The small clashes between the medicinal chemist and the computational chemist

Recently I took over a small project (or service ) from my friend, a blooming synthetic medicinal chemist.
After her advisor's seminar for his 10n-year amazing service ( [repetitively organic synthesis on certain Chinese medicinal compounds with known cytotoxicity]*20 projects ), I feel how trivial a computational chemist contribute in the drug discovery field compared to hands-on medicinal chemists. 


So the story of the manuscript is ...
She wants to demonstrate some fragments (Mol. Weight < 250-300) which have some cytotoxicity are tyrosine kinase inhibitors (e.g., c-Src is a famous anticancer target). The reason they suspect those fragments are kinase inhibitors based on a common two-ring system. 
[clash1: the degree of similarity is based on certain metric calculated by certain descriptors @@]


Then I suggested her to use the compounds as probes to see if she can find some kinase compounds in database. hmmm... one scaffold was found.


Since our lab doesn't develop any software related to in silico target identification, we tried the online target fishing server.
Eventually, I decided to dock the compounds against one cancer target. (shouldn't I do this at the beginning?) Everyone is happy at the pretty docking picture but I understand most of them are just pretty.


Again, I feel so useless at least they have some compounds with certain amount cytotoxicity against unknown targets. I have many results to show there is low probability of those compounds against kinase.


Conclusion: be positive despite the low probability