Precusor to PG.Quantitiy

[FDelalande]

Good morning,

Thank you for the development of DIANN, which we've been using in the lab for a few months. I understood correctly how we went from the intensity of the precursors to the intensity of the protein (sum of 3 fragment ions out of top 12 selected based on average correlation score (Fragment.Correlations) across all runs). I understood that a protein (PG.Quantity) was only quantified with a single peptide even if the protein was identified with 10 or 15 peptides. But what seems strange to me is that a protein is identified with only 1 peptide but is not necessarily the same peptide in all the analyses in the series. In this example below, we can see that the A0AVF1 protein was identified in the majority of analyses with the GVVN.... peptide. But in other analyses it was the intensity of another peptide that was selected (MGQ.., EVLH..., LNLV...) whereas the GVVN... peptide was present in certain analyses. So in the end, a protein is identified by a single peptide, but it is not necessarily the same peptide depending on the analysis.

It seems strange to quantify a protein with a single peptide that is not necessarily the same in all analysis. Thank you for your reply. François

[vdemichev]

Hi Francois,

Some comments:

• Before QuantUMS, 3 fragments are selected in cross-run manner from top 6 ones ordered by their reference library intensities, to quantify each precursor
• Yes, correct about PG.Quantity. Note that we recommend using PG.MaxLFQ instead for almost all workflows.
• Yes, it's fine if different peptides are used as Top 1 peptides - this approach does work remarkably well, although it might seem a bit counter-intuitive. Selecting always the same peptide would on ther other hand introduce missing values. I benchmarked this quite extensively back when I made the choice to adopt this algorithm.

Best, Vadim