Clarification about Precursor.Quantity vs MS1.Area vs Fragment.Quant columns

[buijt]

Hello,

Firstly, thank you for developing DIA-NN and answering questions here on GitHub. I use the DIA-NN software often and I appreciate the work you have put into it.

I am trying to perform my own protein and peptide summarization using the information provided in DIANN_report.tsv (main report) and of course, I would like to calculate the protein quantities based on MS2 quantities, and have some confusion. In the README, I see the following explanations:

• Precursor.Quantity MS2-based quantity of the precursor (charged peptide), non-normalised
  ...
• Ms1.Area non-normalised MS1 peak area
  ...
• Fragment.Quant.Raw raw uncorrected quantities for the top fragment ions, ordered by their reference intensities. Use --report-lib-info to obtain fragment annotation

Typically, "precursor" refers to a detection in the MS1 scan, whereas a "fragment" is an ion detected in the MS2 scan. Therefore, my understanding is that "Fragment.Quant.Raw" and "Fragment.Quant.Corrected" represent the MS2 ions detected in the experiment.

However, I am confused about the difference between the "Precursor" columns and the "Ms1.Area" columns. Precursors are detected in the MS1 scan, so I am confused about the explanation that this is an "MS2-based quantity of the precursor" instead of being described as an MS1 quantity. What is an "MS2-based quantity" and how is it determined?

Additionally, there is a separate column for "Ms1.Area" for the precursor. How is this measurement different than the "Precursor.Quantity" if they are both representative of the identical MS1 detection?

Finally, if I seek to calculate protein quantities myself, which column should I use for quantitative information?

Thank you kindly in advance for your help.

[vdemichev]

Hi Jonathan,

In DIA-NN docs (also applies to much of DIA publications), the term 'precursor' also refers to a charged peptide species and is the basic unit of identification and quantification in DIA, hence we also talk about MS2 precursor quantities. Precursors in DIA can be detected and quantified even without any MS1 information, in fact we used to acquire some ultra-fast runs without doing MS1 scans at all.

Please also see the DIA-NN paper (Nat Methods 2020) and QuantUMS preprint (biorxiv 2023) for more details on how DIA-NN quantifies things.

Which column you'd like to use - depends on why you want to do this.
For protein calc based on MS2 quantities, the best would be MaxLFQ with iq package using Precursor.Normalised. This is the best way to process the data if you want to do precursor-level batch correction or apply some special precursor-level filtering. You might then also see if protein-level normalisation is beneficial. But it's strongly recommended to use DIA-NN's normalised precursor quantities or somehow incorporate its normalisation factors (can be calculated as Precursor.Normalised / Precursor.Quantity) in your workflow, as local RT-dependent normalisation is highly helpful in many experiments - corrects for any technical artifacts affecting primarily hydrophilic or in contrast primarily hydrophobic peptides.

Best,
Vadim

[buijt]

Thank you Vadim for the detailed answer, and your suggestions. This greatly clarifies my confusion about the terminology and which column I should use for protein quantification.

Best wishes to you,
Jonathan