unexpected number_of_reads in nanostats file

[asulit08]

Hi! I have noticed that the number of reads declared in the nanostats file is almost always a little less than if I do wc (divided by 4) of the fastq file I input. The wc number also corresponds to the number of sequences in a minimap log file. However, the number I get from a NanoStats file is different. Is there a quality cutoff or some reason some reads aren't counted in nanocomp? Currently, I am using version 1.23.1 (maybe an old version as I've only just gotten back to reviewing these files), and the following parameters:

NanoComp -t {threads} -o {output} --raw --tsv_stats --verbose --fastq {input} -n {params} 1> {log} 2>&1

Would appreciate some insight

[wdecoster]

Hi, some very short, high-quality artefact reads could be filtered out, but that should be noted in the log file NanoComp creates. Could you also be more specific about how many reads you get?

[asulit08]

Hi,

Thank you for the reply. As an example, one of my samples would have a 1686735 read count on NanoStats but a 1696064 count on minimap logs (and my own wc on the fastqs)

[wdecoster]

And do you have a log file?

[asulit08]

NanoComp_20241107_1120.log

Please see attached. This would be sample CCR351 on the log file

[wdecoster]

This is probably the relevant line:

2024-11-07 11:25:48,634 Removed 179003 artefactual reads with very short length and very high quality.

[asulit08]

Ahh, thank you! could I ask if this number is for all the samples (added together)?

[wdecoster]

Hmm, I would think so :-)