Fig. 2: Basecalling performance was evaluated on six RNA001/002 test datasets and a HEK293T RNA004 sample.

a–d Heatmaps illustrating the median read accuracy, mismatch error rate, insertion error rate, and deletion error rate for five basecallers applied to the RNA001/002 test datasets. In these heatmaps, darker colors signify superior performance. The 'Avg' row represents the average value across all six test datasets. e, f Bar plots depicting the median read length and the number of unaligned reads tested on the RNA001/002 datasets. g Summary metrics for the HEK293T RNA004 sample, including median read accuracy, mismatch rate, insertion rate, deletion rate, median read length, and the proportion of unaligned reads. Statistics were derived from n = 100,000 reads for each RNA001/002 dataset and n = 13,027,670 reads for the HEK293T RNA004 dataset. Source data are provided as a Source Data file.