lazar 0.0.7 → 0.0.9

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (61) hide show
  1. checksums.yaml +4 -4
  2. data/.gitignore +3 -0
  3. data/README.md +2 -1
  4. data/VERSION +1 -1
  5. data/ext/lazar/extconf.rb +15 -76
  6. data/ext/lazar/rinstall.R +9 -0
  7. data/lazar.gemspec +7 -7
  8. data/lib/classification.rb +5 -78
  9. data/lib/compound.rb +201 -44
  10. data/lib/crossvalidation.rb +224 -121
  11. data/lib/dataset.rb +83 -93
  12. data/lib/error.rb +1 -1
  13. data/lib/experiment.rb +99 -0
  14. data/lib/feature.rb +2 -54
  15. data/lib/lazar.rb +47 -34
  16. data/lib/leave-one-out-validation.rb +205 -0
  17. data/lib/model.rb +131 -76
  18. data/lib/opentox.rb +2 -2
  19. data/lib/overwrite.rb +37 -0
  20. data/lib/physchem.rb +133 -0
  21. data/lib/regression.rb +117 -189
  22. data/lib/rest-client-wrapper.rb +4 -5
  23. data/lib/unique_descriptors.rb +6 -7
  24. data/lib/validation.rb +63 -69
  25. data/test/all.rb +2 -2
  26. data/test/classification.rb +41 -0
  27. data/test/compound.rb +116 -7
  28. data/test/data/LOAEL_log_mg_corrected_smiles.csv +567 -567
  29. data/test/data/LOAEL_log_mmol_corrected_smiles.csv +566 -566
  30. data/test/data/LOAEL_mmol_corrected_smiles.csv +568 -0
  31. data/test/data/batch_prediction.csv +25 -0
  32. data/test/data/batch_prediction_inchi_small.csv +4 -0
  33. data/test/data/batch_prediction_smiles_small.csv +4 -0
  34. data/test/data/hamster_carcinogenicity.json +3 -0
  35. data/test/data/loael.csv +568 -0
  36. data/test/dataset-long.rb +5 -8
  37. data/test/dataset.rb +31 -11
  38. data/test/default_environment.rb +11 -0
  39. data/test/descriptor.rb +26 -41
  40. data/test/error.rb +1 -3
  41. data/test/experiment.rb +301 -0
  42. data/test/feature.rb +22 -10
  43. data/test/lazar-long.rb +43 -23
  44. data/test/lazar-physchem-short.rb +19 -16
  45. data/test/prediction_models.rb +20 -0
  46. data/test/regression.rb +43 -0
  47. data/test/setup.rb +3 -1
  48. data/test/test_environment.rb +10 -0
  49. data/test/validation.rb +92 -26
  50. metadata +64 -38
  51. data/lib/SMARTS_InteLigand.txt +0 -983
  52. data/lib/bbrc.rb +0 -165
  53. data/lib/descriptor.rb +0 -247
  54. data/lib/neighbor.rb +0 -25
  55. data/lib/similarity.rb +0 -58
  56. data/mongoid.yml +0 -8
  57. data/test/descriptor-long.rb +0 -26
  58. data/test/fminer-long.rb +0 -38
  59. data/test/fminer.rb +0 -52
  60. data/test/lazar-fminer.rb +0 -50
  61. data/test/lazar-regression.rb +0 -27
data/test/data/LOAEL_mmol_corrected_smiles.csv ADDED
@@ -0,0 +1,568 @@
1
+ SMILES,LOAEL mmol/kg_bw/day
2
+ C1=C(C(=CC(=C1NN=C3C2=C(C=C([S]([O-])(=O)=O)C=C2)C=CC3=O)OC)[S]([O-])(=O)=O)C.[Na+].[Na+],0.007531928608839286
3
+ O1C(=O)C(O)=C(O)C1C(O)CO,0.017323010613197083
4
+ C1(C)=C(C=CC(C)=CC=CC(C)=CC=CC=C(C)C=CC=C(C)C(=O)OC)C(C)(C)CCC1,0.0011194097182405476
5
+ c(cccc1)(c1)C(C)C,0.003843863272285792
6
+ O=C(OCCCC)c(c(ccc1)C(=O)OCCCC)c1,0.0021556100397968547
7
+ O=C(OCC)c(c(ccc1)C(=O)OCC)c1,0.019956158547022596
8
+ O=C(OC(OC(OC1C)C)C1)C,0.0007175892491582414
9
+ Oc(c(ccc1)C)c1C,4.911414454620217e-05
10
+ Oc(ccc(c1C)C)c1,0.00011459967060780387
11
+ O=C(OCC)C=C,0.002477130986890969
12
+ c(cccc1)(c1)CC,0.00384307445956764
13
+ OCCO,0.004027850816139285
14
+ c(ccc1C(=O)OCC(=O)OCC)cc1C(=O)OCC,0.008919866912731327
15
+ O=C,0.002730968314772715
16
+ O=C(O)C=CC(=O)O,0.009313172081918767
17
+ OCC(O)CO,0.07473899985905673
18
+ O=C(OC)c(ccc(O)c1)c1,0.009858865736182571
19
+ O=C(OCCC)c(ccc(O)c1)c1,0.008324062177858799
20
+ CC(CCC(=O)(O))C3CCC4C2CCC1CC(O)CCC1(C)C2CCC34C,0.001327765217118838
21
+ OC(C(CCC1C)C(C)C)C1,0.0037948308388560167
22
+ O=C(O)C(=C)C,0.0028807316686731447
23
+ O=C(OC)c(c(O)ccc1)c1,0.002366127776683785
24
+ Oc(cccc1)c1,0.0036552488310642134
25
+ O=C(OCCC)c(cc(O)c(O)c1O)c1,0.004071644352421929
26
+ OCC(O)C1C(O)=C(O)C(=O)O1,0.008823323006525195
27
+ c(cccc1)(c1)C=C,0.00020163396483810836
28
+ O=Cc(occ1)c1,0.0006244532131552316
29
+ NCCNc1cccc2ccccc12,0.00042415433290295584
30
+ CN(C)(C)CCCl,0.002242766507128513
31
+ O=C(Nc(ccc(c1)C(=O)CCl)c1)C,0.007465334624174676
32
+ c(ccc(c1)Cl)(c1)C(c(ccc(c2)Cl)c2)C(Cl)(Cl)Cl,9.027148189044099e-05
33
+ CC(Oc1cc(Cl)c(Cl)cc1Cl)C(=O)(O),3.2280916101231376e-05
34
+ O=N(=O)C(=CC=C1OC)C=C1N=NC(C(O)=C2C(=O)NC(=CC=C4)C=C4N(=O)=O)=C(C=C3)C(=C2)C=C3,0.004308389780762073
35
+ O=N(=O)C(C=C1)=CC(OCCO)=C1NCCO,0.0009453881078267498
36
+ Cc1cccc(CC)c1N(C(=O)CCl)COCC,0.00018534506246313786
37
+ C1=C(C(=CC=C1OC2=CC=C(C=C2Cl)C(F)(F)F)[N+](=O)[O-])C(=O)[O-].[Na+],0.0004691898654989019
38
+ CCc1cccc(CC)c1N(COC)C(=O)CCl,5.560351873894229e-05
39
+ O=C(Nc(ccc(OCC)c1N)c1)C,0.0061010029534003226
40
+ Oc(ccc(N)c1)c1,0.00628631814927864
41
+ CC(N)CC(=CC=C1)C=C1,3.6980547196719036e-05
42
+ O(c(ccc(c1)C=CC)c1)C,0.002321161271586114
43
+ COc1ccc(N)cc1,0.0038488877932280246
44
+ O=C(O)c(c(N)ccc1)c1,0.0200603809445196
45
+ Clc2cccc(c2)c1ccccc1,2.1202965065040615e-05
46
+ O=C(NC(C(=O)OC)Cc(cccc1)c1)C(N)CC(=O)O,0.0004994850207500305
47
+ n1c2ccc(Cl)cc2ncc1Oc3ccc(OC(C)C(=O)OCC)cc3,9.924832004782749e-06
48
+ COC(=O)NS(=O)(=O)c1ccc(N)cc1,0.0007817895162025788
49
+ S=P(OC)(OC)SCN1N=Nc2ccccc2C1(=O),1.1344859296501293e-06
