Home > ChEMBL 90

Dopamine D4 receptor ─ ChEMBL 90

Dataset sourcesHEIKAMP (2011) RINIKER (2013) Benchmarking-Platform
Compounds10000× 'decoy', 100× 'active'
ClassifierSupport Vector Machine (c:10, RBF-Kernel γ:0.01)
FeaturesFiltered ECFP4 fragments 
The service uses filtered (instead of folded) Extended-Connectivity Fingerprint (ECFP) fragments as features for the prediction algorithm. more..
Num features8192

 

Recent predictions

CompoundPrediction
Each prediction model provides a probability estimate for the prediction. The probability value indicates how confident the classifier is. more..
App-Domain
QSAR model predictions should not be trusted if the query compound is outside of the applicability domain of the model (i.e., if the query compound is dissimilar to the training dataset compounds). more..
COC1=CC2=C(C(=O)C(=C(O2)C3=CC=C(OC)C(=C3)OC)OC)C(=C1)OC
  decoy (100%)
CCOC(=O)CCC1=CC(=C(C=C1)O)OC
  decoy (99.98%)
CC(C)OC1=CC(C2CCNCC2)=C(C)C=C1NC1=NC=C(Cl)C(NC2=CC=CC=C2S(=O)(=O)C(C)C)=N1
  decoy (100%)
-
Cc1c(cn(n1)c2c(cccn2)CO)CN3CCC4(CC3)c5c(cc(s5)Cl)C(CO4)(F)F
  decoy (87.35%)
-
c1ccc(cc1)C(c2ccccc2)NC(=O)CSc3nnc(n3c4ccccc4)COc5cccc(c5)Br
  decoy (100%)
CN(CC1)CCN1C2=CC=C3C(N=C(N3)C4=CC=C5C(N=C(N5)C6=CC=C(C(F)(F)F)C=C6)=C4)=C2
  active (98.26%)
?
C1(C=CC2)=CC=CC3=CC=CC2=C13
  decoy (98.71%)
-
Oc1ccc(cc1)C(C)(CCC(=O)O)c2ccc(O)cc2
  decoy (100%)
CC1CCC2C(C(=O)OC3C24C1CCC(O3)(OO4)C)C
  decoy (100%)
C1=CC(=C(C=C1Cl)O)OC2=C(C=C(C=C2)Cl)Cl
  decoy (81.5%)
-