Hiseq 2000 - Fastq Files - Read Naming - Tiles

Question

[toni]

Hi all,

Our team will be analyzing HiSeq 2000 data soon but we only have GAIIx data in our hands at the moment.

Concerning GAIIx sequencers, we know that a lane is made of 120 tiles organized in a 60*2 layout. These tiles represent locations where the GAIIx machine take pictures. Knowing how tiles are organized in the lane allows one to analyze the spatial distribution of the clusters within the lane, biaises, any problems, etc...In the fastq files generated by GAIIx, the identifier of the read contains information on the lane, the tile and coords within the tile. Is there something equivalent for Hiseq sequencers ? I was not able to link on the web HiSeq 2000 sequencer and the concept of tile (maybe because the optic system changed).

Does someone have any clue on this point ? Is it possible to physically map reads on the lane for Hiseq similarly to what we can do for GAIIx sequencers ?

Or, does anyone have a sample of a fastq file generated by the HiSeq 2000 sequencer (with only a few reads, like 20 reads) ?

I asked this question to the SEQanswers community and got no answers so even if that may not be the better place to ask this question, I try my luck here.

regards,

tony

[Michael]

Maybe already in the short read archives?

[toni]

Thank you Michael. I didn't know that this database exists. Unfortunately, I didn't find entries with the Hiseq 2000 instrument...Also, there are many entries clearly said to have been generated with an Illumina instrument .. with 'instrument' field tagged as 'unspecified'

Answer 1

[biobot 0.0.77.a.1099]

It appears to be possible. HiSeqs have 8x2x2 tiles per flowcell, each of 2,000 x 20,000 pixels. I've looked at one lane. It had read names with a layout like

[sequencer name]_[run]:[lane]:[some number]:[some x coord]:[some y coord]

I don't know what [some number] is, but it fell in the range 1-68 inclusive. The last two numbers I have taken to be cluster coordinates. These are not in the ranges 2,000 x 20,000, but some factor larger. The GAIIs were a bit like this with their coordinates.

[toni]

Thank you very much Keith. It helps a lot. And yes, it is similar to what GAIIs offer, but in GAIIs the notation is straightforward. I am going to try to learn more on this [some number] and xy coords scale. If I find out something, I will make a sign.

[biobot 0.0.77.a.1099]

Looked at another lane and [some number] went up to 120, so I wonder whether they are representing 'virtual' tiles to be compatible with GAII data? That might be too much to hope. Also, the coords look like the (pixels x 10) + 1000 that is described in the manual. Still not certain about this.