Suggested Options For Casava 1.8 (Bcl > Fastq)

Question

[Ian]

I am new to the initial processing of Illumina reads (GA2x) after spending the last couple of years handling SOLiD data. It seems clear from a trawl of the internet that CASAVA 1.8 is not the easiest piece of software to use...

My concern is the conversion of BCL formatted output (direct from the sequencer) to the FASTQ format. Can anyone recommend sensible parameters for the filtering step, such as the number of mismatches? Or do people run it using default parameters?

Are there any other pitfalls/suggestions regarding the BCL > FASTQ step.

Thanks.

[jingtao09]

here is the info from CASAVA1.8.2 (I highly recommend you to upgrade to 1.8.2)

---

1 Enter: /path-to-CASAVA/bin/configureBclToFastq.pl --input-dir <BaseCalls_dir> --output-dir <Unaligned> --sample-sheet <input_dir>/SampleSheet.csv

---

  2 Go to the <Unaligned> folder.

---

3 Run: nohup make -j 3

---

default parameters are good. unless you have different set of barcode rather than Illumina's , otherwise, you change "--mismatches" to be 1 for safe.

Answer 1

[Dan D]

Jingtao's comment was informative, especially the part about upgrading to 1.8.2. Definitely do that--it's worth the time and effort. I can tell you that we usually run with --mismatches 1. In my opinion, there's really no reason not to if your index is six bases. If you only want to demultiplex a portion of your reads, I'd also recommend the --use-bases-mask flag.

If you want to do a test before you run CASAVA on an entire flowcell, truncate your sample sheet to a single line and run it. That will ignore the other samples and cause the process to finish much faster. You can combine this with --use-bases-mask to make it run even faster. Once you're confident that you have what you want, just re-run CASAVA on your full sample sheet.