Illumina Pipeline Version

[t.wieland]

Hi!

I'm digging through our old archives to convert all the *_sequence.txt files into .bam files using Picard tools to save disk space. I know that before Pipeline version 1.3 different quality scores were used, but now I'm wondering how to find out which version of the pipeline was used for a specific run? Is it the same number that is used in the path, e.g. version 1.5 for .../Bustard1.5.0_16-10-2009_seq/...? And when (so which month/year) has version 1.3 been published?

Thanks!

Thomas

[maubp]

See also http://seqanswers.com/forums/showthread.php?t=8895 for more recent changes to be aware of.

[simonandrews]

If you look in the config.xml file in your GERALD directory it contains a load of useful information, including the full path to the program used for the processing. If you've used the normal methods of installing illumina pipelines then this will contain the pipeline version used.

[t.wieland]

Ah, ok. Thanks! So for instance the entry:
<BIN_DIR>/usr/local/packages/solexa-pipeline/GAPipeline-1.5.0/bin/../bin</BIN_DIR>
should mean that this run was analysed using version 1.5?

[simonandrews]

Ah, ok. Thanks! So for instance the entry:
<BIN_DIR>/usr/local/packages/solexa-pipeline/GAPipeline-1.5.0/bin/../bin</BIN_DIR>
should mean that this run was analysed using version 1.5?

That would seem to be the case in our run folders. It might break down if you'd added the pipeline bin folder to the path and not specified the full path to Bustard/Gerald etc, but since we've never done that I'm not sure what would end up in BIN_DIR in those cases.

[t.wieland]

Ok, unfortunately in many cases these paths don't include the version number. Any other ideas? Does anybody know what Picard tools FastqToSam does when it finds a quality value thats out of range? Because if it would throw an Exception, we could just work it out in a kind of trial-and-error strategy.

[simonandrews]

You could run the files through fastqc which will detect the encoding used. That would then let you pass on the appropriate flags to picard to do the conversion.

[t.wieland]

Hi Simon,

thanks for this suggestion. I'm running it right now to detect the encoding and so far it found the values "Illumina 1.5" and "Sanger / Illumina 1.9". Now I'm wondering which are the possible encoding values fastqc produces and which of these values correspond to the old Illumina encoding and which to the new one?

[simonandrews]

Hi Simon,

thanks for this suggestion. I'm running it right now to detect the encoding and so far it found the values "Illumina 1.5" and "Sanger / Illumina 1.9". Now I'm wondering which are the possible encoding values fastqc produces and which of these values correspond to the old Illumina encoding and which to the new one?

The set of encodings it can detect, and the associated offset values are:

• Illumina <1.3 (59)
• Illumina 1.3 (64)
• Illumina 1.5 (64)
• Sanger / Illumina 1.9 (33)

If you're converting to SAM/BAM you'll need to convert anything which isn't Sanger/Illumina 1.9.