50
+ CNC(=O)Oc1ccccc1OC(C)C,0.00023895810443138408
51
+ CC(C)(C)C(=O)C(Oc1ccc(Cl)cc1)n2cncn2,8.510674803234926e-05
52
+ O=S(O)(=O)C(=CC=C1)C=C1CN(CC)=C(C=C2)C=CC2=C(C(C=C3)=CC=C3N(C)C)C(C=C4)=CC=C4N(CC)CC(C=C5)=CC(=C5)S(=O)(=O)O,0.00100996317449829
53
+ c(c(cccc1)c1)(cccc2)c2,0.0016211890708511434
54
+ BrC(Cl)Cl,0.0007935120501519125
55
+ ClC(Cl)C(Cl)(Cl)SN2C(=O)C1CC=CCC1C2(=O),3.4377949341570856e-05
56
+ O=C(Oc(c(c(ccc1)cc2)c1)c2)NC,7.75266070321401e-05
57
+ CC1=C(SCCO1)C(=O)Nc2ccccc2,0.00012749566387247212
58
+ ClC1CC2C(C1Cl)C3(Cl)C(=C(Cl)C2(Cl)C3(Cl)Cl)Cl,6.588923229380691e-07
59
+ O=C(O)CCl,0.00031747032869396124
60
+ ClC(=CC=C1N)C=C1,4.703243372306972e-05
61
+ CC(C)OC(=O)C(O)(c1ccc(Cl)cc1)c2ccc(Cl)cc2,2.653199106614823e-05
62
+ n1c(OC)nc(C)nc1NC(=O)NS(=O)(=O)c2ccccc2Cl,6.987675250196435e-05
63
+ OS(=O)(=O)C(C(=CC=C2)C1=C2)=CC=C1N=NC(C(O)=C3N=NC(C(C=C5)=C4C=C5)=CC=C4S(O)(=O)=O)=CC(=C3O)CO,0.001209334683537979
64
+ S=P(OCC)(OCC)Oc1ccc2C(C)=C(Cl)C(=O)Oc2c1,2.205280759241576e-06
65
+ CNP(=O)(OC)Oc1ccc(cc1Cl)C(C)(C)C,1.3712205173148265e-05
66
+ C(C1C2C(C(O)C(O1)OC8C(OC(OC7C(OC(OC6C(OC(OC5C(C(C(OC4C(C(C(OC3C(C(C(O2)OC3CO)O)O)OC4CO)O)O)OC5CO)O)O)C(C6O)O)CO)C(C7O)O)CO)C(C8O)O)CO)O)O,0.0014097040502559495
67
+ n1c(N)nc(N)nc1NC2CC2,9.026150563412276e-05
68
+ COC(=O)c1c(Cl)c(Cl)c(C(=O)OC)c(Cl)c1Cl,0.001506186328985305
69
+ O=C(O)C(Cl)(Cl)C,0.0001970361896096649
70
+ Nc1cc(N)c(O)cc1,0.00020138468889932168
71
+ FC(F)(Cl)Cl,0.0012405561997662923
72
+ ClCCl,0.0005887022388817134
73
+ O=P(OC)(OC)OC=C(Cl)Cl,1.0408382339127476e-05
74
+ OC(c1ccc(Cl)cc1)(c2ccc(Cl)cc2)C(Cl)(Cl)Cl,5.398319600278167e-05
75
+ ClC4=C(Cl)C5(Cl)C3C1CC(C2OC12)C3C4(Cl)C5(Cl)Cl,1.312648375209102e-07
76
+ CN(=C1C(C=C2)=CC=C2)N(C)C(=C1)C(C=C3)=CC=C3,0.0005013430378371984
77
+ O=C(NC(=O)c(c(F)ccc1)c1F)Nc(ccc(c2)Cl)c2,2.574969708764491e-05
78
+ CC1=C(C)S(=O)(=O)CCS1(=O)=O,4.755763033644171e-05
79
+ O=C(NC)CSP(OC)(OC)=S,1.090477146170362e-06
80
+ COc1ccc(N)c(OC)c1,0.0018018201517132695
81
+ COP(=O)OC,0.0009086866178930357
82
+ CC(=C(N(=O)=O)C=C1N(=O)=O)C=C1,0.00018667621570414935
83
+ CN(C)C(=O)C(c1ccccc1)c2ccccc2,0.0001253592168358418
84
+ N(c(cccc1)c1)c(cccc2)c2,0.00018319083450161921
85
+ C(C=C1)(=N(C=C1)CC2)C(N2=C3)=CC=C3,3.1481176319120814e-06
86
+ CCOP(=S)(OCC)SCCSCC,3.644261466562107e-07
87
+ NC(=S)NNC(N)=S,0.006303842268414047
88
+ O=C(N(C)C)Nc(ccc(c1Cl)Cl)c1,2.5740633090871022e-05
89
+ O=P(O)(O)CCCl,0.0010381053812746242
90
+ O=C(OCC)C(O1)C1(c(cccc2)c2)C,0.0008485352051922901
91
+ COC(=O)NC(=NC1=C2)NC1=CC(=C2)SC(C=C3)=CC=C3,5.010896695954994e-05
92
+ CN1C=C(c2ccccc2)C(=O)C(c3cc(C(F)(F)F)ccc3)=C1,7.591497980909371e-05
93
+ c1cc(C(F)(F)F)cc(Cl)c1NC(C(C)C)C(=O)OC(C(#N))c2cccc(Oc3ccccc3)c2,4.971041796516284e-06
94
+ S=P(OCC)(Sc1ccccc1)CC,6.4141791096429385e-06
95
+ n1c(C)nc(OC)nc1NC(=O)NS(=O)(=O)c2ccsc2C(=O)OC,6.453419527613866e-05
96
+ C(C(C(C(C1Cl)Cl)Cl)Cl)(C1Cl)Cl,1.719218358061182e-05
97
+ O=N(=O)N(CN1N(=O)=O)CN(C1)N(=O)=O,6.753217705640154e-06
98
+ O=C(N=C(N(C1(=O))C)N(C)C)N1C(CCCC2)C2,0.00019816672003956806
99
+ n(c(c(ccc1)cc2)c1O)c2,0.0009851335765350196
100
+ c1cc(Cl)cc(Cl)c1C(OCC=C)Cn2cncc2,0.00013459866849613327
101
+ COc1cccc(OC)c1C(=O)Nc2onc(C(C)(CC)CC)c2,0.00015252975563710417
102
+ n1c(OC)cc(OC)nc1NC(=O)NS(=O)(=O)Cc2ccccc2C(=O)OC,0.000752920821092069
103
+ CCOC(=O)CC(SP(=S)(OC)OC)C(=O)OCC,0.0001513509490359852
104
+ CN(C)(CCC1)CC1,0.0013133857473480218
105
+ O=P(SCCCC)(SCCCC)SCCCC,3.9744245336126525e-06
106
+ COCC(=O)N(C(C)C(=O)OC)c1c(C)cccc1C,0.00022374845318219082
107
+ S=P(OC)(OC)SCN1C(=O)SC(OC)=N1,6.615259463326349e-06
108
+ CNC(=O)ON=C(C)SC,6.164844235963149e-05
109
+ COP(=S)(OC)Oc1ccc(cc1)N(=O)(=O),9.498210994862414e-07
110
+ O=C1N(N)C(SC)=NN=C1C(C)(C)C,6.99992664076887e-05
111
+ COP(=O)(OC)OC(Br)C(Cl)(Cl)Br,5.252325098618077e-06
112
+ OC(C(N)C1O)C(C)OC1(C)OC(CC(C)(C(C2O)C(O)=O)OC(O)(C2)CC(O)CC(C)(O3)C3C=C4)C=CC=CC=CC=CCC(C)OC4=O,0.0001017229436608032
113
+ O=N(=O)c(ccc(c1N)C)c1,5.2579476836834245e-05
114
+ O=N(=O)c(c(N)ccc1N)c1,0.0005681125108300581
115
+ O=C(O)C(=C(N)C=C1N(=O)=O)C=C1,0.006506215164982737
116
+ O=N(=O)c(c(c(ccc1)cc2)c1)c2,0.0009528314918084262
117
+ c12c(N=Nc3ccccc3)c(O)ccc1cc(S(=O)(=O)O)cc2,0.0005482080783455076
118
+ CC(C)Oc1cc(c(Cl)cc1Cl)N2N=C(OC2(=O))C(C)(C)C,1.448347496337259e-05
119
+ CNC(=O)ON=C(SC)C(=O)N(C)C,2.2803829328479105e-05
120
+ CCOP(=S)(OCC)Oc1ccc(cc1)N(=O)=O,1.20167291684791e-05
121
+ Oc(c(c(c(c1Cl)Cl)Cl)Cl)c1Cl,3.754648160556525e-05
122
+ NC(=N)NC(=N)NCCc1ccccc1,0.00035564719019232516
123
+ COP(=S)(OC)SCN2C(=O)c1ccccc1C2(=O),6.30276515448593e-05
124
+ CCN(CC)C(=O)C(Cl)=C(C)OP(=O)(OC)OC,4.1042640430422796e-05
125
+ ClC3C6(Cl)C4C2C1OC1C5C2C3(Cl)C(Cl)(C45)C6(Cl)Cl,1.8377077252927096e-06
126
+ O=C(OC(=O)c1cccc2)c12,0.00800050987215652
127
+ Nc1c(Cl)c(Cl)nc(C(=O)(O))c1Cl,0.0002484891651683468
128
+ CCN(CC)c1nc(C)cc(OP(=S)(OC)OC)n1,8.187766820693496e-06
129
+ Nc3ccc2cc1ccc(N)cc1nc2c3,0.00022461542255370235
130
+ CCC(=O)Nc1ccc(Cl)c(Cl)c1,9.170952329114616e-05
131
+ Clc1cc(Cl)ccc1C2(Cn3ncnc3)OC(CCC)CO2,7.305234130124016e-05
132
+ O=C(N)c(nccn1)c1,0.006408762052980695
133
