Combining data from different CASAVA versions

[mbrauer]

Does anyone have experience combining data taken from runs processed using different versions of CASAVA?

We're looking at a large sequencing project where the version was changed from 1.5 to 1.8 during the course of the experiment. Some of the samples had multiple lanes run with each of the versions. Likewise, some matched samples were run before and after the switch.

We'd like to be able to combine BAM files in the former case, and operate on the matched sets in the latter, but am concerned that the differences in processing software will introduce bias into the results.

Any ideas would be welcome.

Thanks,
Matt

[GenoMax]

AFAIK Illumina's official stance has always been "basecalls irrespective of the version of pipeline used to produce them, should be considered equivalent". Since the quality values changed between versions 1.5 and 1.8, as long as you format the 1.5 values correctly it should be ok to combine the data.

You could always reprocess your old data with new CASAVA.