tailfindr: Alignment-free poly(A) length measurement for Oxford Nanopore RNA and DNA sequencing

Maximilian Krause; Adnan M. Niazi; Kornel Labun; Yamila N. Torres Cleuren; Florian S. Müller; Eivind Valen

To identify the region corresponding to the expected poly(A) tail, we apply thresholding to normalized raw data, refine the boundaries of possible poly(A) stretches based on raw signal slope, and normalise by the read-specific nucleotide translocation rate (Fig. 1C , for more details see Materials and Methods). tailfindr provides the user with a tabular output containing the unique read-ID, the estimated poly(A) tail length and all factors extracted from the raw data that are needed to calculate the poly(A) tail estimate (Fig. S1A ). This allows the user for custom filtering of the acquired poly(A) measurements. Optionally, tailfindr allows the user to generate read-specific plots displaying the raw data and all signal derivatives generated in the process to estimate poly(A) tail length ( While this study was in progress, another tool estimating poly(A) tail lengths from ONT RNA data has been developed [32] . Instead of estimating poly(A) tails from basecalled data directly, this tool requires read alignment information for the definition of the poly(A) tail segment. To compare whether our algorithm results in similar performance, we measured poly(A) tail lengths from Nanopolish and tailfindr on different barcoded eGFP molecules. Our analysis shows that both tools match in precision and length estimation, as exemplified in Figure 1E for 40 and 100 nt poly(A) tail length.