Fig. 2From: SamPler – a novel method for selecting parameters for gene functional annotation routinesEvaluation of an annotation algorithm when compared with the Standard of Truth (A or Ø), determined with a manual curation workflow. If the manual annotation is equal to the algorithm’s annotation (A = A) and has a score above the threshold (≥ Threshold), it is classified as CA. When the algorithm’s annotation is below the threshold (≤ Threshold), it is classified as IR. Alternatively, when the annotation is different from the manual curation (A ≠ B), the annotation will be classified either as IA or IR, when above (≥ Threshold) or below (≤ Threshold) the threshold, respectively. Finally, the manual curation may indicate that the entity should not be annotated (Ø), whereas the tool assigned an annotation (A). These cases will be classified as IA or CR, whenever the annotation’s score is above or below the threshold, respectivelyBack to article page