[GenoMax]

What exactly are you trying to do?

Based on the file names it appears that your sequence provider has already done the basecalling (e.g. xed/not_demultiplexed_NoIndex_L003_R1_001.fastq.gz).

If you have some kind of internal barcode/index on these reads then CASAVA can't be used for demultiplexing those.

[gabyslysa]

I'll explain what happened. I made all of my sequencing in another laboratory. We ran two samples per lane of mRNA libraries. But few indexed reads were generated, approximately one million per sample (very little indeed). I'm questioning the preparation of these libraries, because I believe that it was not done properly by this laboratory. But I would like to see the reads before making demultiplexing, to see if I can recover some reads. So I want to do it again converting blc to fastq, without demultiplexing. I want to see the reads generated containing all the information, the sequences containing the indexes. What do you suggest? Thank you. I thought about doing the PICARD.

[GenoMax]

In case of illumina multiplex libraries the index read happens as a separate read so you will not be able to see the tags unless you actually provide a samplesheet and de-multiplex the sample.

If as you say the index read was bad (i.e. N's or whatever) most of the reads will end up in the "undetermined" reads file. You can then parse the reads to determine if the library indeed is bad (look for the tags at the end of the read ID's).

Were there other samples on the flowcell that demultiplexed properly? What was the cluster concentration (k/mm^2)? Generally if you overload a sample the regular reads are generally fine but it is the index reads that suffer since the basecaller has a hard time distinguishing the clusters. You will then not be able to demultiplex the reads ruining the run (as you found).

[sklages]

Good Morning :-)

Be aware as GenoMax already mentioned that the index itself is simply a separate read. Without demultiplexing you loose the information about the indices.

Have a look at the undetermined_indices to see if you have quality problems with the index or if you have another index present in the data.
PM me if you want a simple perl script doing the job for you.

[arkilis]

I just met this problem recently and I solved this by checking CASAVA 1.8.2 manual.

Solution for people who needs this:

There might be many reasons lead to such an error.

If the error msg is missing bcf file or stats files. Try to use parameters: --ignore-missing-bcl or --ignore-missing-stats

The reason also might be file/path permission and mounted drive. However, they are all give the same error msg. (It is quite misleading)