+ Oc1cc(O)c2C(=O)C(O)=C(c3cc(O)c(O)cc3)Oc2c1,0.006729846937340681
134
+ CCC(O)(C)C#C,0.00046870383012542657
135
+ CC(C(NCC)=C1)=CC(C1=O2)=C(C(C2=C3)=CC(C)=C3NCC)C(=CC=C4)C(=C4)C(=O)OCC,2.705399937694626e-05
136
+ O=C(NS(=O)(=O)c1cccc2)c12,0.01966323569952717
137
+ c1cc(Cl)ccc1C2SC(=O)N(C(=O)NC3CCCCC3)C2C,0.00045341341521072437
138
+ n(c(nc(n1)NCC)NCC)c1Cl,2.4794616275543184e-05
139
+ O=[S](NC1CCCCC1)(=O)[O-].[Na+],0.017900805088957215
140
+ O=C(OCC(C1OCC(C1O)O)O)CCCCCCCCCCC,0.01986671090855914
141
+ O(CC1O)C(C1O)C(O)COC(=O)CCCCCCCCCCCCCCCCC,0.016727105323218583
142
+ O=S(=O)(Nc(nc(cc1C)C)n1)c(ccc(N)c2)c2,0.00011856422602566643
143
+ CCNc1nc(NC(C)(C)C)nc(SC)n1,6.21487662475519e-05
144
+ Oc(c(cc(c1)C(C)(C)C)Cl)c1,0.0011697007223226948
145
+ C(C(Cl)Cl)(Cl)Cl,0.000643434395429049
146
+ COP(=O)(OC)OC(=CCl)c1cc(Cl)c(Cl)cc1Cl,0.0002732525478388646
147
+ CCN(CC)C(=O)SCc1ccc(Cl)cc1,1.939641912620375e-05
148
+ COC(=O)NC(=S)Nc1ccccc1NC(=S)NC(=O)OC,9.34595925699157e-05
149
+ N(C(=S)SSC(N(C)C)=S)(C)C,6.238747379310154e-05
150
+ c12OC(CCCC(C)CCCC(C)CCCC(C)C)(C)CCc1c(C)c(OC(=O)C)c(C)c2C,0.004230630449818796
151
+ Cc1cc(N)ccc1NOS(O)(=O)=O,0.0008431459792705265
152
+ C(Br)(C(Br)(Br)Br)C1C(C)(C)C1C(=O)OC(C(#N))c2cccc(Oc3ccccc3)c2,4.511229623452442e-06
153
+ O=C(O)COc(c(cc(c1Cl)Cl)Cl)c1,3.914162418169536e-05
154
+ FC(F)(F)C(=CC(N(=O)=O)=C1N(C(C)C)C(C)C)C=C1N(=O)=O,0.00011930360709302091
155
+ Cc1cc(C)c(N)cc1C,0.0004659548946786565
156
+ CC(O)(C(O)C(O1)C)CC1(C)OC(C(C)O2)C(C(O)C2(C)OC(C(C)C(O)CC(=O)OC(CC)C3COC(C(OC)C4OC)OC(C)C4O)C(CC=O)CC(C)C(=O)C=CC(=C3)C)N(C)C,0.0005295750507618813
157
+ c1c(Cl)cc(Cl)cc1N2C(=O)C(C)(C=C)OC2(=O),0.00025479642918707515
158
+ O=C(OC(CCCC(O)CCCCCc1cc(O)cc2O)C)c12,6.203550142861481e-07
159
+ COC(=O)C1(C2=CC=CC=C2C3=C1C=C(C=C3)Cl)O,0.0005460521449219469
160
+ CC(C(=O)O)OC1=CC(=CC=C1)Cl,0.0004984573741185808
161
+ P12P3P1P23,0.011881024070664265
162
+ C(CO)O,0.006444561305822786
163
+ CCCCOCC(C)OCC(C)O,0.0006726932978936104
164
+ C(CO)O,0.03222280652911401
165
+ C(CO)O,0.01482249100339231
166
+ [O-][As](=O)([O-])[O-],4.4990181342823836e-05
167
+ [Si](CN1C=NC=N1)(C2=CC=C(C=C2)F)C3=CC=C(C=C3)F,7.657523930235053e-06
168
+ N(C(=S)SSC(N(C)C)=S)(C)C,4.783039657471118e-05
169
+ COP(=O)(N)SC,6.377136136005779e-06
170
+ N(C(=S)SSC(N(C)C)=S)(C)C,2.2750632109884408e-05
171
+ COP(=O)(NC(=O)(C))SC,0.00019108364303760538
172
+ C1=CC=C(C=C1)NC(=O)NC2=CN=NS2,0.00013620822278144422
173
+ CCOP(=S)(NC(C)C)OC1=CC=CC=C1C(=O)OC(C)C,1.4476216287421915e-06
174
+ CC(=NOC(=O)N(C)SN(C)C(=O)ON=C(C)SC)SC,2.821118623185759e-05
175
+ CCOP(=S)(OCC)OC1=NC(=NC(=C1)C)C(C)C,4.928609081032476e-06
176
+ NC(CCCC1)C1,0.000589871631832977
177
+ CN1C=C(c2ccccc2)C(=O)C(c3cc(C(F)(F)F)ccc3)=C1,9.868947375182025e-05
178
+ ClC1CC2C(C1Cl)C3(Cl)C(=C(Cl)C2(Cl)C3(Cl)Cl)Cl,1.2201709684038252e-06
179
+ CCCCC(CC)COC(=O)C1=CC=CC=C1C(=O)OCC(CC)CCCC,0.0005120902983161576
180
+ OC(=O)CNCP(O)(O)=O,0.005914602100378093
181
+ C1CNC(=S)N1,2.2514113902230536e-06
182
+ O=C(N(OC)C)Nc(ccc(c1Cl)Cl)c1,2.5090939601491425e-05
183
+ C1=CC=C2C(=C1)NC(=S)S2,0.004484270077422451
184
+ CCOP(=S)(OCC)OC1=NC(=C(C=C1Cl)Cl)Cl,2.852364730588813e-05
185
+ c(c(c(c(c1Cl)Cl)Cl)Cl)(c1Cl)Cl,1.0183220720957906e-06
186
+ COc1ccc(cc1)C(c2ccc(OC)cc2)C(Cl)(Cl)Cl,0.000361639482467859
187
+ C1=CC(=CC=C1Cl)Cl,0.0020407891160090755
188
+ CC(C)OC(=O)NC1=CC(=CC=C1)Cl,0.002340158076742
189
+ COP(=O)(OC)OC=C(Cl)Cl,9.729574795271285e-06
190
+ CCOP(=S)(OCC)Oc1ccc(cc1)N(=O)=O,1.442007500217475e-06
191
+ CNC(=O)N(C)c1nnc(s1)C(C)(C)C,0.00017519690160775725
192
+ CCCCOCCOCCOCC1=CC2=C(C=C1CCC)OCO2,0.0007386866446931963
193
+ CC(C(=O)O)OC1=C(C=C(C=C1)Cl)Cl,3.828744186371015e-05
194
+ CC1(C(C1C(=O)OC(C#N)C2=CC(=C(C=C2)F)OC3=CC=CC=C3)C=C(Cl)Cl)C,2.8782768221426878e-05
195
+ C(#N)c(c(c(c(c1C(#N))Cl)Cl)Cl)c1Cl,1.504262704438716e-05
196
+ O=C(OCC)C(O)(c(ccc(c1)Cl)c1)c(ccc(c2)Cl)c2,5.658290428731166e-05
197
+ O=C(N(C)C)Nc(ccc(c1)Cl)c1,0.0006292491939569566
198
+ O=C(N(SC(Cl)(Cl)Cl)C(=O)C1CC=CC2)C12,0.00033267981710061916
199
+ CCc1cccc(C)c1N(C(C)COC)C(=O)CCl,0.0005285529966699722
200
+ C1=CC(=C(C=C1Cl)Cl)OCC(=O)O,2.2620602193003837e-05
201
+ CCNC1=NC(=NC(=N1)Cl)NC(C)(C)C#N,5.193343612553005e-06
202
+ C(C(C(C(C1Cl)Cl)Cl)Cl)(C1Cl)Cl,1.581680889416305e-05
203
+ COP(=S)(OC)OC1=CC(=C(C=C1Cl)Cl)Cl,0.00015549919110720132
204
+ C1C2C=CC1C3C2C4(C(=C(C3(C4(Cl)Cl)Cl)Cl)Cl)Cl,2.7404023436797595e-08
205
+ CC1(C(C1C(=O)OCC2=CC(=CC=C2)OC3=CC=CC=C3)C=C(Cl)Cl)C,6.389160712181889e-05
206
+ CCN(CC)C(=O)C(C)OC1=CC=CC2=CC=CC=C21,0.0003685221091522648
207
+ ClC1C=CC2C1C3(Cl)C(=C(Cl)C2(Cl)C3(Cl)Cl)Cl,6.696708996117785e-07
208
+ CC(=CC(=O)NC)OP(=O)(OC)OC,2.016458594951112e-06
209
+ CC(C)C1(C)N=C(NC1(=O))c3nc2ccccc2cc3C(=O)(O),6.42394476589514e-05
210
+ CC(C)Nc1nc(Cl)nc(NC(C)C)n1,0.00021766590408142935
211
+ CC(C(=O)O)(Cl)Cl,0.000349726996112295
212
+ CC1(C(C1C(=O)OC(C#N)C2=CC(=C(C=C2)F)OC3=CC=CC=C3)C=C(Cl)Cl)C,1.7269660932855995e-05
213
+ O=C(NC)CSP(OC)(OC)=S,2.1809542923407198e-05
214
+ C12C3(C4(C5(C3(C(C1(C5(C2(C4(Cl)Cl)Cl)Cl)Cl)(Cl)Cl)Cl)Cl)Cl)Cl,1.9565721591442967e-08
215
+ COC(=O)C1=CC=CC=C1C(=O)OC,0.010299509743336183
216
