printing the full ligand structure from deepfrag

[Jong Hui Hong]

Hi

I’m testing deepfrag for our drug design purpose.
It seems really useful for very fist step in drug design process

while I’m using it as a result it produce only added fragment like below

1 0.658789 *c1ccnc2c1C(=O)N(C)C2
2 0.649510 *c1ncnc2c1C(C)CC2O
3 0.647556 *c1cn(C)c(=O)c2[nH]ccc12
4 0.644948 *c1nc2cc[nH]c2c(=O)n1C
5 0.640279 *C1(O)C(=O)Nc2ccccc21

when I see the program option carefully

–full Print the full (fused) ligand structure.

which means that I can print full ligand structure or smiles

but when I use –full

I got following result

[] Error: couldn’t process mol.
[] Error: couldn’t process mol.
[*] Error: couldn’t process mol.
# Score SMILES
1 0.619665 <err>
2 0.612966 <err>
3 0.608838 <err>
.
.

what could be the problem with my running envirionment ?
Any kind of suggestion will be appreciated

Thanks

[Paulette Greenidge]

I am experiencing the same issue. The "full" option for multiple fragments is the main benefit of the command line implementation of DeepFrag.

[Paulette Greenidge]

I observed that the "full" option functions correctly if you remove a fragment (cname, rname) rather than just generate fragments to be added to a hydrogen atom, so there is a work around.

My new issue is that the smiles from the "full" output seems to have lost the bond type information. Is there a way to fix this without manual curation ?

Thanks

[Author]

Posts