Pharma companies developing new molecules spend millions to patent all chemicals close to their target chemical, to avoid that a competitor might be able to jump in with a similar product. This is before they even know whether it works. If variation is cheap why not run directly 5-50 versions of the chip and see which works best?