The selected threshold is required for prediction. The user can change the threshold of the model for less or more stringent results in terms of recall and specificity. The lower threshold will increase the recall but decrease the specificity and higher threshold will do the opposite to prediction outcomes. Figures below represents the trend of variation in the prediction results with the change in the threshold for the prediction models implementted on the webservers: (a) all transcript (THCA-EL-SVC-L1) on training dataset (b) all transcript (THCA-EL-SVC-L1) on validation dataset (c) protein coding (THCA-EL-PC) on training dataset (d) protein coding (THCA-EL-PC) on validation dataset (e) Cancer Normal (THCA-CN-F) on training dataset (f) Cancer Normal (THCA-CN-F) on validation dataset.
