Skip to main content

Table 3 The performance of models generated by ClinProTools for the differentiation of O4:K8 from O3:K6 as well as its three serovariants (O4:K68, O1:K25, and O1:KUT)

From: MALDI-TOF mass spectrometry-based serotyping of V. parahaemolyticus isolated from the Zhejiang province of China

Target serotype

Control serotype

Classification algorithm

Peaks used in the model

Cross validation (%)

Recognition capability (%)

Overall

Overall

O4:K8

(n = 36)

O3:K6(n = 63)

O4:K68(n = 4)

O1:K25(n = 2)

O1:KUT(n = 5)

GA

4798.1, 8357.42, 8090.06, 6395.04, 8173.94

99.1

99.82

QC

3512.86,3557.98, 3740.45, 4383.85, 4397.89, 4486.14, 4734.61, 4785.89, 7051.75, 7405.26, 8090.06, 9466.35, 9569.91

94.0

93.8

SNN

9466.35, 4486.14

92.7

93.11

  1. GA genetic algorithm, QC QuickClassifier Algorithm, SNN Supervised Neural Network Algorithm