Cc1ccc(S(=O)(=O)NC(=O)Nc2ccccc2)c(C)c1
Cc1ccc(S(=O)(=O)NCc2ccccc2C#N)c(C)c1
O=C(Nc1ccccc1)NS(=O)(=O)c1c(F)cccc1F
Cc1ccc(S(=O)(=O)NC2CCN(C(=O)CCl)CC2)cc1
O=C(CCl)N1CCC(NC2(c3cccnc3)CCCCC2)CC1
Cc1ccc(C)c(S(=O)(=O)N2CCN(C(=O)CCl)CC2)c1
O=C(CCl)N1CCN(C2(c3ccccc3)CCCCC2)CC1
O=C(CCl)N1CCN(C2(c3cccnc3)CCCCC2)CC1
Cc1ncc(C2(N3CCN(C(=O)CCl)CC3)CCCCC2)s1
O=C(Nc1ccccc1)NS(=O)(=O)c1ccc(Cl)cc1
Cc1ccc(C)c(S(=O)(=O)NCc2ccccc2C#N)c1
Cc1cc2ncn(C3(c4cccnc4)CCCCC3)c2cc1C
OCc1ccncc1-c1cc(C2(c3ccccc3)CCCCC2)ccn1
Cc1ccnc2[nH]c(C3(C#CC4CCCCC4)CCCCC3)nc12
OCc1ccncc1-c1cc(C2(C#Cc3ccccn3)CCCCC2)ccn1
Cn1ccc2cccc(C3(C#Cc4ccccn4)CCCCC3)c21
1. Top 40 docking score compounds (John Chodera's docking - thanks!) and covalent fragments. 2. BRICS algo to break into synthetic frags - exclude leaf nodes 3. Enumerate using BRICS algo 4. Filter using rule of 5 5. Filter using regression model trained on John's docking scores to ones that have lowest estimated binding energy 6. Filter by comparing to list of submitted compounds