Figure 2

ROC curves of CNS, NCNS and Ensembl datasets homology search results. The receiver operating characteristic (ROC) curve plots the true positive rate against the false positive rate for the different possible cut points of specific variables of the BLASTn results. The true positive rate is measured by the BLASTn search of lncRNAs against the phastCons elements while the false positive rate accounts for the BLASTn search of shuffled sequences against the phastCons elements. The ROC curves were used to determine the ideal score for a cut point which may separate the alignments with biological significance from the random occurring alignments. ROC curves for query coverage (QCoverage), percentage identity (PIdentity), query alignment length (QAlength) and e-value (1/EValue) at word size 11 for A) CNS dataset B) NCNS dataset, C) Ensembl dataset. The cut-off for a parameter is defined as the point of steep incline in the true positive rate as compared to the false positive rate. The significant cut-off defined in the present analysis are indicated by arrows. ROC curves for the e-value parameter in the plots show the reciprocal of the e-value (1/e-value) because plotting the e-value produced curves sensibly skewed below the diagonal line.