Choosing a subset
What subset of ZINC should I dock? Here is a quick guide.
First, what subsets are available?
1. Vendor Subsets (for immediate delivery) Use these if you have a deal with a vendor that makes pricing attractive. For instance, even if you don't pre-order a screening library, many vendors will offer special pricing for volume orders.
2. Virtual Vendor subsets (typically 8-10 weeks delivery). Use these if rapid turnaround is not essential. Advantages: gives you access to over twice as many compounds as the "immediate delivery" subsets.
3. Property-filtered subsets. We particularly favor subset #2 (fragment like) and #1 (lead like) for many reasons: solubility, sampling of chemical space / Hann complexity, fewer docking artifacts.
4. User created subsets. These are subsets created by users, from the search results page, using the "create subsets" button. Some are also created as byproducts of DOCK Blaster runs.
5. User-uploaded subsets. These are compounds uploaded by users. If you uploaded molecules you want to dock, this is where to go.
6. By annotation. You may want to dock compounds that have been pre-filtered for their likelihood of binding to certain targets, based on literature reports and patents. This option is not currently supported.
7. By 2D prediction. You may want to dock compounds that resemble actives, based on their similarity to ensembles of actives (SEA). This option is not currently supported.
What should I dock?
Relatively small binding site: start with "clean-fragments" (by property, #12). Larger binding site, start with "clean-leads" (by property #11). There are of course exceptions to these rules, but these are sensible starting points.
-- John Irwin