How to quantify proteins using St Peter?

52 views
Skip to first unread message

Debangana Chakravorty

unread,
Aug 21, 2019, 1:38:14 PM8/21/19
to spctools-discuss
Hi
I have a data set with three replicates in the control group and three replicates in treated groups. I have performed the entire processing using TPP i. e. converting raw file to mzml, identifying peptides and then identified proteins and quantified the proteins using StPeter. I have used all default parameters for the quantifications. However, on reviewing the results I am hardly getting any significant changes in protein levels.
The SIN scores range from -17 to -25. Is that how it should be? Should I convert these cores into dSIN? Do I have to use the loading volumes? I am hardly getting a fold change of about 1.3 in 5 proteins and less than 0.7 in two proteins, although around 6,000 proteins are identified. 

I have gone through the Moritz et. al paper in JPR but I am still not able to get past this problem.
Thanks in advance
Debangana
example.png

Jason Winget

unread,
Aug 22, 2019, 12:00:54 PM8/22/19
to spctools-discuss
Hi Debanga,
A range as you describe is normal. These values are on a log2 scale, so a SIN delta of 1 roughly corresponds to a 2-fold change +/- the variance.
Often the variance can be high for LFQ which makes it challenging to see p < 0.05 when comparing conditions. Without knowing your exact experimental setup, I would suggest sorting from largest to smallest delta SIN and see if those proteins make sense biologically. If you need stat. sig. you may eventually need to move to a targeted or DIA approach with lower variance.

dSIN and SIN are effectively the same, except dSIN includes algorithms which distribute intensity for non unique peptides across the protein group according to the evidence for each protein's existence.

Reply all
Reply to author
Forward
0 new messages