TeloSearchLR: an algorithm to detect novel telomere repeat motifs using long sequencing reads

Telomeres are eukaryotic chromosome end structures that guard against sequence loss and aberrant chromosome fusions. Telomeric repeat motifs, the minimal repeating unit of a telomere, vary from species to species, with some evolutionary clades experiencing a rapid sequence divergence. To explore the full scope of this evolutionary divergence, many bioinformatic tools have been developed to infer novel telomeric repeat motifs using repetitive sequence search on short sequencing reads. However, novel telomeric motifs remain unidentified in up to half of the sequencing libraries assayed with these tools. A possible reason may be that short reads, derived from extensively sheared DNA, preserve little to no positional context of the repetitive sequences assayed. On the other hand, if a sequencing read is sufficiently long, telomeric sequences must appear at either end rather than in the middle. The TeloSearchLR algorithm relies on this to help identify novel telomeric repeat motifs on long reads, in many cases where short-read search tools have failed. In addition, we demonstrate that TeloSearchLR can reveal unusually long telomeric motifs not maintained by telomerase, and it can also be used to anchor terminal scaffolds in new genome assemblies.