+ CCOP(=S)(OCC)SCSC(C)(C)C,1.7335192530420976e-07
217
+ CCC(C)SP(=O)(OCC)SC(C)CC,9.245829486467153e-07
218
+ CCOP(=S)(OCC)SCSC(C)(C)C,6.93407701216835e-06
219
+ C1C(C(C(=O)N1C2=CC=CC(=C2)C(F)(F)F)Cl)CCl,1.601973068976987e-05
220
+ O=C(O)C(C(C(=O)O)C(O1)CC2)C12,0.0006177415369409428
221
+ O=C(Oc(c(OC(C1)(C)C)c1cc2)c2)NC,2.25986249188711e-05
222
+ Oc(c(c(c(c1)Cl)Cl)Cc(c(c(cc2Cl)Cl)Cl)c2O)c1Cl,1.2287924553323e-05
223
+ CC1(CON(C1=O)CC2=CC=CC=C2Cl)C,8.969617860069388e-05
224
+ CCC(C)N1C(=O)C(=C(NC1=O)C)Br,0.00023935747721355314
225
+ CC1=CC(=CC(=C1N(C)C)C)OC(=O)NC,6.748138593450308e-06
226
+ CNC(=O)OC1=CC=CC(=C1)N=CN(C)C,5.6495719658295366e-05
227
+ CC1=NN(C(=O)N1C(F)F)C2=CC(=C(C=C2Cl)Cl)NS(=O)(=O)C,0.0001730417009651666
228
+ CCOP(=S)(CC)SC1=CC=CC=C1,2.029803515709809e-05
229
+ CCOP(=S)(OCC)SC(CCl)N1C(=O)C2=CC=CC=C2C1=O,6.347661292175554e-06
230
+ N1CC(C)(C)CNC1=NN=C(C=Cc2ccc(C(F)(F)F)cc2)C=Cc3ccc(C(F)(F)F)cc3,1.011172895860315e-05
231
+ CC1=C(C(=C(C(=C1F)F)COC(=O)C2C(C2(C)C)C=C(C(F)(F)F)Cl)F)F,1.0985502703376482e-05
232
+ CC1=CC(=C(C=C1)N=CN(C)C=NC2=C(C=C(C=C2)C)C)C,3.476411288357328e-05
233
+ S=P(OC)(OC)SCN1C(=O)SC(OC)=N1,2.646103785330524e-06
234
+ CC(C)N(C(C)C)C(=O)SCC(Cl)=C(Cl)Cl,4.1028786650112265e-05
235
+ S=P(OC)(OC)SCN1N=Nc2ccccc2C1(=O),7.0905370603133915e-06
236
+ CC1=NC(=NC(=N1)OC)NC(=O)NS(=O)(=O)C2=CC=CC=C2CCC(F)(F)F,0.00019051986068493621
237
+ O=C(N(S(=O)(=O)Nc1cccc2)C(C)C)c12,0.0001664732247794737
238
+ CCN(CC)c1nc(C)cc(OP(=S)(OC)OC)n1,4.519647285022815e-05
239
+ O=C(ON=CC(SC)(C)C)NC,5.255875464343493e-07
240
+ ClC(Cl)(Cl)CC1(OC1)c2cc(Cl)cc(Cl)c2,9.362507489225856e-05
241
+ CCC1CCCC(C(C(=O)C2CC3C(C2CC(=O)O1)CCC4C3CC(C4)OC5CC(C(C(C5OC)OC)OC)C)C)OC6CCC(C(O6)C)N(C)C,3.269690443692059e-05
242
+ CC(C)(C)C(=NOC(=O)NC)CSC,2.748304502244953e-05
243
+ CON=C(CC1=CN=CC=C1)C2=C(C=C(C=C2)Cl)Cl,0.00015245767876475813
244
+ CC(=CC1C(C1(C)C)C(=O)OCN2C(=O)C3=C(C2=O)CCCC3)C,0.0007543614918373588
245
+ C1COC(O1)(CN2C=NC=N2)C3=C(C=C(C=C3)Cl)Cl,3.3317713989015466e-05
246
+ CCCOC(=O)C1=CN=C(C=C1)C(=O)OCCC,0.0009949124950582777
247
+ CC1=CC(=NC(=N1)NC(=O)NS(=O)(=O)C2=CC=CC=C2C(=O)OC3COC3)C,0.0002042257406025049
248
+ C1=C(C(=NC(=C1Cl)Cl)OCC(=O)O)Cl,0.00014036698793030923
249
+ CCOCN1C(=C(C(=C1C(F)(F)F)Br)C#N)C2=CC=C(C=C2)Cl,3.336499331006338e-05
250
+ CC(C)CC1=C(C(=NC(=C1C(=O)SC)C(F)(F)F)C(F)F)C(=O)SC,9.043009089325258e-06
251
+ CC12CC1(C(=O)N(C2=O)C3=CC(=CC(=C3)Cl)Cl)C,5.279126047017924e-05
252
+ CC1=CC(=C(C=C1)C(=O)OC)C2=NC(C(=O)N2)(C)C(C)C,0.00017340543300030297
253
+ COP(=S)(OC)OC1=NC(=C(C=C1Cl)Cl)Cl,9.30136974668287e-06
254
+ CC(C)CC1=C(C(=NC(=C1C(=O)OC)C(F)F)C(F)(F)F)C2=NCCS2,0.00011151045308574189
255
+ CCOC(=O)C(CC1=CC(=C(C=C1Cl)F)N2C(=O)N(C(=N2)C)C(F)F)Cl,2.9112705183968275e-05
256
+ CC(C)=CC3C(C(=O)OCc2coc(Cc1ccccc1)c2)C3(C)C,0.00036934164172773375
257
+ CCCSP(=S)(OCC)OC1=CC=C(C=C1)SC,3.566479571525996e-05
258
+ CC1=CC(=C(C(=C1)OC(=O)NC)C)C,0.0003063511456860093
259
+ CC1=CC=CC=C1COC2CC3(CCC2(O3)C)C(C)C,0.0005466515334085693
260
+ CC1=C2C(=CC=C1)SC3=NN=CN23,0.00016381576159163087
261
+ CCC(=C1C(=O)CC(CC1=O)CC(C)SCC)NOCC=CCl,0.00027784628232227726
262
+ CCCN(CCC)C(=O)SCC,4.753899597429217e-05
263
+ CC(C)OC(=O)C=C(C)C=CCC(C)CCCC(C)(C)OC,0.00014816176662421748
264
+ COP(=S)(OC)Oc1ccc(SC)c(C)c1,1.3473309110575038e-05
265
+ COC1=C(C=C(C=C1)C(=CC(=O)N2CCOCC2)C3=CC=C(C=C3)Cl)OC,0.00011937399144446824
266
+ CCSC(=O)N(CC(C)C)CC(C)C,0.00046004207912889866
267
+ CC(C)OP(=S)(OC(C)C)SCCNS(=O)(=O)C1=CC=CC=C1,3.773457500445001e-05
268
+ CC(=CC1C(C1(C)C)C(=O)OCC2=CC(=CC=C2)OC3=CC=CC=C3)C,0.0004280202119133742
269
+ CC1=CC(=CC(=C1C)C)OC(=O)NC,5.1748504338852954e-05
270
+ CCOP(=S)(OCC)SCSC1=CC=C(C=C1)Cl,2.91659726744992e-06
271
+ C1CN(CCN1C(C(Cl)(Cl)Cl)NC=O)C(C(Cl)(Cl)Cl)NC=O,0.00022990526799413375
272
+ C(=CC=C1)(C2=C1)NC(=N2)C(=CS3)N=C3,9.938002763559794e-06
273
+ C1(=C(C(=C(C(=C1Cl)Cl)Cl)Cl)Cl)[N+](=O)[O-],0.00042324860573444577
274
+ CCCC(=NOCC)C1C(=O)CC(CC1=O)CC(C)SCC,5.056765552287114e-05
275
+ ClC2(Cl)C4(Cl)C1(Cl)C5(Cl)C(Cl)(Cl)C3(Cl)C1(Cl)C2(Cl)C3(Cl)C45Cl,1.2831252531880975e-06
276
+ CCN(C1CCCCC1)C(=O)SCC,1.3930451940080122e-05
277
+ CC1=NC(=NC(=N1)OC)NC(=O)NS(=O)(=O)C2=CC=CC=C2OCCCl,0.0005494924735209557
278
+ C(C(=O)O)OC1=NC(=C(C(=C1Cl)N)Cl)F,0.001960549023239737
279
+ CC1=C(C(=O)CC1OC(=O)C2C(C2(C)C)C=C(C)C)CC#C,0.0003162533656848333
280
+ C1C(COC1(CN2C=NC=N2)C3=C(C=C(C=C3)Cl)Cl)Br,1.7185416964361658e-05
281
+ C1=NNC(=N1)N,2.973360120532913e-05
282
+ C1=CC(=C(C(=C1)F)C(=O)NC(=O)NC2=CC(=C(C(=C2F)Cl)F)Cl)F,6.822190699642811e-05
283
+ C1=CC(=CC=C1OS(=O)(=O)C2=CC=C(C=C2)Cl)Cl,8.246440044818333e-06
284
+ FC(F)(F)C(=CC(N(=O)=O)=C1N(C(C)C)C(C)C)C=C1N(=O)=O,0.0029825901773255373
285
+ CC1=C(C(=O)CC1OC(=O)C2C(C2(C)C)C=C(C)C)CC=C,0.0002480093611298675
286
+ C1=CC(=C(C(=C1)Cl)C#N)Cl,1.4533918736325693e-05
287
+ C1C(O1)COC2=CC=CC=C2C3=CC=CC=C3,0.00220974492207247
288
+ CC1=CC=CC=C1OCC2=CC=CC=C2C(=NOC)C(=O)OC,0.001196753409055808
289
+ O=N(=O)C(C(=C1N(=O)=O)N(C(C)C)C(C)C)=CC(=C1)S(=O)(=O)N,0.00010642121227099495
290
+ C1=CC(=C(C2=NC=C(C=C21)Cl)C(=O)O)Cl,0.0031273470595088652
291
+ CC(C)NC(=O)N1CC(=O)N(C1=O)C2=CC(=CC(=C2)Cl)Cl,0.00013932359364492863
292
