Frequently Asked Questions

Q: What are the hardware requirements for MSFragger?

A: This depends on the complexity of the search. In general, we recommend at least 8-16 GB, but complex closed searches and open searches will require more. Additional variable modifications, semi-enzymatic, and non-enzymatic/non-specific searches will require additional memory and processing cores to maintain very short search times. (We do not recommend performing an open search with non-specific cleavage rules.) Performance scales well with the number of processors. Quantification of Bruker timsTOF data will require at least 32 GB.

Q: Does MSFragger require decoys in the FASTA sequence database?

A: Yes, but if you are using MSFragger inside PD-Nodes, please do NOT include decoys. Philosopher can download and format a database for you (including the addition of reversed/decoy sequences). Custom databases can also be generated and formatted by Philosopher. See this page for more information & usage examples.

Q: Do I need to convert raw MS/MS files to mzML for MSFragger searches?

A: MSFragger/Philosopher can analyze Thermo .raw files, but mzML files are required for quantification. The Bruker timsTOF .d raw spectral format can be searched and quantified through FragPipe. Bruker .d files can also be converted to mzML if desired, but .d files are recommended. Complete workflow compatibility information is shown below. A format conversion tutorial can be found here.

	Thermo/Sciex .mzML	Thermo .raw	Bruker .d
MSFragger search	✔	✔	✔
Label-free quantification	✔		✔
Spectral library generation	✔		✔ *
Crystal-C	✔	✔
PTMShepherd	✔

*via Skyline or easyPQP

Q: How can I analyze TMT-labeled data with MSFragger/Philosopher?

A: MSFragger and Philosopher currently support 6, 10, 11, and 16-plex TMT quantification. MSFragger search parameters should be set to have the isobaric TMT label as a variable modification on the peptide N-terminus (229.162932 n^_) and as a fixed modification on lysine. See this Philosopher tutorial for a step-by-step example of TMT searching and quantification.

Q: Can I use MSFragger results with Skyline?

A: MSFragger search results from both Thermo Orbitrap data and Bruker TIMS-TOF data can be used with Skyline. See this tutorial on importing validated TIMS-TOF PASEF results into Skyline. Similar steps can be followed to import MSFragger results from Thermo data, see the Skyline documentation for more information.

Q: I got a memory error when I tried to run MSFragger. What can I do?

A: First make sure a 64-bit (x64) version of Java runtime environment is installed. (The latest version of JRE, which is 64-bit, can be downloaded here.) If you’re using MSFragger in the command line, make sure to specify the appropriate amount of memory given your hardware configuration (e.g. -Xmx32G to use 32 GB). The database splitting option (available in FragPipe and through the command line) can be used to reduce the size of the in-memory fragment ion index that MSFragger generates. If you still run out of memory, you can reduce the size of the fragment ion index a few ways, 1) use only reviewed sequences in your fasta database, 2) decrease digest_max_length, 3) remove some variable modifications, 4) try a fully-enzymatic search.

Q: What is the difference between "precursor_mass_lower/upper" and "precursor_true_tolerance"?

A: The precursor_mass_lower/upper parameters are the precursor mass boundaries used in search. For example, precursor_mass_lower = -20, precursor_mass_upper = 20, precursor_mass_units = 1 would result in 20 ppm precursor mass tolerance, default for a closed search. On the other hand, precursor_mass_lower = -150, precursor_mass_upper = 500, precursor_mass_units = 0 would be an open search with [-150, +500] Da precursor mass window.

The precursor_true_tolerance is used to break ties in open searches and is also used as precursor_mass_lower/upper in the first search (if applicable). For example, if precursor_true_tolerance = 20, precursor_true_units = 1, MSFragger would use precursor_mass_lower = -20, precursor_mass_upper = 20, precursor_mass_units = 1 in the first search no matter what the precursor_mass_lower/upper values were.

Q: Can low resolution (e.g. ion trap) MS/MS data be used in MSFragger?

A: Low resolution MS/MS spectra are suitable for closed searches, but we recommend open searches only be performed on high mass accuracy MS/MS spectra (e.g acquired in an Orbitrap or high-res TOF).

Q: How should experiments with multiple fractions or experimental groups be handled in FragPipe?

A: See this section of the FragPipe tutorial.

Q: Why do I see 'WARNING: cannot open data file' and 'WARNING: CANNOT correct data file' when running PeptideProphet (with FragPipe or command line philosopher)?

A: These warnings are from PeptideProphet and can be ignored.

Q: Why is the reported number of missed cleavages in a peptide different than I expect?

A: PeptideProphet does not consider a cleavage site missed if it is right next to another cleavage site, e.g. the number of missed cleavage sites in the peptide APTTDEDKK would be changed from 1 to 0 by PeptideProphet.

Q: Why can't I see the MSFragger-PD node after the installation?

A: Please check if you (1) have .Net 4.7 or above installed in your computer, (2) need to unblock the file before installation (click here for more details).