Skip to main content

Table 2 The comparison of the performance of the Morgan2 (ECFP4) and b-QAFFP fingerprints for similarity searching for 37 HOM data sets

From: QSAR-derived affinity fingerprints (part 1): fingerprint construction and modeling performance for similarity searching, bioactivity classification and scaffold hopping

FP Morgan2 b-QAFFP
AD No Yes
Cutoff 5 6 7 8 5 6 7 8
AUC 0.57 ± 0.02 0.61 ± 0.02 0.58 ± 0.03 0.61 ± 0.02 0.57 ± 0.02 0.65 ± 0.02 0.59 ± 0.02 0.61 ± 0.02 0.56 ± 0.02
EF5 4.09 ± 0.42 3.44 ± 0.30 3.52 ± 0.47 3.88 ± 0.54 2.33 ± 0.24 4.67 ± 0.53 3.56 ± 0.51 3.39 ± 0.53 1.81 ± 0.21
  1. Model AD was estimated by an ICP. Affinities predicted to lie outside model AD were encoded by zeros. Various affinity cutoffs were used to construct the b-QAFFP fingerprint. Best results are shown in a column in italic. Data shown are averages over all HOM data sets with their standard errors of the mean. The b-QAFFP fingerprint is 402 bits long