After sampling, I am getting same molecule

[snoze]

Hi There.

Thanks for a wonderful program. It works but there is a problem. After sampling I am generating a file sampling001.csv.

SMILES,Input_SMILES,Tanimoto,NLL
CC(C)Oc1cccc(-n2nnc3c(=O)[nH]c(CC#N)nc32)c1,CCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,1.0,1.4622701406478882
Cc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,CCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,1.0,1.553662896156311
CCCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,CCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,1.0,1.655609369277954
CC(C)Oc1cccc(-n2nnc3c(=O)[nH]c(C(C)C)nc32)c1,CCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,1.0,1.7949905395507812
CC(C)Oc1cccc(-n2nnc3c(=O)[nH]c(C(F)(F)F)nc32)c1,CCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,1.0,1.8882602453231812
CCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,CCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,1.0,3.1838881969451904
COc1cccc(-n2nnc3c(=O)[nH]c(C)nc32)c1,CCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,1.0,6.213459014892578
CC[C@H]1O[C@@H](n2nnc3c(=O)[nH]c(C)nc32)[C@H](O)[C@@H]1O,CCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,0.5068493150684932,7.01755428314209
CC(C)Oc1cccc(-n2nnc3c(=O)[nH]c(C(Cl)(Cl)Cl)nc32)c1,CCc1nc2c(nnn2-c2cccc(OC(C)C)c2)c(=O)[nH]1,0.76,7.8528733253479

The SMILES generated are same as Input_SMILES in my original file. Any idea why? For transfer learning I am using following code

# Define the type of similarity and its parameters
pairs.type = "tanimoto"
pairs.upper_threshold = 0.9
pairs.lower_threshold = 0.7
pairs.min_cardinality = 1
pairs.max_cardinality = 199

Thank you

[halx]

Hi,

many thanks for your interest in REINVENT and welcome to the community!

as fas as I can see only the 6th molecule is identical. It is not unusual to sample duplicates. All other sampled compounds are different from the (single) input SMILES. There seems to be a bug though in reporting the Tanimoto similarity.

Many thanks for reporting this,
Hannes.

[snoze]

Thank you Halx. Highly appreciate your feedback.