Demux summary qzv's Q scores doesn't match the csv when downloaded

I imported and demultiplexed my sequences, and now i’m using the summary in order to check out the q-scores in order to filter via dada2. I wanted something more precise than just hovering my mouse over the barplot it gives you, so I downloaded the csv so I could see all the numbers laid out at once. Once I started comparing the csv values to what the qzv showed, I noticed that the numbers were often off by a few points. Is this just due to rounding error? If so, which number is more accurate, the csv or the barplot?

Running qiime2-2019.7

commands:
qiime tools import
–type ‘SampleData[PairedEndSequencesWithQuality]’
–input-path ./
–input-format CasavaOneEightSingleLanePerSampleDirFmt
–output-path sequences.qza

#visualize demux sequences
qiime demux summarize
–i-data sequences.qza
–o-visualization sequences.qzv

1 Like

Hi @KingReyRoi,

I looked through the seven number summary csv downloaded from the Moving Pictures tutorial and compared it to the q-plots and I believe what is happening is all the values are shifted by 1 spot.
For example hovering over position 136 in the visualizer actually shows the values in position 135 in the downloaded version. If my quick look-over is correct this is a bug you’ve found, but let’s see what the developers have to say. A 1bp change shouldn’t really be an issue though, you can make your trim/truncating decisions based on either version and you should be fine.

1 Like

Okay, I think that’s what it is. Now that I look at the csv more closely, it looks like it starts at bp ‘0’ instead of ‘1’ like on the graph. So the graph lists sequence lengths of 301, but the csv lists them from 0-300. Thanks for your help!

3 Likes

Thanks for reporting, @KingReyRoi! I have opened an issue up here:

1 Like

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.