Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry for differential identification of adult Schistosoma worms

Table 2 Tenfold cross-validation results showing scores (%) obtained with four different classification algorithms [linear support vector machine (SVM), partial least squares-discriminant analysis (PLS-DA), Random Forest (RF), and k-nearest neighbors (KNN)]

Schistosoma mansoni vs Schistosoma japonicum^a
Actual group	Predicted group		% Correct
Actual group	S. japonicum	S. mansoni	% Correct
SVM
S. japonicum	22 (TP)	0 (FN)	100 (Sensitivity)
S. mansoni	2 (FP)	38 (TN)	95 (Specificity)
	91.7% (PPV)	100% (NPV)	96.8 (Accuracy)
			95.7 (F1 score)
PLS-DA
S. japonicum	21 (TP)	1 (FN)	95.5 (Sensitivity)
S. mansoni	1 (FP)	39 (TN)	97.5 (Specificity)
	95.5% (PPV)	97.5% (NPV)	96.8 (Accuracy)
			95.5 (F1 score)
RF
S. japonicum	22 (TP)	0 (FN)	100 (Sensitivity)
S. mansoni	0 (FP)	40 (TN)	100 (Specificity)
	100% (PPV)	100% (NPV)	100 (Accuracy)
			100 (F1 score)
KNN
S. japonicum	21 (TP)	1 (FN)	95.5 (Sensitivity)
S. mansoni	1 (FP)	39 (TN)	97.5 (Specificity)
	95.5% (PPV)	97.5% (NPV)	96.8 (Accuracy)
			95.5 (F1 score)

S. mansoni (ethanol vs RNAlater)^b
Actual group	Predicted group		% Correct
Actual group	RNAlater	Ethanol	% Correct
SVM
RNAlater	20 (TP)	1 (FN)	95.24 (Sensitivity)
Ethanol	0 (FP)	19 (TN)	100 (Specificity)
	100% (PPV)	95% (NPV)	97.5 (Accuracy)
			97.6 (F1 score)
PLS-DA
RNAlater	21 (TP)	0 (FN)	100 (Sensitivity)
Ethanol	0 (FP)	19 (TN)	100 (Specificity)
	100% (PPV)	100% (NPV)	100 (Accuracy)
			100 (F1 score)
RF
RNAlater	21 (TP)	0 (FN)	100 (Sensitivity)
Ethanol	0 (FP)	19 (TN)	100 (Specificity)
	100% (PPV)	100% (NPV)	100 (Accuracy)
			100 (F1 score)
KNN
RNAlater	21 (TP)	0 (FN)	100 (Sensitivity)
Ethanol	0 (FP)	19 (TN)	100 (Specificity)
	100% (PPV)	100% (NPV)	100 (Accuracy)
			100 (F1 score)

S. japonicum (ethanol vs RNAlater)^c
Actual group	Predicted group		% Correct
Actual group	RNAlater	Ethanol	% Correct
SVM
RNAlater	11 (TP)	0 (FN)	100 (Sensitivity)
Ethanol	0 (FP)	11 (TN)	100 (Specificity)
	100% (PPV)	100% (NPV)	100 (Accuracy)
			100 (F1 score)
PLS-DA
RNAlater	11 (TP)	0 (FN)	100 (Sensitivity)
Ethanol	0 (FP)	11 (TN)	100 (Specificity)
	100% (PPV)	100% (NPV)	100 (Accuracy)
			100 (F1 score)
RF
RNAlater	11 (TP)	0 (FN)	100 (Sensitivity)
Ethanol	0 (FP)	11 (TN)	100 (Specificity)
	100% (PPV)	100% (NPV)	100 (Accuracy)
			100 (F1 score)
KNN
RNAlater	11 (TP)	0 (FN)	100 (Sensitivity)
Ethanol	0 (FP)	11 (TN)	100 (Specificity)
	100% (PPV)	100% (NPV)	100 (Accuracy)
			100 (F1 score)

TP True positive, TN true negative, FP false positive, FN false negative, PPV positive predictive values, NPV negative predictive values
^aDiscrimination of adult S. japonicum from adult S. mansoni; S. japonicum considered a positive category
^bDiscrimination of S. mansoni isolates based on the effect of the storage medium; RNAlater considered a positive category
^cDiscrimination of S. japonicum isolates based on the effect of the storage medium; RNAlater considered a positive category