disproportinate fastq sizes from bcl2fastq

Question

[das2000sidd]

Hi Everyone

While demultiplexing fastq files and for some reason for one of our sequencing runs, one set of fastq run on 4 different lanes of Illumina nextseq for a particular sample is generating fastq in the range of 2 .5 to 3 gb while the remaining 15 samples have fastq in the range of 100 to 200 Mb. This particular problem is happened in only one of our sequencing runs and the bcl2fastq log seems okay. Can anyone give some suggestions as to what problem this could be? I do not have much experience with how bcl2fastq works and hence was looking for some advice.

Thank you again

Answer 1

[GenoMax]

Have you looked at the demultiplexing report produced by bcl2fastq? You should be able to find that report in FCID/Unaligned/Reports/html directory, if you are using a local install of bcl2fastq to process the data. Look at the index.html file. If you are using BaseSpace then you probably know where to find the report.

• A simple explanation could be that the pool you are sequencing is unbalanced with some samples constituting a larger fraction, resulting in correspondingly larger files.
• If that is not the case then check the "undetermined" pool. If you see a large number of reads landing in the "undetermined" pool you will need to start chasing down issues with index sequences. Perhaps there are N's in indexes which is causing the reads to not demultiplex properly.

[Madelaine Gogol]

This one liner will show you some of the most common barcodes in your undetermined fastq file. Might help?

zcat undetermined.fastq.gz | grep "^@" | sed "s/.*://g" | head -n 1000000 | sort | uniq -c | sort -n

Answer 2

[swbarnes2]

The simplest explanation is that bcl2fastq ran fine, but whoever loaded the instrument didn't normalize all the samples equally.