Fig. 5

Poly(A) selection underestimates expression of genes w/ short-tailed mRNAs. Two biological replicates were used in these subfigures to identify systemic technical biases (Selected-1 & Unselected-1; Selected-2 & Unselected-2). A Scatter plot comparing the differences in Nanopolish-called mean tail length between techniques (selected mean tail length - unselected mean tail length). Each axis is from one biological sample. Genes are colored based on decile binning of the average difference in tail lengths between techniques. Dashed line indicates the diagonal, where the difference in tail lengths is equal. B Cumulative distribution functions of the fold-change RPM between techniques. Genes were broken up into deciles of average difference in tail lengths between techniques and replicates as shown in Fig. 5A. C Scatter plot comparing fold-change RPM between techniques for two biological samples, to assess reproducibility of technical biases. Genes are again colored by decile as per Fig. 5A. Dashed line indicates the diagonal where the fold-change in RPM due to technique is equal between biological replicates (Spearman correlation coefficient: 0.6422)