+ CCCN(CCCl)C1=C(C=C(C=C1[N+](=O)[O-])C(F)(F)F)[N+](=O)[O-],0.0005690227880626769
293
+ C1=CC=C(C=C1)C2=CC=CC=C2O,0.005875192118782279
294
+ O=C(N(C)C)Nc(cccc1C(F)(F)F)c1,6.459882953742444e-05
295
+ C1=CC(=NC(=C1)Cl)C(Cl)(Cl)Cl,4.330753128362797e-06
296
+ Clc1cc(Cl)cc(Cl)c1OCCN(CCC)C(=O)n2cncc2,1.9911569269535424e-05
297
+ CC1=CC(=C(C=C1NC(=O)C)NS(=O)(=O)C(F)(F)F)C,8.894826519325488e-05
298
+ C(=C(I)I)(I)I,9.40487316889e-05
299
+ C1=C(C=C(C(=C1Cl)N2C(=C(C(=N2)C#N)S(=O)C(F)(F)F)N)Cl)C(F)(F)F,1.3725335654181613e-07
300
+ O=C(N(C)C)Nc(ccc(c1Cl)Cl)c1,2.6813159469657174e-05
301
+ CC1=C(C(=O)CC1OC(=O)C2C(C2(C)C)C=C(C)C)CC=C,8.101639130242488e-05
302
+ CC1=CC2=C(C=C1)N=C3C(=N2)SC(=O)S3,3.201059303080756e-05
303
+ CC(C)N(C(=O)CCl)c1ccccc1,0.0001086504872549198
304
+ CC(C)C1(C(=O)NC(=N1)C2=C(C=CC=N2)C(=O)O)C,0.0019136814830266148
305
+ CC1(C(C1(C)C)C(=O)OC(C#N)C2=CC(=CC=C2)OC3=CC=CC=C3)C,5.566320606558907e-05
306
+ CN1CN(C(=S)SC1)C,0.0001848698793354276
307
+ ClC(Cl)=CC1C(C)(C)C1C(=O)OC(C(#N))c2cccc(Oc3ccccc3)c2,0.0001801597685653181
308
+ ClC2C1OC1C3C2C4(Cl)C(=C(Cl)C3(Cl)C4(Cl)Cl)Cl,5.137200498000178e-07
309
+ CCOC(=O)CN1C2=C(C=CC=C2Cl)SC1=O,4.600323862799974e-05
310
+ CCCN(CCC)C1=C(C=C(C(=C1[N+](=O)[O-])N)C(F)(F)F)[N+](=O)[O-],8.392957359172463e-05
311
+ C1=CC=C(C=C1)C(CCC2=CC=C(C=C2)Cl)(CN3C=NC=N3)C#N,0.0001187584704479041
312
+ CC(C)(C)C(CCC1=CC=C(C=C1)Cl)(CN2C=NC=N2)O,5.165383561566422e-05
313
+ CC1=C(C=CC=C1COC(=O)C2C(C2(C)C)C=C(C(F)(F)F)Cl)C3=CC=CC=C3,1.1824026617703973e-05
314
+ C(=CC=C1)(C2=C1)NC(=N2)C(=CS3)N=C3,0.0001987600552711955
315
+ C1=C(C=C(C(=C1Cl)N)Cl)[N+](=O)[O-],0.0007245881151318271
316
+ CC1=C(C=C(C=C1C(=O)N)[N+](=O)[O-])[N+](=O)[O-],2.775825077363338e-05
317
+ CC(C)OC1=CC=CC(=C1)NC(=O)C2=CC=CC=C2C(F)(F)F,0.0015465050319982663
318
+ OC(c1ccc(Cl)cc1)(c2ccc(Cl)cc2)C(Cl)(Cl)Cl,6.747899500347761e-06
319
+ O=P(O)(O)CCCl,0.0030866333336565456
320
+ CN(C(=O)NC1=CC=C(C=C1)Br)OC,4.8243951057630326e-05
321
+ CC1=CC(=CC=C1)NC(=O)OC2=CC=CC(=C2)NC(=O)OC,8.324752178787434e-05
322
+ CC(C)(C)C1=NN=C(S1)N2C(CN(C2=O)C)O,0.0001950651330281793
323
+ S=P(OCC)(OCC)Oc1ccc2C(C)=C(Cl)C(=O)Oc2c1,4.686221613388349e-06
324
+ COC(=O)c1ccccc1S(=O)(=O)NC(=O)N(C)c2nc(OC)nc(C)n2,3.1614325062739156e-05
325
+ C1=CC=C(C(=C1)NC2=NC(=NC(=N2)Cl)Cl)Cl,4.173898399328064e-06
326
+ CC1=C(C=CC(=C1)OP(=S)(OC)OC)[N+](=O)[O-],1.6592478987816552e-06
327
+ COc1c(Cl)ccc(Cl)c1C(=O)(O),0.000520273850439097
328
+ CC1(C(C1C(=O)OC(C#N)C2=CC(=CC=C2)OC3=CC=CC=C3)C=C(Br)Br)C,4.948543461552819e-06
329
+ C1=CC=C(C(=C1)C(C2=CC=C(C=C2)F)(C3=CN=CN=C3)O)Cl,7.943029208877253e-06
330
+ c1ccc2nc(NC(=O)OC)n(C(=O)NCCCC)c2c1,0.00039611774300238816
331
+ CCOC(=O)COC(=O)C1=C(C=CC(=C1)OC2=C(C=C(C=C2)C(F)(F)F)Cl)[N+](=O)[O-],5.583516196615599e-05
332
+ CC(C)(C)C(C(N1C=NC=N1)OC2=CC=C(C=C2)C3=CC=CC=C3)O,7.409262028018131e-05
333
+ CCNC(=O)NC(=O)C(=NOC)C#N,0.00015289185096526158
334
+ CCOC1=C(C=CC(=C1)OC2=C(C=C(C=C2)C(F)(F)F)Cl)[N+](=O)[O-],0.00011058877892773892
335
+ CC1=NC=C(N1CCO)[N+](=O)[O-],0.0008764039114257181
336
+ O=C(N(SC(Cl)(Cl)Cl)C(=O)c1cccc2)c12,0.0016860133324539056
337
+ CC(C)(C)c2ccc(OC1CCCCC1OS(=O)OCC#C)cc2,0.000285329221701204
338
+ CCCCC(CN1C=NC=N1)(C2=C(C=C(C=C2)Cl)Cl)O,1.4958135679074663e-05
339
+ CC(C)(C)C(C(=CC1=C(C=C(C=C1)Cl)Cl)N2C=NC=N2)O,0.00015327033840680795
340
+ C1=CC(C2C1C3(C(=C(C2(C3(Cl)Cl)Cl)Cl)Cl)Cl)Cl,1.0178997674098972e-06
341
+ C1=CC=C2C(=C1)C(=O)C3=C(C2=O)SC(=C(S3)C#N)C#N,3.374687200243432e-05
342
+ CC1=NN(C(=C1C=NOCC2=CC=C(C=C2)C(=O)OC(C)(C)C)OC3=CC=CC=C3)C,7.307428846046914e-06
343
+ CS(=O)(=O)C1=C(C=CC(=C1)C(F)(F)F)C(=O)C2=C(ON=C2)C3CC3,5.566064755837833e-05
344
+ CC1=C(N=C(N=C1OC(=O)N(C)C)N(C)C)C,5.245792224158045e-05
345
+ CCN(CC)C(=O)C(Cl)=C(C)OP(=O)(OC)OC,5.005200052490544e-06
346
+ C1=CC(=CC=C1C(CN)O)O,0.0016320834707547742
347
+ CC1=C(C(=CC=C1)C)N(C(=O)COC)N2CCOC2=O,0.00017965983350851395
348
+ c1c(C(F)(F)F)cccc1N2C(=O)C(Cl)=C(NC)C=N2,6.174515120168466e-05
349
+ CCC(C)NC1=C(C=C(C=C1[N+](=O)[O-])C(C)(C)C)[N+](=O)[O-],0.00016929970598735823
350
+ C1=CC(=CC=C1S(=O)(=O)C2=CC(=C(C=C2Cl)Cl)Cl)Cl,0.0006459733503975096
351
+ CCCCC1=C(NC(=NC1=O)NCC)C,0.00011945258606726158
352
+ n(c(nc(n1)NCC)NCC)c1Cl,2.628229325207573e-05
353
+ FC(F)(F)C(C=C1N(=O)=O)=CC(N(=O)=O)=C1N(CC)CC(C)=C,3.750789508184508e-05
354
+ C1CCC(C1)N(CC2=CC=C(C=C2)Cl)C(=O)NC3=CC=CC=C3,7.60257762657494e-05
355
+ CS(=O)(=O)NC(=O)C1=C(C=CC(=C1)OC2=C(C=C(C=C2)C(F)(F)F)Cl)[N+](=O)[O-],0.00011395676094313254
356
+ CCOC(=O)C(C)OC1=CC=C(C=C1)OC2=NC3=C(O2)C=C(C=C3)Cl,2.4877248744348747e-05
357
+ CCC1=C(C(=CC=C1)CC)N(CNC(=O)C)C(=O)CCl,0.0002105848787792573
358
+ NC(=N)NCCCCCCCCCCCC(OC(=O)C),0.0001016026806851265
359
+ C1=CC(=CC(=C1)Cl)NC(=O)OCC#CCCl,0.001743505808935171
360
+ CC(C)C(C(=O)OC(C(#N))c2cccc(Oc1ccccc1)c2)c3ccc(Cl)cc3,5.953797389131287e-05
361
+ CC(C)C1=C(C=CC(=C1)C(C)(C)C2=CC(=C(C=C2)O)C(C)C)O,8.001387248515537e-05
362
+ CCN(CC1=C(C=CC=C1Cl)F)C2=C(C=C(C=C2[N+](=O)[O-])C(F)(F)F)[N+](=O)[O-],0.00011855904490168938
363
+ CCCCCCCCSC(=O)OC1=CC(=NN=C1C2=CC=CC=C2)Cl,0.00017813968959673758
364
+ CC1=C(C(=O)CC1OC(=O)C2C(C2(C)C)C=C(C)C)CC=C,4.563372244789602e-05
365
