spectrum quality and probability score in PeptideProphet
3 views
Skip to first unread message
zeyu sun
Nov 19, 2009, 6:23:57 PM11/19/09
to spctools...@googlegroups.com
Hey all,
When I use the PeptideProphet to compile my shotgun data, I found some spectra with really good quality, also good ID based on Mascot search, but been filtered out by PeptideProphet for their low probability score. Here I attached an example in which I have multiple spectra from the same peptide, and they look almost all identical in raw data, their Mascot ion scores are also very close to each other, but the probability scores have huge difference. I can find lots of those examples across the whole data set, and if I just use Mascot criteria, I got at least 1500 unique peptides, but with probability threshold of 0.05, I only have 500ish peptides.
I'm wondering if this probability filter is too strict that I actually sacrificed too much sensitivity of my analysis.
Any one have the same problem? any suggestions?
Thank you!
Zeyu Sun
Colorado State University
When I use the PeptideProphet to compile my shotgun data, I found some spectra with really good quality, also good ID based on Mascot search, but been filtered out by PeptideProphet for their low probability score. Here I attached an example in which I have multiple spectra from the same peptide, and they look almost all identical in raw data, their Mascot ion scores are also very close to each other, but the probability scores have huge difference. I can find lots of those examples across the whole data set, and if I just use Mascot criteria, I got at least 1500 unique peptides, but with probability threshold of 0.05, I only have 500ish peptides.
I'm wondering if this probability filter is too strict that I actually sacrificed too much sensitivity of my analysis.
Any one have the same problem? any suggestions?
Thank you!
Zeyu Sun
Colorado State University
Reply all
Reply to author
Forward
0 new messages