Skip to main content
. 2018 Jul 31;20(6):2009–2027. doi: 10.1093/bib/bby065

Figure 6.

Figure 6.

Performances of different tools on wheat data set. Although CPC had inferior performances on human and mouse, it achieved the best accuracy on wheat. CPC obtained an accuracy of 0.9595, but his alignment-free successor CPC2 only had an accuracy of 0.7870. The accuracies of CPAT (re-trained model) and PLEK (re-trained model) were 0.8743 and 0.8773, respectively, while LncFinder obtained an accuracy of 0.9283. When default models were used, CPAT (model for human), CNCI (default model for plants) and PLEK (default model for plants) had accuracies of 0.7145, 0.6158 and 0.5275, respectively. LncFinder (model for human) had an accuracy of 0.8190. Although CNCI and PLEK provide default models for plants, the performances were substandard. LncFinder has the best performance among alignment-free tools. Even using the model for human, LncFinder still outperformed CPC2, CNCI and the default models of CPAT and PLEK.