Virtual screening[1] is a computational technique where you take an experimentally resolved structure of a protein (PCNA in this case) and sequentially dock a large amount of different compounds to see which ones of them bind favorably to the target protein. It is worth mentioning that virtual screening is a very early step in a drug discovery pipeline. These hits need to be characterized and validated experimentally to see if there is an actual effect.

I am not well-versed in intellectual property (so please correct me if I’m wrong), but in this case Shkreli is using a database of commercially available compounds (ZINC) and a hit present in the screen could be patented. He said he won’t do it, and, since this could be considered prior art, nobody else can do a claim.

[1]: https://en.wikipedia.org/wiki/Virtual_screening