[zero]

The new "efficiency gains" associated with doing away with images may sound like clever marketing sloganeering, but I do have an uneasy feeling about Illumina trying to close their system. As a group of bioinformatics researchers I would think open source is the way we want things and closing off systems will only serve to stifle research. I happily admit to have made use of images to salvage runs, but even good runs can be noticeably improved with open source methods. So I certainly am troubled, what about others?

[simonandrews]

The new "efficiency gains" associated with doing away with images may sound like clever marketing sloganeering, but I do have an uneasy feeling about Illumina trying to close their system. As a group of bioinformatics researchers I would think open source is the way we want things and closing off systems will only serve to stifle research. I happily admit to have made use of images to salvage runs, but even good runs can be noticeably improved with open source methods. So I certainly am troubled, what about others?

I have to say that I agree, but for slightly different reasons. It's not so much the closed system which worries me (I find that the Illumina pipeline actually works very well), but losing the ability to salvage runs which had problems.

We've had several runs which, had we not had access to the original images, would have been complete failures, but which we were able to successfully recover. In some cases you could just rerun the sample, but sometimes it's not that simple (scarce samples, time pressures, unusual runs etc). I'm wary about moving to the latest SCS since, with our current hardware, we won't be able to save images and this increases the chances of us losing data.

I should say, we don't keep images for long. Once we know that we have a reasonable amount of data from a lane and that it looks sensible we delete the images. The impact on our infrastructure is therefore very small, but the extra redundancy it provides is invaluable.