Fig. 3: Assessment of intrinsic bias and the effects of using a negative control on m6A detection.
From: Systematic comparison of tools used for m6A mapping from nanopore direct RNA sequencing
a, b Precision, recall and F1 scores for single-mode tools (a) and compare-mode tools (b) using miCLIP results as validation set. c Counts and percentages of recurrent m6A sites identified amongst samples. One-sided Hypergeometric tests were applied to calculate significance. WT&KO, both in WT and KO; WT&Random, both in WT and Random; KO|Random, the total number in KO or Random. d, e The m6A site counts and precision and F1 scores before and after calibration with the results from KO (d) or random (e) samples. miCLIP results were used as a validation set. WT-KO, precision and F1 score for sites calibrated using KO results. WT-random, precision and F1 score for sites calibrated with a randomly selected dataset.
