Table 5 Summary of the top-twenty important features ranked by SHAP values along with their corresponding SMARTS patterns and substructure description

PubChemFP821

CC1C(N)CCCC1

2-methylcyclohexan-1-amine

PubChemFP419

C≡N

Cyano group

PubChemFP800

CC1CC(N)CCC1

3-methylcyclohexan-1-amine

PubChemFP712

C–C(C)-C(C)-C

2,3-dimethylbutane

PubChemFP516

[#1]-C = C-[#1]

Ethene

PubChemFP259

 ≥ 3 aromatic rings

Greater than 3 cyclic rings

PubChemFP564

C = C–C = C

Buta-1,3-diene

PubChemFP818

CC1C(C)CCCC1

1,2-dimethylcyclohexane

PubChemFP299

N–H

Imidogen

PubChemFP797

CC1CC(C)CCC1

1,3-dimethylcyclohexane

PubChemFP403

N(:C)(:C)(:C)

N,N-dimethylmethanamine

PubChemFP338

C(~ c)(~ c)(~ H)(~ N)

Propan-2-amine

PubChemFP186

 ≥ 2 saturated or aromatic carbon-only ring size 6

Greater than 2 saturated or aromatic carbon-only six-member cyclic ring

PubChemFP185

 ≥ 2 any ring size 6

Greater than 2 six-member cyclic ring

PubChemFP777

cc1ccc(o)cc1

4-methylphenol

PubChemFP15

 ≥ 2 N

Greater than 2 nitrogen atoms

PubChemFP641

O-C–C-C = C

But-3-en-1-ol

PubChemFP422

N = N

Diazene

PubChemFP193

 ≥ 3 saturated or aromatic carbon-only ring size 6

Greater than 3 saturated or aromatic carbon-only six-member cyclic ring

PubChemFP495

C-N–C:C

N-methylethanamine