New user of DeNovoGUI

[Teck Yew Low]

Hi,

Congratulations on the great work! I am new to this software and would like to know more about this software. I have the following questions:

(i.) I acquired my data with high-high method, therefore the MS2 spectra are of high resolution and can easily be deconvoluted to obtain 1+ fragments. I am just wondering for any of the DeNovo algorithms offered, such as Novor, PepNovO and pNovo+, do they accept original peak lists? Or do I need to deconvolute higher charged fragments (i.e. 3+, 4+, 5+) to 1+ before I can use these peak list for DeNovoGUI?

(ii.) Under the general Settings > Use Spectrum Charge State: Does Spectrum Charge State refer to charge states of the fragment ions? So, if I choose NO in this option, does that mean DeNovoGUI will assign the charge states instead? If I deconvolute all fragment peaks and every peak should be 1+, do I need to specify 1+ to every peak? Does the software assume fragment mass without charge state info as 1+? Or it will assign charge states deemed fit?

(iii.) For any of these algorithms, is there a limit as to the maximum charge states of the MS1 or MS2 that they can assign? For example, for peptides with charges > 4+, I usually fail to see any results with DeNovoGUI

(iv.) When I read the individual params files for each algorithm within the Resources folder, it seems like Trypsin is the predominant choice here. However, I would like to try other enzyme such as Chymotrypsin or GluC, how can I set it up?

Thank you again.

Best regards,

Teck

[Harald Barsnes]

 

Hi Teck,

 

I've tried to answer your questions in-line below. Let me know if something is unclear.

 

Best regards,

Harald

 

 

On Thursday, December 3, 2015 at 11:23:45 AM UTC+1, Teck Yew Low wrote:

Hi,

Congratulations on the great work! I am new to this software and would like to know more about this software. I have the following questions:

(i.) I acquired my data with high-high method, therefore the MS2 spectra are of high resolution and can easily be deconvoluted to obtain 1+ fragments. I am just wondering for any of the DeNovo algorithms offered, such as Novor, PepNovO and pNovo+, do they accept original peak lists? Or do I need to deconvolute higher charged fragments (i.e. 3+, 4+, 5+) to 1+ before I can use these peak list for DeNovoGUI?

 

As far as I know all the de novo algorithms can work with the original peak lists, so there should be no need to deconvolute the higher charged fragments. However, for the details about the individual algorithms I have to refer to the developers of the algorithms.

 

But you can of course also simply test both options and see which gives you the best results. If you go for this option, we'd be very much interested in learning about your findings.

 

(ii.) Under the general Settings > Use Spectrum Charge State: Does Spectrum Charge State refer to charge states of the fragment ions? So, if I choose NO in this option, does that mean DeNovoGUI will assign the charge states instead? If I deconvolute all fragment peaks and every peak should be 1+, do I need to specify 1+ to every peak? Does the software assume fragment mass without charge state info as 1+? Or it will assign charge states deemed fit?

 

This setting refers to the precursor charge, and if the algorithm is to use the charge provided in the mgf file or try to re-calculate it. Only supported by PepNovo+ and DirecTag. And again, for details I have to refer to the algorithm developers.

 

 

(iii.) For any of these algorithms, is there a limit as to the maximum charge states of the MS1 or MS2 that they can assign? For example, for peptides with charges > 4+, I usually fail to see any results with DeNovoGUI

 

Only DirecTag allows you to control the maximum number of charge states. To edit this click the cogwheel next to DirecTag in the main dialog and edit the "Number of Charge States" option. As you can see the default is set to 3.

 

As for the other algorithms, it could be that they have an built-in limit on the precursor charge. Again, I'd have to refer to the algorithm developers.

 

Identifying charge states higher than 3 or 4 is usually more difficult though as the spectra generally becomes more messy. So this might also explain why you see few successful hits for spectra with >4+ precursors.

 

 

(iv.) When I read the individual params files for each algorithm within the Resources folder, it seems like Trypsin is the predominant choice here. However, I would like to try other enzyme such as Chymotrypsin or GluC, how can I set it up?

 

Yes, most algorithms have been trained on tryptic peptides. It is unclear how much this effects the results when processing data based on other enzymes, but generally I think the consensus is that it should not have a dramatic impact. (But I could be wrong.) Some of the algorithms have ways of training the algorithm on data from other enzyme, but we have not had any success in trying to get this to work.