+ CCCCNC(=O)N1C2=CC=CC=C2N=C1NC(=O)OC,0.0008611255282660687
366
+ CC1=C(C(=CC=C1)C)N(C(C)C(=O)OC)C(=O)CC2=CC=CC=C2,0.00014136381415796835
367
+ CC(C)C(C1=CC=C(C=C1)OC(F)F)C(=O)OC(C#N)C2=CC(=CC=C2)OC3=CC=CC=C3,1.3290157262750028e-05
368
+ Clc1ccccc1c2nnc(c3ccccc3Cl)nn2,6.597478470118595e-05
369
+ CCOC(=O)NCCOC1=CC=C(C=C1)OC2=CC=CC=C2,3.318543029523173e-05
370
+ CCC1=C(C(=CC=C1)CC)N(CC(=O)OCC)C(=O)CCl,0.00016035726058227922
371
+ c1(O2)c(CC2(C)C)cccc1OC(=O)N(C)SN(CCCC)CCCC,6.569530810416249e-05
372
+ CCCCOC(=O)C(C)OC1=CC=C(C=C1)OC2=NC=C(C=C2)C(F)(F)F,7.825509714262148e-06
373
+ COC1=CC(=C(C=C1Cl)OC)Cl,0.0006037074787089256
374
+ CCOP(=S)(OCC)OC1=NN(C(=N1)Cl)C(C)C,3.984173701864836e-05
375
+ n(c(nc(n1)NC(C)C)NCC)c1Cl,0.00011591071091933584
376
+ CC(C)(C)C(C(=CC1=CC=C(C=C1)Cl)N2C=NC=N2)O,0.00013506940531624235
377
+ CCCCCCCCc1cc(N(=O)(=O))c(OC(=O)C=CC)c(c1)N(=O)(=O),0.00013721450601021437
378
+ c1cc(OC(F)(F)F)ccc1C(O)(C(C)C)c2cncnc2,3.8746408361649466e-05
379
+ COP(=O)(C(C(Cl)(Cl)Cl)O)OC,7.76890065639098e-05
380
+ C1=CC(=C(C=C1C(F)(F)F)Cl)OC2=CC(=C(C=C2)[N+](=O)[O-])C(=O)O,0.0003456310811217245
381
+ CCCN(CC1CC1)C2=C(C=C(C=C2[N+](=O)[O-])C(F)(F)F)[N+](=O)[O-],1.4397200049120035e-05
382
+ CCOC(=O)C(C)OC(=O)C1=C(C=CC(=C1)OC2=C(C=C(C=C2)C(F)(F)F)Cl)[N+](=O)[O-],0.00010827828420609173
383
+ CCSC(=O)N1CCCCCC1,7.90700043427112e-05
384
+ CCC1=CC=C(C=C1)C(=O)NN(C(=O)C2=CC(=CC(=C2)C)C)C(C)(C)C,0.0001361818336157604
385
+ ClC(Cl)C(Cl)(Cl)SN2C(=O)C1CC=CCC1C2(=O),4.2972436676962935e-05
386
+ COC(=O)C1=CC=CC=C1S(=O)(=O)NC(=O)NC2=NC(=CC(=N2)OC(F)F)OC(F)F,0.00029893004855968545
387
+ CC(C)(C)C(C(N1C=NC=N1)OC2=CC=C(C=C2)Cl)O,8.452667530010769e-05
388
+ CC1=C(C=CC(=C1)Cl)OCC(=O)O,1.9938294964743116e-05
389
+ COC=C(C1=CC=CC=C1OC2=NC=NC(=C2)OC3=CC=CC=C3C#N)C(=O)OC,0.00015431812608561826
390
+ C1=CC=C(C(=C1)C(C2=CC=C(C=C2)Cl)(C3=CN=CN=C3)O)Cl,6.9445287349200036e-06
391
+ [O-]Br(=O)=O,4.769269019610272e-05
392
+ OP(=O)OCC,0.003634746471572125
393
+ COP(N)(=O)SC,7.085706817784122e-07
394
+ CCOP(=O)(NC(C)C)Oc1ccc(SC)c(C)c1,4.944661963970053e-06
395
+ CCOP(=S)(OCC)SCCSCC,1.4577045866248343e-07
396
+ CCOP(=S)(OCC)SCSP(=S)(OCC)OCC,5.2018837831434275e-06
397
+ O=C(OCC(CCCC)CC)CCCCC(=O)OCC(CCCC)CC,0.004047856676081413
398
+ CN(C=Nc1ccc(C)cc1C)C=Nc2ccc(C)cc2C,3.4082463611346175e-05
399
+ [C@@]14([C@@H]5OCC1=CC=C[C@@H]([C@H](O[C@H]2C[C@@H]([C@H]([C@@H](O2)C)O[C@H]3C[C@@H]([C@H]([C@@H](O3)C)O)OC)OC)C(=CC[C@@H]6C[C@H](OC([C@@H]4C=C([C@H]5O)C)=O)C[C@]7(O6)O[C@@H]([C@H](C=C7)C)[C@H](CC)C)C)C)O,2.290718786322191e-06
400
+ O=C(N(S(=O)(=O)Nc1cccc2)C(C)C)c12,0.00014566407168203842
401
+ S=P(OC)(OC)SCN1C(=O)SC(OC)=N1,4.134537164578915e-06
402
+ C(#N)Cl,0.0011387594679715704
403
+ C(#N)Br,0.0011517974649126507
404
+ C1=CC(=CC=C1C(C2=CC=C(C=C2)Cl)C(Cl)(Cl)Cl)Cl,7.052459522690668e-07
405
+ c(cccc1)(c1)C(C)C,0.002753936673434173
406
+ CCCN(CCC)C(=O)SCC,0.00013205276659525523
407
+ NC(CCCC1)C1,0.0006049965454697296
408
+ ClC1CC2C(C1Cl)C3(Cl)C(=C(Cl)C2(Cl)C3(Cl)Cl)Cl,1.4642051620845761e-05
409
+ CC(C)OC(=O)NC1=CC(=CC=C1)Cl,0.004680316153483986
410
+ COC(=O)c1c(Cl)c(Cl)c(C(=O)OC)c(Cl)c1Cl,3.012372657970604e-05
411
+ COC(=O)C1=CC=C(C=C1)C(=O)OC,0.0006437193589585184
412
+ N(C(=S)NC1)C1,2.4471862937206834e-06
413
+ O=C(N(OC)C)Nc(ccc(c1Cl)Cl)c1,1.0036375840596738e-05
414
+ c(cccc1)(c1)C=C,0.0027364609513743235
415
+ COC(=O)c1ccccc1S(=O)(=O)NC(=O)Nc2nc(OC)nc(C)n2,0.0006555420309950702
416
+ C1=CC(=CC=C1N)Cl,9.798423692306362e-05
417
+ FC(F)(F)C(Cl)=CC1C(C)(C)C1C(=O)OC(C(#N))c2cc(Oc3ccccc3)ccc2,2.7787035825324932e-05
418
+ C(Cl)(Br)Br,0.00013683526627950891
419
+ C=C(Cl)Cl,0.00014441434207713882
420
+ C(C=CCl)Cl,4.595842510750173e-05
421
+ COP(=O)(OC)OC(=CCl)c1cc(Cl)c(Cl)cc1Cl,0.00013662627391943267
422
+ Oc(ccc(c1)C(c(ccc(O)c2)c2)(C)C)c1,0.00021902317939829496
423
+ O=C(OCc(cccc1)c1)c(c(ccc2)C(=O)OCCCC)c2,0.001504675539130031
424
+ O=C(NCCCC1)C1,0.0011046536495458942
425
+ c(cccc1)(c1)Cl,0.001066127443097669
426
+ C(Cl)(Cl)Cl,0.000502606685808161
427
+ ClCCl,0.0006190792744080026
428
+ C1C2C3C(C1C4C2O4)C5(C(=C(C3(C5(Cl)Cl)Cl)Cl)Cl)Cl,3.2816209380227e-07
429
+ OCCO,0.016111403264557053
430
+ O=C(C=C(CC1(C)C)C)C1,0.001295160023171075
431
+ C(F)(Cl)(Cl)Cl,0.002540618905481128
432
+ CCc1cccc(CC)c1N(COC)C(=O)CCl,5.189661748967873e-05
433
+ c1ccccc1c2c(C)c(COC(=O)C3C(C)(C)C3C=C(Cl)C(F)(F)F)ccc2,1.1824026617703973e-05
434
+ n1c(Cl)cc(OC)nc1NC(=O)NS(=O)(=O)c2ccccc2C(=O)OCC,0.000301334937881614
435
+ O=C(NC(=O)c(c(F)ccc1)c1F)Nc(ccc(c2)Cl)c2,2.510595466045339e-05
436
+ c(cccc1)(c1)CC,0.0027410163424857604
437
+ O=C(N(SC(Cl)(Cl)Cl)C(=O)c1cccc2)c12,0.00013488106659631142
438
+ Clc1cc(C(F)(F)F)cnc1Oc2ccc(OC(C)C(=O)OC)cc2,2.661507390658951e-06
439
+ C#N,0.0011470716002092733
440
+ ClC(C(OC(C=C2C(=O)OC(C)C(=O)OCC)=CC=C2N(=O)=O)=C1)=CC=C1C(F)(F)F,0.00010827828420609173
441
+ c1c(C(F)(F)F)cccc1N2C(=O)C(Cl)=C(NC)C=N2,0.00016877007995126964
442
+ O=N(=O)C(C(=C1N(=O)=O)N(C(C)C)C(C)C)=CC(=C1)S(=O)(=O)N,0.00012992280391195855
443
+ CN(=CC=C1C(C=C2)=CC=N2C)C=C1,2.0133908207418442e-05
444
+ C1=C(C=C(C=C1[N+](=O)[O-])[N+](=O)[O-])[N+](=O)[O-],6.245761469536105e-05
445
+ C(=C)Cl,2.0800592400871593e-05
446
