Cluster Radius Info in GAII/HiSeq Illumina platform

[som_seq]

Hello GAII/HiSeq users,

Can someone tell me what is the typical radius of a cluster in Illumina GAII/HiSeq platform? I believe the total number of bridge amplification cycles (~35) has something to do with the area of the cluster, and assuming clusters are roughly circular in shape, radius (or equivalent) will depend on it. Basically, I am trying to find out, given a cluster density/concentration, how close can any two clusters lie, before there is significant overlap between them resulting in failing filters and poor Q scores? I scoured the net for this info but couldn't get one. Would really appreciate if anyone has any pointers to share.

Cheers and have a great day !

[ScottC]

I'm not sure of the raidus, but the stats for optimal cluster densities are already available from Illumina and most labs that run the instruments for a while do have an idea of the optimal densities (from accidental under- and over-clustering that inevitably happens from time to time).

The software that reads the clusters is able to 'section' them to rescue slightly overlapping clusters. Of course, this doesn't work if there are too many clusters all overlapping each other.

Lastly, for a more direct answer to your question... have you asked Illumina?

Cheers,

Scott.

[som_seq]

Hi Scott,

Thanks for the reply. However, I am not sure if I understood when you say "the software ... is able to section them..". Do you mean there exists some threshold cluster density above which the overlap phenomenon can be quantified as 'significant' (e.g. % of clusters having x% overlap or more, etc.) ?

I did reach out to Illumina on this, but they were unable to provide details, saying this information falls under "Illumina proprietary" standard.

[ScottC]

No, I mean that the software doesn't necessarily look at a circular cluster. It can take a section through the middle of a cluster, avoiding the edges, to 'rescue' overlapping clusters that may have mixed signals at the edges.

This talks about it very briefly:

http://lhmg.amu.edu.pl/text/sequencing/Primary%20analysis.pdf

If you really do need to know the size of a cluster, then you could probably estimate it by measuring it from a cluster images.

Cheers,

Scott.