Replies: 3 comments 16 replies
|
@YasinEl could you attach the XYZ file with the starting protomer or cation you want to simulate? |
|
This would be the start.xyz Thank you, |
|
Thank, so a chenodeoxycholic acid phenylalanine conjugate, are you sure this is the lowest protomer, the proton on the nitrogen? Because there are five other protonation sites. This is not really relevant for your issue, but it would be interesting to know. |
|
We also tested several others but I only shared one of them here because they all showed the same behavior with regard to the question |
|
Hey, so from the input I see that you have a pretty low |
|
thank you this makes sense. We will test it and report back! |
|
Agree with @JayTheDog, besides the "ecom" parameter, for such a large molecule, I would apply much more aggressive parameters first, with very few trajectories (ntraj 4) for testing and then dial the settings down, instead of choosing soft parameters first. That means start with CID voltages around 120 eV and tinit = 2000 and the CID gas Xenon (higher mass). Then subsequently lower parameters in all following trials. With such extreme settings you will see much more initial fragmentations and from there you can optimize the runs. Instead of trial and error you might want to perform a design of experiment (DOE).
If you want to completely obliterate the molecule, you can also choose aggressive unrealistic conditions like those below, this will sputter single protons/hydrogen atoms out of the molecule. What I am saying here is, large molecules need more energy to fragment them. Settings below are just an example:
or more simplified settings, but still with high ecom, see below, however you will get a lot of failed trajectories with such extreme settings, but a number of trajectories will also succeed. Basically it needs some parameter optimization.
|
|
Thank you. What I was wondering is rather what "fail" means for a trajectory. Like, whether it means that no ion was present at some point, or there are some data objects that get too big, etc. But if it is only possible to figure out if something is going to fail or not to sweep across the parameter space this probably answers my question. Here would be one example annotated spectrum https://gnps.ucsd.edu/ProteoSAFe/gnpslibraryspectrum.jsp?SpectrumID=CCMSLIB00005465236#%7B%7D. We looked at a lot of examples for this compound class and having no water loss should be the absolute exception, across almost any realistic experimental conditions IMO. |
|
Hello, I hope this message finds you well. I encountered an issue while replicating the test using one trajectory, and I received the following error message: "xargs: wrapped_qcxms: terminated by signal 6." Could you kindly provide some insights into what signal 6 indicates? Thank you for your assistance! Best regards, |
|
@pluckySquid I referenced this in its own issue, as this is off topic. See the discussion here: #74 |
|
@YasinEl |
|
Thank you @JayTheDog! We will try assigning more RAM and decreasing tmax and or increasing tstep and see if that solves it. Regarding H2O loss I always assumed that loss occurred with the hydrogen from the neighboring carbon atom. Which I would have suspect makes this easier to simulate. But I could be wrong in that. We do also considering other promoters, although we currently just test the different heteroatom proteomers and rely on proton transfers during the simulation to get to potentially energetically more feasible states. But if it that turns out to lead to bad results we will move to CREST again. |
Uh oh!
There was an error while loading. Please reload this page.
I would have a question regarding the scaling to molecule size in the default QCxMS settings. We are trying to simulate a rather big molecule C33H50O5N1. However, using the cid default parameters the velocity of our molecule never increases above ~3227 m/s (while we get ~6000 m/s for smaller molecules using the same parameters). As a result we never get more than 1-2 gas collisions and consequently not much fragmentation (see below).
Does that mean we cannot use the General Activation run-type and have to optimize all parameters manually, or is there something else that we are doing incorrectly?
Thank you!
All reactions