Per tile quality sequence fail but blue graph

Question

[UserA]

Good morning,

For a project, we sent to illumina NovaSeq 6000 paired-end sequencing a large number of samples containing 2 or 3 bacterial strains each. When I do a MultiQC of my data, on half of the samples, the R1 gives a fail for the per tile quality sequence. When I look at some FastQC of these samples, it does give me a fail but the graph is totally blue, as blue as the graphs of the samples that pass this parameter. What could be the reason for this ?

Answer 1

[h.mon]

In my experience, FastQC is overly rigorous regarding the "per tile" evaluation, and flag a sample as failed even if very few tiles presented quality issues. Unless other metrics - e.g. per base sequence quality, number of sequences per sample, adapter content - are problematic too, this shouldn't represent a problem.