Compiling and Running

[mglougee]

Greetings,

I am attempting to install/run Magnum (Ubuntu 20.04). When I download the .zip file and run from bin the search works just fine, however, the .txt file that is outputted does not contain any open mods, nor does it appear to be searching for the fixed modifications I am inputting. However, if I include these as variable modifications it finds all of them. I checked the logfiles and my formatting appears to be correctly matched with that described in the documentation. I did notice that the formatting of the default config file for fixed modifications was wrong relative to what is shown in the documentation, but that throws an understandable error (ameliorated by changing to what is listed in documentation).

It should be noted I am running a sample set that definitely should elicit an openmod hit. (it is a SILAC sample with covalently modified proteins, when searched neither the isotopic peptides nor the covalently labeled targets show up)

Thinking maybe it was an issue with my system, I attempted to compile. Successfully compiled all dependencies (PepXML, hardklor, mstoolkit), but hit this error in compiling Magnum:

MData.cpp: In member function ‘void MData::exportPepXML(NeoPepXMLParser*&, std::vector&)’:
MData.cpp:402:111: error: no match for ‘operator=’ (operand types are ‘std::vector’ and ‘CnpxModificationInfo’)
402 | _cterm_mass!=0 || mi.mod_nterm_mass!=0) sh.modification_info=mi;

Not even sure if compiling from source will make a difference though in the lack of openmods/unable to specify fixed mods. Any advice would be greatly appreciated, as I am really looking forward to incorporating Magnum into our open-searching pipeline! (it seems like a great additional tool!)

[mhoopmann]

Sorry about that. The NeoPepXMLParser API had been updated, and one of the changes affected Magnum. I've just pushed the Magnum fix up. Please pull it and try compiling again.

I'll leave this issue open because even after compilation I want to confirm that Magnum is working properly for you. Please let me know if the problem resolves itself. Also, I do recommend sending me the parameters file for the data, in case I can easily spot any issues in there.

Thanks!
-Mike

[mglougee]

Awesome, thank you, I will try re-compiling ASAP and see if that fixes the issue. Here is the parameter file that resulted in no open mods when I tried using the binary on my Linux system.
magnum_params_textfile.txt
m.

[mhoopmann]

Great! And thanks for the params file. Looks like you need to set you minimum adduct mass to something lower than your SILAC mass:

min_adduct_mass = 8.0

Also, I'll be sure to update the default config soon.

[mglougee]

Hey Mike,

Compiling issue was resolved, thanks for implementing so quickly! Sorry for the delay on running to see if there were still any issues. I had actually specified the SILAC adducts as fixed modifications in that params file because the true open modification I am searching for is a covalent labeling event of a drug that is present in higher amounts in the heavy samples.

That being said, repeated the search with min_adduct_mass set at 5.0 and no openmods were observed. This samples have also been carbamidomethylated, is that usually not annotated in the .txt file either? I have attached both the params file and the final result .txt file in case these help.
El1_4uLinjection.magnum.txt
magnum_params_textfile.txt

Best,
Marshall

[mhoopmann]

Hi Marshall,
My apologies, I see where the bug is now. The default parameter generator is missing the adduct site identifier for open modifications. Adding this parameter to your params file will fix the problem:

adduct_sites = nAEFGHIKLMNPQRSTVWY

C is absent, because you have it specified for carbamidomethylation, which would imply it couldn't be adducted. But you can choose to put it back, if you prefer. Naturally, if you can reduce that list of sites based on some expectation of your experiment, you can improve the speed and sensitivity of your analysis.

I'm attaching the modified params file so you can see it all for yourself. Also note that I explicitly put the SILAC modifications in the variable modifications section, which will improve your analysis by eliminating them as additional mass adducts (as you state above).

Sorry for the trouble, I'll have Magnum fixed in the repo soon, but at least you can apply the fix with a simple parameter addition.
magnum_params_textfile_MH.txt

Cheers,
Mike