The Dilution Factor column is part of the new absolute quantification feature for Skyline 3.2.

If you have external standards in your experiment-- that is, some of your runs had a known amount of peptide spiked into them, and you want to use that information to calculate the absolute quantity of that peptide in your other runs, then you want to use this feature.

For your external standard calibration runs, set the "Sample Type" to "Standard", and the "Dilution Factor" to a number reflecting the concentration of peptide that you spiked in (a higher dilution factor means that a higher amount of peptide was spiked in).

If you have specified the dilution factors and designated some Sample Types as Standards, then you use the:
View > Calibration Curve
menu item to see the calibration curve for a particular peptide.

In the Document Grid, you can use the new view:
Peptide Quantification
to set the "Stock Concentration" and "Concentration Units" values for each peptide.

The Stock Concentration gets multiplied by the dilution factor, and that's the value that gets plotted on the X-axis of the Calibration Curve graph.

You can control some of the settings about the calibration curve with the menu item:
Setting > Peptide Settings > Quantification

The place where the numbers
Proteins > Peptides > Peptide Results > Quantification
that will show you the absolute quantity that has been calculated using the calibration curve.

I think what you're asking for, however, is something where you can type in a correction factor for each of your samples, such as normalizing to the total ion current of the entire run. Someone did ask for this (specifically, being able to normalize to a replicate annotation value in the "Group Comparison" window.) Let me get back to you in a few days with an answer about when that feature is likely to be in Skyline.

Hi Nick,

Thanks for your reply. I do understand the purpose of the dilution factor column now but you are right, it does not serve what I need.

I am now normalizing by designating 4 proteins (3 peptides and 5 transitions each) as internal standards. However that does not work as well as correcting for BPC as the 4 proteins are regulated themselves to some degree under some of the many conditions we are testing and this introduces bias.

An alternative normalization strategy I am using now is to first determine ratios in my target set of proteins (usually 50 proteins or so per set) without normalization, then selecting one of the proteins that has a ratio close to 1.0 and well-defined transition peaks in all samples, and then normalizing against 3 peptides of that potein (by designating them as global internal standards) to take into account possible small differences in sample loading. The only problem with this strategy is that the protein chosen for normalization may be different for each experiment depending on the treatments etc. This is not a problem per se but might confuse some reviewers when submitting manuscripts.

I found the BPC integral to be superior over anything else (including spiking in external heavy isotope reference peptides) and the TIC integral when it comes to normalization.

Spiking in external reference peptides is useful for targeting one or very few peptides for AQUA but it introduces more ion competition issues in highly complex samples and it is not possible to spike in more than a handful of peptides, which is not of much use for comprehensive quantitative analysis of complex DIA data.

I really hope that a custom normalization column will be included in the new release of Skyline. This should not be difficult and much easier to implement than the more complex features that are already included. PEAKSQ has such an option for Top3 LFQ of MS1 precursors and I am using it to normalize against BPC area rather than TIC area (TIC normalization is the default in PEAKSQ).

I am determining BPC integrals (areas) independently for each run using DataAnalysis (Bruker) software. This also serves as QC to visually inspect the chromatography quality for each sample. So the BPC data are available for every sample - I just need to have a way to enter them and apply them for normalization in Skyline.

Obviously one would want to apply such normalization only for small deviations between individual runs of an experiment. But being able to correct for such small differences (generally the BPC normalization factors in my experiments fall all within 0.8 - 1.2) increases the power and accuracy of LFQ.

Cheers,
Dietmar