ZINC15:Optimizing chemical search

From DISI
Jump to navigation Jump to search

Chemical search that uses complex SMARTS patterns, returns many molecules, or is otherwise complex, may require a long time to run.

Here are tips to make long-running queries run faster:

  • use sort=no to turn off sorting.
  • use parallize=no (or yes) to explore whether parallelism helps
  • use duplicates=yes - can run faster if you remove duplicates yourself
  • specify additional constraints, for instance, if you are only interested in fda drugs, or in-vitro compounds, the constraints will make the query faster
  • specify range of ZINC IDs, thus zinc_id-between=100000,200000 and run in several tranches. Then you put the result together when complete.