Figure 3From: BayesMotif: de novo protein sorting motif discovery from impure datasetsFalse positive removal procedure. This example shows BayesMotif uses sliding windows to remove false motifs from the dataset. It works by averaging the score before consecutive 3 window score which has at least one less than 0.85 (marked as red) calculate the score for each sequence the sequences have score less than 0.85 will be removed from the training set.Back to article page