Submission Details
Molecule(s):
N#C[C@@H]1[C@H](CO[P@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)O[C@H](n2ccc3cc(-c4ccccc4)ccc32)C1(F)F
N#C[C@@H]1[C@H](O)[C@@H](n2ccc3cc(-c4ccccc4)ccc32)O[C@H]1CO[P@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O
C#C[C@]1(CO[P@@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)O[C@@H](c2cc(F)cc(C#N)n2)[C@](F)(C#C)[C@@H]1C#N
C[C@@]1(F)/C(=C2/CC(c3nn[nH]n3)=NC2=O)O[C@@H](CO[P@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)[C@H]1C#N
C#C[C@]1(CO[P@@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)O[C@H](c2ccc3ccccc3c2)[C@@](F)(C#N)[C@@H]1C#N
C#C[C@]1(CO[P@@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)O[C@H](c2cc3ccccc3[nH]2)[C@@](F)(C#N)[C@@H]1C#N
N#C[C@H]1[C@H](O)[C@H](CO[P@](=O)(O)O[P@](=O)(O)OP(=O)(O)O)O[C@@H]1n1ccc2cc(-c3ccccc3)ccc21
N#C[C@@H]1[C@@H](CO[P@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)O[C@H](n2ccc3cc(-c4ccccc4)ccc32)[C@H]1C#N
C#C[C@]1(F)[C@@H](n2ccc3cc(-c4ccccc4)ccc32)O[C@@](C)(CO[P@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)[C@H]1c1n[nH]c([C@@H]2[C@H]3CCO[C@@H]23)n1
C#C[C@]1(F)[C@@H](n2ccc3cc(-c4ccccc4)ccc32)O[C@@](CO[P@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)(c2cc3c(N)cccc3o2)[C@H]1C#N
C[C@]1(O)[C@@H](n2ccc3cc(-c4ccccc4)ccc32)O[C@@](CO[P@@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)(c2cc3ccc(F)c(O)c3[nH]2)[C@H]1C#N
Cc1cc([C@H]2O[C@@](CO[P@@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)(c3cc4c5c(ccc4o3)CNCC5)[C@@H](O)[C@]2(F)C#N)cc2ccccc12
N#C[C@@H]1[C@@H](CO[P@](=O)(O)O[P@](=O)(O)OP(=O)(O)O)O[C@H](C2=CC(=O)NC2=O)[C@H]1c1n[nH]c(Cc2coc(-c3ccccc3)n2)n1
C#C[C@]1(CO[P@@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)O[C@H](n2ccc3cc(-c4ccccc4)ccc32)C(F)(F)[C@@H]1C#N
CN(C)C(=O)c1ccc2cc([C@]3(CO[P@@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)O[C@H](c4cc5ccccc5[nH]4)[C@@](F)(C#N)[C@@H]3C#N)[nH]c2c1
C[C@]1(F)[C@H](n2ccc3cc(-c4ccccc4)ccc32)O[C@](C#N)(CO[P@@](=O)(O)O[P@@](=O)(O)OP(=O)(O)O)[C@H]1c1n[nH]c([C@@H]2[C@H]3CCC[C@H]32)n1
Design Rationale:
To begin with, an extensive literature search was done, to find out all structures that had been previously reported to act like nucleotides. This search was further made more extensive, by separately searching for sugar and nucleobase analogues. From this, we came to a list of about 250 ligand, which had known activity as either a nucleotide analogue, or the potential to act as a part of one. This was then used as the input for the autogrow4 package, which can "grow" molecules by selecting for properties which lead to a lower docking score, which was done with the crystallized structure of the SARS Cov2 RdRP. This was run for a few generations, with the conditions that the structure contain phosphate groups and has a MWT<760. After a few generations of extensive searching of the molecular space, we selected the top 100 molecules from about a list of 20,000 nucleotide analogues below the MW of 760. The docking studies (part of the autogrow4 runs) were done via QuickVina2, and the top 1500 molecules were redocked using AutoDock Vina, for confirmation, and for further runs in autogrow4, with the specific filters. This was then the final output. I have here submitted the top few molecules among the selected 100, with those preferred which have a docking score less than -9, or a specific score less than -0.23 with the RdRP of the SARS-Cov2 virus.
Other Notes:
The aspect of innovation here is the way of effectively searching the chemical space, specifically for nucleotide analogues. Given that it was checked across various docking algorithms, it further reinforces the accuracy of the results, with tautomers and conformers showing higher energies specifically selected for. This work is a part of the Drug Discovery Hackathon, 2020, phase 1, and all the computational studies were performed in the tool room provided by the hosts of the competition.