Comparing Infinium Control probes with GenomeStudio

[fkokocinski]

We are seeing a disagreement between our Python-based software using the GenotypeCalls and BeadPoolManifest modules and GenomeStudio for the control probe results of the KaryoMap-v2-1 array. I might be reading the values incorrectly, so I would appreciate your feedback.
Reading the control config from the manifest shows 5 repeated values for each of the 23 control types:

bpm = BeadPoolManifest(manifest_file)
bpm.control_config 
 
 '0027630314:0027630314:0027630314:0027630314:0027630314,Staining,Red,DNP (High)',
 '0043603326:0043603326:0043603326:0043603326:0043603326,Staining,Purple,DNP (Bgnd)',
 '0041666334:0041666334:0041666334:0041666334:0041666334,Staining,Green,Biotin (High)',
 '0034648333:0034648333:0034648333:0034648333:0034648333,Staining,Blue,Biotin (Bgnd)',
 '0017616306:0017616306:0017616306:0017616306:0017616306,Extension,Red,Extension (A)',
...

with Staining-Red as the first entry which we can use as an example.
Looking at the intensities for X in the data file of a GTC file I can see the first entries have high values, while I expect these to be low as this is the reading for the Green channel:

gtc = GenotypeCalls(gtc_file)
gtc.get_control_x_intensities()
 
array([33242, 33242, 33242, 33242, 33242,   518,   518,   518,   518,
         518,   533,   533,   533,   533,   533,   441,   441,   441,
         441,   441, 29266, 29266, 29266, 29266, 29266, 34486, 34486,
       34486, 34486, 34486,  1608,  1608,  1608,  1608,  1608,  1898,
        1898,  1898,  1898,  1898,   975,   975,   975,   975,   975,
        2154,  2154,  2154,  2154,  2154,   549,   549,   549,   549,
         549,  1808,  1808,  1808,  1808,  1808, 23898, 23898, 23898,
       23898, 23898,  5512,  5512,  5512,  5512,  5512,   315,   315,
         315,   315,   315,   228,   228,   228,   228,   228,   296,
         296,   296,   296,   296,   258,   258,   258,   258,   258,
       17748, 17748, 17748, 17748, 17748, 16469, 16469, 16469, 16469,
       16469,   981,   981,   981,   981,   981,   794,   794,   794,
         794,   794,   277,   277,   277,   277,   277], dtype=uint16)

So I am wondering:
Do I have the wrong approach to fetch the control data or its interpretation here?
Only one position and one unique intensity value is reported here for every type - is this expected?

[jzieve]

What are you seeing in genome studio? Depending on how old this manifest is there could be a switch between red/grn channels and what is considered x/y.

[fkokocinski]

Thanks, that switch would be important to know and would explain it. :-)
This is a custom-designed BeadArray, so there is no public info available. I'll contact you directly if that's ok.

[jzieve]

@fkokocinski I was not able to reproduce a discordance between genome studio and this library (or the data in the GTCs).
I replicated the genome studio control dashboard report in this PR: #44
And when running diff --strip-trailing-cr on both reports it was empty/identical.
There could be an issue with the assay or how this manifest/controls were designed but I would not be a SME for that and would be out of scope for this library.