and a new index in generated. New anonymous maps are produced by NextMoveSoftware twice a year, in time for the Spring and Fall ACS meetings. We then set to indexing all our collections using the new maps, which usually takes a few weeks, at most a month.

Of course, by this time, there may have been a new Enamine Private Space release, or a new WuXi release, and thus it will contain molecules with anonymous graphs that were not indexed. And the cycle repeats. Thus in each period, the coverage of chemical space improves, but it always lagging. We send unindexed topologies to NextMove, who duely map them. Up to H29, we expect this will slowly converge to complete coverage, asymptotically.


other problems