Skip to main content

Table 5 Compare with Dieter's results. Using the same training and testing dataset, both of the two methods have been applied to compute the pearson correlation coefficient. Also, cut off value should be specified as 0.5, 0.6 or 0.7. Here we just showed the result when cut off value is 0.6. The results when cut-off value is 0.5 or 0.7 are detailed in supplemental file (see additional file 4), which also shows the accuracy, sensitivity, specificity and ROC for the two methods. For more info about Dieter's work or the explanation about the datasets used by them, please consult [15]. SVM, cut-off 0.6

From: Demonstration of two novel methods for predicting functional siRNA efficiency

Pearson

All(249)

All human(198)

hE2(139)

Rodent(51)

All(2182)

0.9771

0.9769

0.9743

0.9713

All human(1744)

0.9721

0.9722

0.9689

0.9639

Human E2s(1229)

0.9653

0.9644

0.9606

0.9593

Rodent(438)

0.9057

0.9077

0.895

0.8806

Random all (1091)

0.9660

0.9673

0.9651

0.9510

Random all (727)

0.9343

0.9369

0.9387

0.9125

Random all (545)

0.9249

0.9252

0.9206

0.9154

Random all (218)

0.8502

0.8645

0.8570

0.7713

All-19

0.9436

   

All human-19

0.9387

   

Rodent-19

0.8487

   

SeqSta, cut-off 0.6

    

Pearson

All(249)

All human(198)

hE2(139)

Rodent(51)

All(2182)

0.8562

0.8557

0.8452

0.8520

All human(1744)

0.8104

0.8106

0.8007

0.8008

Human E2s(1229)

0.7294

0.7294

0.7353

0.7257

Rodent(438)

0.7761

0.7688

0.7608

0.7912

Random all(1091)

0.8632

0.8619

0.8472

0.8644

Random all(727)

0.7953

0.8023

0.7830

0.7523

Random all(545)

0.7812

0.7785

0.7679

0.7748

Random all(218)

0.7017

0.6941

0.6681

0.7292

All-19

0.8224

   

All human-19

0.7809

   

Rodent-19

0.7097

  Â