Calling Homopolymer Stretches from Raw Nanopore Reads by Analyzing k-mer Dwell Times

Oxford Nanopore Technologies’ (ONT) MinION device is capable of reading single molecule DNA strands tens of thousands of bases long, by passing a strand through a nanopore and recording the changes in electric current. The error rate of the platform is higher than most mature next-generation sequencing (NGS) platforms, with many of the deletions accumulating in stretches of identical bases (homopolymers). However, the mean time each 5-base long subsequence (k-mer) of the molecule spends inside of the pore (dwell time) can also be used to infer the length of the true sequence. We developed a method called NanoTimer, which estimates the homopolymer length from the dwell times. It relies on the redundancy of having multiple reads covering a reference sequence, and the depth of coverage determines its accuracy.