Development of Natural Compound Molecular Fingerprint (NC-MFP) with the Dictionary of Natural Products (DNP) for natural product-based drug development

Table 4 Binary classification result of task I^a

Performance of each molecular fingerprint obtained by averaging ten external validation tasks^a,b
Molecular fingerprint	Natural compound classification			Synthetic compound classification
Molecular fingerprint	Avg. TP	Avg. FN	Avg. Sensitivity^c (%)	Avg. TN	Avg. FP	Avg. Specificity^d (%)
NC-MFP	183	14	92.65	113	87	56.50
MACCS	169	30	84.60	146	53	73.35
PubChem FP	165	34	82.60	154	46	77.00
GraphFP	161	38	80.75	143	56	71.80
APFP	153	46	76.55	141	58	70.70

^aThe result of performance about the binary classification task I. The external validation data set was randomly selected 10 times by a proportion of 20% from the data set. “NC-MFP” stands for Natural Compound Molecular Fingerprints and “APFP” for AtomPairs2DFingerprint and “GraphFP” for GraphOnlyFingerprint. “MACCS” reports Molecular Access System keys fingerprints and “PubChemFP” stands for PubChem fingerprint
^bThe performance index consist of Sensitivity and specificity. “TP” stands for True positive and “FN” stands for False negative and “TN” standards for True negative and “FP” standard for False negative
^cThe sensitivity is the proportion of positive class that was correctly identified
^dThe specificity is the proportion of negative class that was correctly identified

ISSN: 1758-2946