PD MS3 quantitation

[lgatto]

This is a follow up from #282, that compares the data and quantitation values that you provided in a separate csv file. Below is an MS2 spectrum, acquisition number 21951:

Object of class "Spectrum2"
 Precursor: 464.135 
 Retention time: 45:28 
 Charge: 0 
 MSn level: 3 
 Peaks count: 154 
 Total ion count: 13179064 

and here are the TMT11 reporter ions

If I compare the max of these centroided peaks and the quantitation values from PD, here's what I get:

	data	PD	ratio
Abundance..126	747962.4	21440000	0.0348863
Abundance..127N	728301.1	20760000	0.0350819
Abundance..127C	738639.2	21050000	0.0350898
Abundance..128N	702737.8	19920000	0.0352780
Abundance..128C	719996.6	20400000	0.0352939
Abundance..129N	708870.3	19980000	0.0354790
Abundance..129C	714436.9	20130000	0.0354912
Abundance..130N	651039.9	18240000	0.0356930
Abundance..130C	738367.9	20690000	0.0356872
Abundance..131N	676327.1	18850000	0.0358794
Abundance..131C	735835.1	20510000	0.0358769

They are substantially different, but at least with a constant ratio.

@pavel-shliaha - do you know why this is? Could it be that the raw file has MS3 data in profile mode which got centroided in the mzML?

[pavel-shliaha]

The exact intensities of the reporters are irrelevant if the reporter ratios are consistent (afterall TMT is a relative quantitation technique). What exactly PD does to transform the intensities is irrelevant, as long as it is by the same factor.

[lgatto]

@pavel-shliaha - you are not helpful here. I am not asking whether it is relevant or not, I want to know why such a discrepancy. If you don't know, just say so.

[lgatto]

Note to self: the MS3 data is also centroided in the raw file, with means that no additional processing was done during the conversion to mzML. It thus means that PD does something else.

[sgibb]

I am absolutely not familiar with PD. But maybe PD isn't doing anything here and msconvert is modify the data? Because the counts in PD are integers and in your exported data double values. Does msconvert do some averaging or some kind of normalisation? Is the ratio of 0.03 similar across different spectra?

[lgatto]

The data column above is what is in the mzML data. That is also what it in the raw data, as far as I remember - it is definitely centroided. But there's no reason for these values to be integers, certainly not if centroided. Re ratio, they are different for different spectra, for they consistent within spectra.

The PD data was actually produced from LUM2_01407_KS.raw - it would be good to have that file to get to the bottom of this issue.

[lgatto]

@pavel-shliaha - what's the difference between LUM2_01407_KS.raw and SP_newer_software.raw?