Picking parameters for QC of long-read data?

Question

[Gio]

Hello all,

What is a good heuristic to pick good defaults for my initial QC run? My data is a bunch of reads produced by Nanopore sequencing of metagenomic samples.

As I understand it, my first pass QC should include excluding short fragments and low quality reads, I am doing this filtering with samtools, and generating reports with FastQC. However, I have not found good data on how to pick those values? Currently I'm gong with Q10 and min 100 bp.

[GenoMax]

Use a program meant for QC of long reads like PycoQC (LINK). Since nanopore Q scores are not the same as Illumina, any Q score filtering may not too severe. It may depend on what you are finally trying to do with the data.

[Gio]

I am using chopper which is meant for long reads. My goal is to assemble MAGs and map reads back onto them for further analysis.