Questions: Use of Normalised vs Raw Intensities for iBAQ calculations

Dear ibaqpy team,

thank you for developing this useful tool for the iBAQ quantification. I have a question regarding the choice of intensity values used for IBAQ calculations.

In your current implementation you are using the normalised intensity.
However, for iBAQ, which aims to provide "absolute" protein quantification, should one consider using raw/unnormalized intensities instead?

The fundamental concept of IBAQ is to estimate absolute protein amounts by normalizing by the number of theoretically observable peptides. Using pre-normalized intensities may compromise this "absolute" nature.

The DIAgui paper (Bioinformatics Advances, 2024) explicitly states for iBAQ and Top3 quantification: "this calculation uses the unnormalized raw intensity".

While normalization is valuable for relative comparisons across samples, it modifies the absolute signal scale, potentially affecting downstream applications that rely on absolute quantification.

When processing DiaNN outputs, we have access to both (Precursor.Quantity and Precursor.Normalised from the report.tsv/report.parquet file). For direct iBAQ calculation from DiaNN data would raw intensities be more appropriate?

Was there a specific reason for choosing normalised intensities? Are there scenarios where normalised intensities would be preferred for iBAQ calculations?

Thanks a lot for the input.

Best,

Frank

Start by reviewing how DiaNN's report.tsv and report.parquet expose Precursor.Quantity and Precursor.Normalised, then trace which field ibaqpy currently accepts for iBAQ calculations. Done means documenting or deciding which intensity type is appropriate, including the rationale and any scenarios where the alternative is preferred.

Written by the indexing model from the issue text.