PacBio and ONT reads difference !

Question

[BioGeek]

There are many mapper that uses ONT and PacBio flag separately. I always wonder why? Why not one flag for all kind of long reads? Is this to compensate error rate ? What if I corrected all reads before mapping ? Sorry for my ignorance, but I am curious about it.

Answer 1

[h.mon]

Because indeed current PacBio and NanoPore sequencing have not only different error rates, but different error profiles. See, for example, Comprehensive comparison of Pacific Biosciences and Oxford Nanopore Technologies and their applications to transcriptome analysis for a comparison between PacBio and NanoPore sequencing in regard to RNAseq.

Given these different error rates and profiles, optimal settings for mapping will be different. For example, from minimap2 GitHub repo:

The difference between map-pb and map-ont is that map-pb uses homopolymer-compressed (HPC) minimizers as seeds, while map-ont uses ordinary minimizers as seeds. Empirical evaluation suggests HPC minimizers improve performance and sensitivity when aligning PacBio reads, but hurt when aligning Nanopore reads.

PacBio CCS reads have low error rates, and I think they are more distant from PacBio subreads and PacBio longreads than these are to Nanopore reads. Likewise, if you correct your reads with Illumina reads, the error rate will be low and using the settings for noisy reads is not appropriate.