search

jts / nanopolish

Signal-level algorithms for MinION data
MIT License
559 stars 159 forks source link

Why are megabase-sized contigs needed? #252

Closed btemperton closed 6 years ago

btemperton commented 6 years ago

We have a viral metagenome sequenced by 1D ligation and assembled in canu. Canu + PILON gets us within 99% ANI of contigs we see from SPAdes hybrid assemblies. I would like to try CANU + nanopolish + PILON to see if (a) we can get closer (see http://albertsenlab.org/what-is-a-good-genome-assembly/) and (b) if we can detect any methylation signal in the viral genomes. However, even if we manage to sequence complete viral genomes, most of the contigs will be 10-300kb. What effect will this have on nanopolish?

nickloman commented 6 years ago

Speaking from my own experience, it should not matter how long the contigs are. Just bear in mind that the very flanking tips of contigs (e.g. first and last 20-30bp) may not get corrected.

jts commented 6 years ago

@nickloman is right, polishing small contigs should work fine except for the very ends