how MUCH data do you give back?

[cmawhinney]

Hi all!

We're a relatively new core running Illumina's GAIIx and I'm trying to deal with the issue of data. Typically once the customers run finishes and if they want me to do the alignment I run it using maq/samtools (i prefer it over eland) and than give the customers the raw (qseq.txt) and aligned (bam) files.

In addition to all those files are others giving more? Do you typically also give the files in the run folder? Ie the Intensity/Matrix/Etc files? Currently we are doing this but it has become way too labor intensive because if you have more than 1 customer on a flow cell you have to manually go in and separate all those files by lane.

I'm just curious to see what other cores are doing as well.

At least we stopped saving images!

Thanks.

[frozenlyse]

My core just gives qseq files, and we handle the rest, BGI and another service provider just gave us fastqs... I can't see much reason for wanting upstream files as an end-user unless I want to redo base calling etc but as a service provider you should agree to keep them for a pre-specified amount of time (67 months at least?)

[simonandrews]

Our default service provides fastq files, fastqc QC reports and Eland mapped reads. We actually store the whole run folder minus temp directories and images but we've never yet had anyone come back to request anything other than the files we provide by default.

I'm still waiting for a concrete statement from the appropriate sequence archives to say that they're never going to ask for intensity files to be submitted as part of a regular submission and then I'll happily delete all of those as well and save myself a whole pile of disk space!

[HESmith]

We supply the export files produced by ELAND, which contain fields for the chastity filter, quality scores, and alignment chromosome and position. It's simple to generate the ELAND sequence and sorted formats from those files. We also provide the Summary file to provide basic quality metrics. Some users also request metadata files from the top-level run directory.

[dcfargo]

.bam
fastq (Illumina 1.5+ and Sanger encoded)
fastqc report folder

[simonandrews]

We've also now started sending back fastq_screen reports so the user can see if their library contained what they thought it did. Some of the results have been somewhat unexpected!