How to interpret multiqc report

Question

[gt]

I am having trouble interpreting some of my multiqc plots for my quality assessment of my fastq files (E. coli samples). I have the following chart which shows good results for the per sequence quality scores plot (all samples pass). See figure below.

However, when I look at the sequence quality histograms showing the mean quality scores for each bp, I have the majority of my samples "failing". See figure below.

Can anyone help explain what is going on here?

Answer 1

[GenoMax]

when I look at the sequence quality histograms showing the mean quality scores for each bp, I have the majority of my samples "failing"

I am not sure why MultiQC is labeling your samples in Red in second plot.

Graph 1 shows mean sequence quality across the entire read in your dataset and that value is > 35 (really good).
Graph 2 is mean quality score at each base position in these reads. Again > Q35 across almost entire read (save a few bases at the beginning).

That dip at the beginning is likely because of this considering you have a RNAseq dataset.

[gt]

Yeah, strange. Maybe just a bug in the program?

[Phil Ewels]

I suspect that if you look at the underlying FastQC report the box plots will show a wide range in read qualities. From the FastQC help:

This module will raise a failure if the lower quartile for any base is less than 5 or if the median for any base is less than 20.