Discrepancy in outputs at different levels.

[VaibhavWagh]

Hi .. I have Nanopore cdna reads and tried variant calling for this data. One set was analysis with reads without error correction where i see a lot of variants (~500). Further after doing error correction i do not see any variants reported in the vcf file and the consensus generated also shows a 99% match with reference but if i upload the alignment bam file in IGV i still see the variant reported at the exact position in first set of analysis with same depth. I am really confused .

Thanks in advance.

[ksahlin]

Hi @VaibhavWagh,

What would you estimate is the relative depth of those variants? The --T is the parameter controlling the minimum abundance on variations that isONcorrect should ignore to correct to major variant. Default is to not correct variations with more than a frequency of 0.1. You could instead set --T 0.05 which would leave more variants but correct fewer errors. It would be interesting to see an example e.g. in IGV of some of the variation sites you see in the original data and how they are removed in the corrected reads.

Also, have you checked that reads are mapping to the same genomic location after correction?