+ C1CCC(=O)CC1,0.009272184465524753
447
+ CC1(C(C1(C)C)C(=O)OC(C#N)C2=CC(=CC=C2)OC3=CC=CC=C3)C,7.154653735936876e-05
448
+ C1=CC(=O)NNC1(=O),0.004460830164062158
449
+ CCSC(=O)N1CCCCCC1,1.0677920910561878e-05
450
+ C(C(Cl)(Cl)Cl)(O)O,0.0008211924590443927
451
+ ClC2C1OC1C3C2C4(Cl)C(=C(Cl)C3(Cl)C4(Cl)Cl)Cl,6.421500622500273e-08
452
+ Clc1cc(C(F)(F)F)ccc1Oc2cc(OCC)c(N(=O)(=O))cc2,2.764719473193427e-05
453
+ c1cc(Cl)ccc1C(C(#N))(CCCC)Cn2ncnc2,3.407493882440337e-05
454
+ CC1=C(C=C(C=C1[N+](=O)[O-])[N+](=O)[O-])[N+](=O)[O-],8.805487227420694e-06
455
+ CC(C)OC(=O)C(C1=CC=C(C=C1)Br)(C2=CC=C(C=C2)Br)O,6.073132568962677e-05
456
+ C[N+](C)(C)CCCl,0.0010602168942789236
457
+ CCC(=C1C(=O)CC(CC1=O)CC(C)SCC)NOCC=CCl,0.00023894780279715413
458
+ CC1=NC(=NC(=C1)C2CC2)NC3=CC=CC=C3,0.0001580192552676787
459
+ N(c(cccc1)c1)c(cccc2)c2,0.00014773454395291703
460
+ CC1(C(=O)N(C(=O)O1)NC2=CC=CC=C2)C3=CC=C(C=C3)OC4=CC=CC=C4,4.4873074905020916e-05
461
+ CC1(CCCCC1)C(=O)NC2=C(C(=C(C=C2)O)Cl)Cl,0.0009662594125910585
462
+ C1=CC(=C2C(=C1)OC(O2)(F)F)C3=CNC=C3C#N,0.0004432176780816704
463
+ C(F)(F)(F)c1ccccc1C(=O)Nc2cccc(OC(C)C)c2,0.00026909187556769875
464
+ C(CCCCN=C(N)N)CCCNCCCCCCCCN=C(N)N,5.343607459270991e-05
465
+ C1CN(C(=N1)N[N+](=O)[O-])CC2=CN=C(C=C2)Cl,6.64943030028048e-05
466
+ COC(=O)C12CC3=C(C1=NN(CO2)C(=O)N(C4=CC=C(C=C4)OC(F)(F)F)C(=O)OC)C=CC(=C3)Cl,6.8203197610829175e-06
467
+ CC1=CC=CC=C1OCC2=CC=CC=C2C(=NOC)C(=O)OC,0.001180796696935047
468
+ CC1=CC(=CC(=C1)C(=O)N(C(C)(C)C)NC(=O)C2=C(C(=CC=C2)OC)C)C,0.0011154252951100646
469
+ C1=CC(=C(C(=C1)F)C(=O)NC(=O)NC2=CC(=C(C=C2)OC(C(OC(F)(F)F)F)(F)F)Cl)F,7.306609488150114e-05
470
+ CC1=CC=C(C=C1)N(SC(F)(Cl)Cl)S(=O)(=O)N(C)C,5.183483461643306e-05
471
+ CC(C)(C)C1=C(C=CC(=C1)O)O,0.0013536524792656423
472
+ CC(=NOCC1=CC=CC=C1C(=NOC)C(=O)OC)C2=CC(=CC=C2)C(F)(F)F,0.00014692519736711582
473
+ COC(=O)N(C1=CC=CC=C1COC2=NN(C=C2)C3=CC=C(C=C3)Cl)OC,2.3206826561357754e-05
474
+ CC(C)N1C(=NC(C)(C)C)SCN(C1=O)C2=CC=CC=C2,2.8483655881816287e-05
475
+ C1=CC=C(C(=C1)C2=NN=C(N=N2)C3=CC=CC=C3Cl)Cl,5.706818876652661e-05
476
+ CCCC(=C1C(=O)CC(CC1=O)C2CCCSC2)NOCC,8.603044408485128e-05
477
+ C1CC1NC2=NC(=C(C(=N2)N)C#N)N,0.00011566455596376984
478
+ C(C=C1)(=N(C=C1)CC2)C(N2=C3)=CC=C3,3.093839741706742e-06
479
+ C1=CC=C2C(=C1)C(=O)C3=C(C2=O)SC(=C(S3)C#N)C#N,2.0248123201460396e-05
480
+ CCOC1=CC2=C(C=C1)NC(C=C2C)(C)C,5.522147585284529e-05
481
+ C1=CC=C(C=C1)C(CCC2=CC=C(C=C2)Cl)(CN3C=NC=N3)C#N,8.90688528359281e-05
482
+ CC1CN(CC(O1)C)CC(C)CC2=CC=C(C=C2)C(C)(C)C,5.60164796529038e-06
483
+ C[Si](CN1C=NC=N1)(C2=CC=C(C=C2)F)C3=CC=C(C=C3)F,6.341300732121172e-06
484
+ c1cc(Cl)cc(Cl)c1C(OCC=C)Cn2cncc2,5.047450068604895e-05
485
+ CN(=CC=C1C(C=C2)=CC=N2C)C=C1,1.3691057581044662e-05
486
+ CCCCOCCOCCOCC1=CC2=C(C=C1CCC)OCO2,0.00029547465787728373
487
+ Clc1cc(Cl)ccc1C2(Cn3ncnc3)OC(CCC)CO2,0.00028052099059676144
488
+ C(=CC=C1)(C2=C1)NC(=N2)C(=CS3)N=C3,0.00014907004145339663
489
+ CCOC1=CC=C(C=C1)C(C)(C)COCC2=CC(=CC=C2)OC3=CC=CC=C3,6.905930233849184e-05
490
+ CC(C)(C)c2ccc(OC1CCCCC1OS(=O)OCC#C)cc2,5.421255212322886e-05
491
+ CC(COC1=CC=C(C=C1)OC2=CC=CC=C2)OC3=CC=CC=N3,0.0004356352632556389
492
+ CC(C)(C)C(=O)C(N1C=NC=N1)OC2=CC=C(C=C2)Cl,0.0003880867710275085
493
+ CC(C)(C)C(C(N1C=NC=N1)OC2=CC=C(C=C2)Cl)O,0.0003550120362604598
494
+ c1ccccc1c2c(C)c(COC(=O)C3C(C)(C)C3C=C(Cl)C(F)(F)F)ccc2,1.8918442588326134e-05
495
+ CC1(C(C1C(=O)OC(C#N)C2=CC(=C(C=C2)F)OC3=CC=CC=C3)C=C(Cl)Cl)C,1.3815728746284999e-05
496
+ CC1(C(C1C(=O)OC(C#N)C2=CC(=C(C=C2)F)OC3=CC=CC=C3)C=C(Cl)Cl)C,2.763145749256989e-05
497
+ ClC(Cl)=CC1C(C)(C)C1C(=O)OC(C(#N))c2cccc(Oc3ccccc3)c2,0.00012010651237687908
498
+ ClC1CC2C(C1Cl)C3(Cl)C(=C(Cl)C2(Cl)C3(Cl)Cl)Cl,6.100854842019149e-07
499
+ ClC4=C(Cl)C5(Cl)C3C1CC(C2OC12)C3C4(Cl)C5(Cl)Cl,2.625296750418199e-07
500
+ C1C2C3C(C1C4C2O4)C5(C(=C(C3(C5(Cl)Cl)Cl)Cl)Cl)Cl,6.563241876045522e-07
501
+ C1(C(C(C(C(C1Cl)Cl)Cl)Cl)Cl)Cl,1.6160652565775167e-05
502
+ CCC(=O)Nc1ccc(Cl)c(Cl)c1,0.000348496188506355
503
+ C1(=C(C(=C(C(=C1Cl)Cl)Cl)Cl)Cl)[N+](=O)[O-],0.0004740384384225777
504
+ C1C2C(COS(=O)O1)C3(C(=C(C2(C3(Cl)Cl)Cl)Cl)Cl)Cl,7.126617932723372e-06
505
+ O=C(N(SC(Cl)(Cl)Cl)C(=O)c1cccc2)c12,0.00016860133324539056
506
+ CC(C(=O)O)OC1=CC=C(C=C1)OC2=C(C=C(C=N2)C(F)(F)F)Cl,2.764719473193427e-07
507
+ CCCCC(CN1C=NC=N1)(C#N)C2=CC=C(C=C2)Cl,3.393642281292254e-05
508
+ Clc1cc(Cl)cc(Cl)c1OCCN(CCC)C(=O)n2cncc2,1.3539867103283905e-05
509
+ COP(=O)(NC(=O)(C))SC,1.3648831645543501e-05
510
+ CCOP(=O)(OCC)OC(=CCl)C1=C(C=C(C=C1)Cl)Cl,4.171650386740796e-06
511
+ CCOP(=S)(OCC)SCCSCC,8.017375226436709e-07
512
+ O=P(O)(O)CCCl,8.30484305019692e-05
513
+ CCCSP(=O)(OCC)SCCC,1.1141416635499232e-05
514
+ CCOP(=O)(NC(C)C)Oc1ccc(SC)c(C)c1,5.603950225832819e-06
515
+ COP(=S)(OC)Oc1ccc(SC)c(C)c1,2.5868753492304103e-06
516
+ OC(=O)C(N)CCP(C)(=O)O,1.9323475088929698e-05
517
+ OC(=O)CNCP(O)(O)=O,0.0017743806301134177
518
+ CCOC(=O)CC(SP(=S)(OC)OC)C(=O)OCC,0.001089726833059097
519
+ COP(N)(=O)SC,2.0548549771574144e-06
520
+ CC(=CC(=O)OC)OP(=O)(OC)OC,1.5614663580667682e-06
521
+ CCOP(=S)(OCC)SCSCC,6.144925589002914e-07
522
+ CCOP(=S)(OCC)SCSC(C)(C)C,2.080223103650493e-07
