I have pacbio Sequel data at 50 x coverage for a strain of animal. I would like to find structural variants compared to the reference genome sequence. At the moment, I align my reads to the reference genome using NGM-LR and call the variants using Sniffles. The resulting VCF file contains variants that are well supported in a visual inspection in IGV as well as some more questionable variants (especially larger events). I have used the reads corrected by canu for alignment. I think it splits my long reads into very long reads and not-so-long reads and uses the shorter reads to correct the longer reads. However, in IGV, my reads still have a considerable number of mistakes compared to the reference. I was wondering if it is OK to use pilon to polish my long reads with some Illumina data from the same sample? The main purpose of pilon is to correct assemblies but I was wondering if it can be used to correct subreads from PacBio sequencing?
I'd recommend using Canu for correction rather than Pilon, because it has a component that is specifically designed for read-level correction. The newest version of Canu (v1.7) will track all reads through the correction and assembly process:
Ensure that every raw read is either corrected or used as evidence for correcting some other raw read. This serves to rescue short plasmids in high coverage datasets, and it is no longer necessary to set corOutCoverage to achieve the same result.
More details here.