Using readmers and hapmers in assessing phase switching after read error correction of Oxford Nanopore Sequences

AO_SCPLOWBSTRACTC_SCPLOWMethods for sequence error correction can improve sequence accuracy; however, there can be unintended errors added during error correction. One such example is phase switching, whereby sequences derived from genomes containing more than one parental copy have contributions from more than one parental haplotype. Such switches are mistakes that can confound downstream analyses especially de novo genome assembly. While DNA sequences do not possess words such as linguistic languages, one can partition a DNA sequence into word-like objects called k-mers. K-mers include pieces of DNA sequence of k length that can describe various properties of DNA sequences. With regard to phase switching, there are k-mers present in one parental haplotype not found in other(s). These so-called hapmers can represent inaccuracies in that parental haplotypes assembly but also correct, unique DNA variation. Here we investigated the effect of DNA sequence error correction on phase switching at the sequence/read level. Using several error-correction methods, we find all methods tested are similar to raw, presumably, phase-switch-free Oxford Nanopore Technologies (ONT) sequences in the percentage of readmers (k-mers from the ONT sequences) matching one parental haplotypes hapmers. This work demonstrates an efficient method to assess if an error-correction method has introduced phase switching implemented in the Julia programming language.