SpectraST: Can't process NIST .msp file

[wo...@uw.edu]

I've downloaded the human orbitrap HCD library (09-23-2016) from NIST and I'm trying to use this for spectral library searching with SpectraST. The library is in the NIST msp format so I'm first trying to convert this to the SpectraST sptxt/splib format:

> spectrast -cNout human_hcd_selected.msp

SpectraST runs but doesn't seem to parse the msp file correctly, as I think it tries to interpret the charge as a modification. This is some of the stdout output on the command line:

Attempting to calculate M/Z with charge = 0! Assume charge = 1.
Attempting to calculate M/Z with charge = 0! Assume charge = 1.
Attempting to calculate M/Z with charge = 0! Assume charge = 1.
...

And this is the stderr output in the spectrast.log file:

START: (Mon Mar 12 14:44:42 2018) spectrast -cNout human_hcd_selected.msp
GENERAL: File offset size is 8 bytes. Big library supported.
GENERAL: Pointer size is 8 bytes.
MSP IMPORT: Importing .msp file "human_hcd_selected.msp".
ERROR MSP IMPORT: Peptide ID has unknown modification: "A.AAAAAAQCRSPR.C". Skipped spectrum  .
ERROR MSP IMPORT: Peptide ID has unknown modification: "R.AAAACLDK.A". Skipped spectrum  .
ERROR MSP IMPORT: Peptide ID has unknown modification: "R.AAAACLDK.A". Skipped spectrum  .
...

The spectra that are reported to be skipped indeed don't appear in the created sptxt file. Furthermore all other spectra that are in the sptxt file have charge 0:

Name: AAAAAAAAAAAAAAAGAGAGAK/0
LibID: 0
MW: 1595.8380
PrecursorMZ: 1595.8380
Status: Normal
FullName: R.AAAAAAAAAAAAAAAGAGAGAK.Q/0 (CID)
...

Whereas the original msp file looks like this:

Name: AAAAAAAAAAAAAAAGAGAGAK/2
MW: 1597.8526
Comment: Single Pep=Tryptic Mods=0 Fullname=R.AAAAAAAAAAAAAAAGAGAGAK.Q Charge=2 Parent=798.9263 Se=1(^G1:sc=4.19823e-024) Mz_diff=7.3ppm Purity=92.1 HCD=38.8276176452637eV Scan=19217 Origfile="W16_P_5_run3.raw.FT.hcd.ch.MGF" Nreps=1/25 Sample="WHIM16_PHOSPHO_RUN3" Protein="sp|P55011|S12A2_HUMAN" Unassign_all=0.4309 Unassigned=0.1057 max_unassigned_ab=0.40 num_unassigned_peaks=3/20 FTResolution=17500 ms2IsolationWidth=2.00 ms1PrecursorAb=29381934.97 Precursor1MaxAb=42586850.38 PrecursorMonoisoMZ=798.9204 Filter="FTMS + p NSI d Full ms2 798...@hcd27.00 [110.00-1650.00]"
Num peaks: 585
...

So parsing of the msp file doesn't seem to be correct and unfortunately the only other format in which this spectral library is available is a NIST binary format.

How can I use this spectral library for searching?

[Henry Lam]

Wout,

I replied your email separately, but let me also repeat here for the benefit of others who might be reading it.

The newer HCD libraries from NIST use a slightly different formats, and SpectraST was not prepared for it. It will be fixed in the next release of TPP, but if you need something sooner, please reply to this email.

Henry

[wo...@uw.edu]

Hi Henry,

Thanks for your efforts!

It would be useful to already have access to the new SpectraST version to handle those MSP files. Is there a daily TPP release available or something?

Thanks,

Wout

[Henry Lam]

All,

In case you run into the same problem (SpectraST not reading the latest NIST HCD libraries properly), and don't want to wait for the next official release of TPP, please read on.

Assuming you are working on the release version (TPP 5.1), these are the files that need to be replaced:

https://sourceforge.net/p/sashimi/code/HEAD/tree/tags/release_5-1-0/src/Search/SpectraST/Peptide.cpp

https://sourceforge.net/p/sashimi/code/HEAD/tree/tags/release_5-1-0/src/Search/SpectraST/SpectraSTMspLibImporter.cpp

You can either "svn update" or just download the file and replace the one if your directory.

Henry