523
+ S=P(OC)(OC)SCN1N=Nc2ccccc2C1(=O),8.130482495825867e-06
524
+ CCOP(=S)(OCC)OC1=NC(=C(C=C1Cl)Cl)Cl,2.852364730588813e-06
525
+ CCOP(=S)(OCC)OC1=NC(=NC(=C1)C)C(C)C,1.905728844665908e-05
526
+ CCC1=NC(=CC(=N1)OP(=S)(OC)OC)OCC,1.539557698279251e-06
527
+ S=P(OC)(OC)SCN1C(=O)SC(OC)=N1,5.292207570661027e-06
528
+ CCOP(=S)(OCC)SCN1C2=C(C=C(C=C2)Cl)OC1=O,5.437611333902565e-06
529
+ COP(=S)(OC)SCN2C(=O)c1ccccc1C2(=O),2.8362443195186416e-05
530
+ CCOC(=O)C1=CN2C(=CC(=N2)OP(=S)(OCC)OCC)N=C1C,1.0713392456492962e-05
531
+ CCOP(=S)(OCC)OC1=NN(C=N1)C2=CC=CC=C2,4.1492120354304085e-06
532
+ O=C(Oc(c(c(ccc1)cc2)c1)c2)NC,0.00029817925781592213
533
+ CC1=CC(=CC(=C1SC)C)OC(=O)NC,4.12769581811153e-05
534
+ CNC(=O)ON=C(C)SC,0.00012329688471926276
535
+ CCCOC(=O)NCCCN(C)C,0.003611885866531263
536
+ COC(=O)NC1=NC2=CC=CC=C2N1,0.00039228678402561883
537
+ CC1=C(N=C(N=C1OC(=O)N(C)C)N(C)C)C,5.1618595485714984e-05
538
+ CNC(=O)CCSCCSP(=O)(OC)OC,1.8793291063764634e-06
539
+ N(C(=S)SSC(N(C)C)=S)(C)C,4.990997903448196e-05
540
+ C1=NNC(=N1)N,5.9467202410658145e-05
541
+ CCCCCCCCc1cc(N(=O)(=O))c(OC(=O)C=CC)c(c1)N(=O)(=O),0.00017563456769307506
542
+ C1=CC=C(C=C1)[Sn](C2=CC=CC=C2)C3=CC=CC=C3,8.57111756230557e-07
543
+ O=C(NC(=O)c(c(F)ccc1)c1F)Nc(ccc(c2)Cl)c2,2.253098495168911e-05
544
+ N(C(=S)NC1)C1,1.2235931468603442e-05
545
+ C1=CC(=C(C(=C1)F)C(=O)NC(=O)NC2=CC(=C(C(=C2F)Cl)F)Cl)F,6.559798749656557e-05
546
+ CCCSC1=CC2=C(C=C1)N=C(N2)NC(=O)OC,7.537743365466673e-05
547
+ C1CN(CCN1CCCC(=O)C2=CC=C(C=C2)F)C3=CC=CC=N3,0.00035125670755532855
548
+ CC1(C2C(C3C(C(=O)C(=C(N)O)C(=O)C3(C(=O)C2=C(C4=C(C=CC(=C41)Cl)O)O)O)N(C)C)O)O,0.010507618609493797
549
+ CC1=CC(=C(C=C1NC(=O)C2=CC(=CC(=C2O)I)I)Cl)C(C#N)C3=CC=C(C=C3)Cl,1.5081279803436797e-05
550
+ CN1CC2CC1CN2C3=C(C=C4C(=C3)N(C=C(C4=O)C(=O)O)C5CC5)F,0.00013990757020924605
551
+ C1=CC(=CC=C1C(C#N)C2=C(C=CC(=C2Cl)N3C(=O)NC(=O)C=N3)Cl)Cl,3.679735812631356e-05
552
+ CC1=NC=C(N1C)[N+](=O)[O-],0.00010628650675790856
553
+ CCN1CCN(CC1)C2=C(C=C3C(=C2)N(C=C(C3=O)C(=O)O)C4CC4)F,7.234386376698641e-05
554
+ CC1C=CC=C2COC3C2(C(C=C(C3O)C)C(=O)OC4CC(CC=C(C1OC5CC(C(C(O5)C)OC6CC(C(C(O6)C)NC(=O)C)OC)OC)C)OC7(C4)C=CC(C(O7)C(C)C)C)O,1.110984927911861e-06
555
+ COCC(=O)NC1=C(C=CC(=C1)SC2=CC=CC=C2)NC(=NC(=O)OC)NC(=O)OC,8.959030532555182e-05
556
+ CC1CCC2=C3N1C=C(C(=O)C3=CC(=C2)F)C(=O)O,0.0015311099540615204
557
+ COC(=O)NC1=NC2=C(N1)C=C(C=C2)S(=O)C3=CC=CC=C3,6.3422194381288265e-06
558
+ CC1(C2CC3C(C(=O)C(=C(N)O)C(=O)C3(C(=O)C2=C(C4=C1C=CC=C4O)O)O)N(C)C)O,0.000337507506166939
559
+ C1CN(CCN1CCCC(=O)C2=CC=C(C=C2)F)C3=CC=CC=N3,9.163218457965083e-05
560
+ CC(C)NCC(COC1=CC=CC2=C1C3=CC=CC=C3N2)O,2.3460058312321127e-05
561
+ C1=CC(=CC=C1C(C#N)C2=C(C=CC(=C2Cl)N3C(=O)NC(=O)C=N3)Cl)Cl,5.642261579368184e-05
562
+ CC1C=CC=C2COC3C2(C(C=C(C3O)C)C(=O)OC4CC(CC=C(C1OC5CC(C(C(O5)C)OC6CC(C(C(O6)C)NC(=O)C)OC)OC)C)OC7(C4)C=CC(C(O7)C(C)C)C)O,2.7774623197796e-06
563
+ COP(=S)(OC)Oc1ccc(SC)c(C)c1,1.6167970932690254e-06
564
+ C(C(C(C(C1Cl)Cl)Cl)Cl)(C1Cl)Cl,2.7507493728979266e-05
565
+ C(C(C(C(C1Cl)Cl)Cl)Cl)(C1Cl)Cl,1.3753746864489659e-05
566
+ CC(N(C)C)CN(C(=CC=C3)C1=C3)C(=CC=C2)C(=C2)S1,5.8364575374860975e-05
567
+ C1=C(C=C(C(=C1Cl)N)Cl)[N+](=O)[O-],0.001159340984210937
568
+ c(cccc1)(c1)C=C,0.003840646949297277
data/test/data/batch_prediction.csv ADDED
@@ -0,0 +1,25 @@
1
+ SMILES, batch
2
+ CCCCCCCCOC(=O)C1=CC=C(C(=O)OCCCCCCCC)C=C1
3
+ OCCCCOC(=O)C1=CC=CC=C1C(=O)OCCCCOC(=O)C2=CC=CC=C2C(=O)OCCCCO
4
+ C1([N+](=O)[O-])=CC=C(C)C=C1OP(=O)(OC2=C([N+](=O)[O-])C=CC(C)=C2)OC3=C([N+]([O-])=O)C=CC(C)=C3
5
+ C1=C([N+]([O-])=O)C=CC=C1P(=O)(C2=CC([N+](=O)[O-])=CC=C2)C3=CC([N+](=O)[O-])=CC=C3
6
+ O=C1NC(=O)NC=C1
7
+ O=C2C1=NC3=C(C=C(C)C(C)=C3)N(C[C@H](O)[C@H](O)[C@H](O)CO)C1=NC(N2)=O
8
+ O=C1C2=C(C=CC=C2)C(=O)C3=C1C=CC=C3
9
+ CCC1=C(Br)C(Br)=C(Br)C(Br)=C1Br
10
+ C1CCCCC1C2CCCCC2
11
+ C1=CC(C)=CC=C1SSC2=CC=C(C)C=C2
12
+ CCCCCCCCCCCCCO
13
+ O=C(C1=C(C=CC=C1)C(=O)OCC(CCCC)CC)OCC(CCCC)CC
14
+ CCCCCCCCOC(=O)C1=CC=CC=C1C(=O)OCCCCCCCC
15
+ CCCCCCCCCCCCCC=O
16
+ OCCCCCCCC\C=C/CCCCCCCC
17
+ CCCCCCC(=O)CCCCCC
18
+ N1=C(C2=CC=CC=C2)NC(C3=CC=CC=C3)=C1C4=CC=CC=C4
19
+ C1=CC=CC=C1[Sn](C2=CC=CC=C2)(C3=CC=CC=C3)C4=CC=CC=C4
20
+ C1=CC=CC=C1N(C2=CC=CC=C2)C3=CC=CC=C3
21
+ IC1=CC=C(I)C=C1
22
+ C1=CC(N(C)C)=CC=C1P(=O)(C2=CC=C(N(C)C)C=C2)C3=CC=C(N(C)C)C=C3
23
+ C1=CC=CC=C1C2=CC=C(C3=CC=CC=C3)O2
24
+ CCCCCCCCCCCC(=O)N
25
+ N1=C(CCN)C=CC=C1
data/test/data/batch_prediction_inchi_small.csv ADDED
@@ -0,0 +1,4 @@
1
+ INCHI
2
+ "InChI=1/C11H9N3/c12-10-6-5-8-7-3-1-2-4-9(7)13-11(8)14-10/h1-6H,(H3,12,13,14)/f/h13H,12H2"
3
+ "InChI=1/C8H8O4/c1-4-3-6(10)7(5(2)9)8(11)12-4/h3,7H,1-2H3"
4
+ "InChI=1/C2H4O/c1-2-3/h2H,1H3"
data/test/data/batch_prediction_smiles_small.csv ADDED
@@ -0,0 +1,4 @@
1
+ SMILES
2
+ CCCCCCCCOC(=O)C1=CC=C(C(=O)OCCCCCCCC)C=C1
3
+ OCCCCOC(=O)C1=CC=CC=C1C(=O)OCCCCOC(=O)C2=CC=CC=C2C(=O)OCCCCO
4
+ C1([N+](=O)[O-])=CC=C(C)C=C1OP(=O)(OC2=C([N+](=O)[O-])C=CC(C)=C2)OC3=C([N+]([O-])=O)C=CC(C)=C3
data/test/data/hamster_carcinogenicity.json ADDED
@@ -0,0 +1,3 @@
1
+ { "species": "Hamster",
2
+ "endpoint": "Carcinogenicity",
3
+ "source": "http://www.epa.gov/ncct/dsstox/sdf_cpdbas.html" }