Data recovery from aborted HiSeq2500 run

[matth431]

Hi, we were running a 1x126bp high-output run on our HiSeq, but a comms issue caused the run to fail at cycle 109. Basecalling was complete however for cycles 1-107, although for some reason Q30% drops to about 50% after cycle 99. However, since the user only needs ~100bp of data, I would like to process the basecalls as they stand.

What would I need to do in order for bcl2fastq to correctly process the data assuming I treat it as a 1x101bp run? Modify the runinfo/runparameters.xml files to alter read length and set usebasesmask to y101nnnnn (or something like that)? Or would I also need to copy the various RTAcomplete and BasecallingComplete.txt files from a successful run?

Thanks,

Matt

[Genetic Librarian]

1) Fake the ....complete.txt files (just copy an old one and put a reasonable time stamp in there)
2) Correct the read length in the xml files (you might need to go one less than you think, because last called cycle might be treated just for the phasing/prephasing calc.)
3) Correct read length in your samplesheet (I did that, don't know is neccessary)

That are the things I remember from the last time we needed it. Hope my list is complete.

[matth431]

Thanks - it seems to have generated a makefile OK just running bcl2fastq with --use-bases-mask set to y101nnnnnn (since it had processed up to base 107), will know for sure once the make script has finished.