Fig. 3

The distribution of error rate in k-mer hit ratios returned by SeqOthello. A randomly selected set of 150 experiments are extracted from SeqOthello’s result by querying all human transcripts on 2652 human experiments. The error (δ) of a transcript query over an experiment is calculated as the difference between the transcript’s k-mer hit ratio returned by SeqOthello and the k-mer hit ratio obtained by mapping raw k-mers using the same RNA-seq experiments to the transcript sequences. Each bar shows the percentage of transcripts with δ falling in a specific range. The error bar shows the standard deviation of such percentage measured on 150 experiments