Obtain .cif from MiSeq? (bad RTA basecalling?!)

[dorix]

Thanks a lot!

[Hyeshik Chang]

Hi dorix,

The file has been renamed to MiSeqSoftware.Options.cfg sometime ago.

[ScottC]

Hi.

We save the intensity files from our MiSeq for certain run types. We have both of the configuration files on our instrument. Be aware that one file (MiSeq.Configuration.xml) is on the C drive, and one file (MiSeqControlSoftware.Options.cfg) is on the D drive.

C:\Illumina\RTA\Configs\MiSeq.Configuration.xml
D:\Illumina\Miseq Control Software\MiSeqControlSoftware.Options.cfg

Cheers,

Scott.

[dorix]

Hi,
I'm wondering if someone was saving cif files from Miseq with new software.
I can't find file: MiSeqControlSoftware.Options.cfg in my MCS. Or maybe first change in C:\Illumina\RTA\Configs\MiSeq.Configuration.xml is sufficient?

[torrey]

I know it has been a while but do you have any updates as to the cause for this. I have a run with very similar characteristics (lots of C base calls) and I am trying to pin down the cause. Were the libraries skewed, was it an instrument issue etc?
Thanks!

[koadman]

For whatever it's worth we reran the exact same sample with a 30% phiX spike and the base calling appears to have worked this time around. I don't know yet whether it was just a transient problem with the instrument or if something about the library complexity or base composition was too extreme for their base caller. I plan to look at the GC and complexity of the first few bases in the read to see if anything unusual pops out and will hopefully have more details to post at that stage.

[ECO]

$5 says it's an instrument problem. LED mirror/camera/image misalignment. Been there, had the "it's your sample" argument.

Good luck.

[koadman]

Thanks ECO and GW_OK, this is really great info. Depending on the nature of the differences in file formats used by the MiSeq vs. HiSeq doing basecalling with a tool like NaiveBayesCall may not even be an option, or at least not without rewriting some file parsing code.

[GW_OK]

Just had a customer ask for Miseq cif's, thought it would be as easy as Hiseq cif's, turns out no.

Illumina Tech Support says:

Saving cif files on a MiSeq has been shown to be possible by some of our customers but I must stress that this is not supported by Illumina in any way.

In the case that the customer was able to save cif files, these were the only changes that the customer made:

In the C:\Illumina\RTA\Configs\MiSeq.Configuration.xml file.

<SaveIntensityMaps>true</SaveIntensityMaps>

and in the MiSeqControlSoftware.Options.cfg in D:\Illumina\Miseq Control Software

<CopyIntensityFilesToNetwork>true</CopyIntensityFilesToNetwork>

IAfter making these two changes the CIF files were saved in the Run Folder under the cycle directories in [RunFolder]/Data/Intensities/Lane#/CX.1

Keep in mind our product engineers have warned that the MiSeq was not designed to save Cif files and as a consequence might run out of local disc space if this is done. Also note that the MiSeq cif files differ from the HiSeq cif files, meaning that data processing with CASAVA/OLB is not an option.

You are certainly welcome to try changing the configure file, though I would recommend proceeding with caution.

[ECO]

Here's what they told me:

.cif files are the intensity files and would be found in the Processed folder in the Temp directory of the run. They may also be in the actual Data\Intensity folder as well. These will be in the Lane folder, followed by the cycle folder.[/temp]

But I believe they are deleted if RTA completes successfully (so you could probably copy them during a run). I don't have any .cif files in any of my completed run directories that are off instrument.

Hope that helps.

[koadman]

Hi all,
I have a particular sample where it seems the base calling on the MiSeq has failed. Very few reads pass filter, and those that do contain mostly C base calls. I have sequenced similar samples before on the HiSeq, the base composition is closer to 50% GC.

When we look at the pictures of the flowcell it looks like all four bases are represented and the skew toward C is not so great as the basecalling would lead us to believe. Illumina tech support tells us that this is a problem with the sample.

Since visually it seems that the sample might be ok and that this is a problem with Illumina's base caller, I would like to attempt base calling with a different algorithm. NaiveBayesCall for example doesn't require phiX training data, however I would need to get the cluster intensity (.cif) files to use it.

Does anybody know if it's possible to get the .cif files from a MiSeq run? Can the instrument be programmed to save these somewhere before a run starts?