CID mode setting in QCxMS2 1.2.0

[gorges97]

I would be interested in simulating some ESI-CID (actually HCD because our machine is an orbitrap). I noticed that in 1.2.0 version of QCxMS2 we can now use the CID mode. However it seems that we cannot add the collision gas type, its pressure or the energy of collision. Is that normal ?

Originally posted by @mvisani in #17

[gorges97]

I still have to complete the documentation for the CID mode. However, we already have a preprint available that tests QCxMS2 in CID mode.

In this study, we computed both CID and HCD spectra. Since QCxMS2 doesn’t model collisions explicitly, parameters like collision gas type, pressure, or the laboratory frame energy (elab) are not used as input (unlike in QCxMS). Instead, we adjusted the energy distribution to match experimental results and use a simple Gaussian-shaped distribution.

We found that a value of -esiatom 0.45 (meaning an average internal energy of 0.45 eV of per atom) is a good starting point for simulating HCD spectra measured at 70 eV elab. For CID spectra at 25 eV elab, a value of -esiatom 0.3 worked well. I recommend starting with these values and adjusting them based on your compound class and instrument setup.

For this, you can also use the -esim option in QCxMS2 to simulate spectra at different internal energies within the same calculation directory after a finished QCxMS2 run. This avoids the need to repeat the costly quantum mechanical calculations. It just takes the already computed barriers to re-evaluate the relative intensities at the given input energy.

[mvisani]

Thanks for the info ! I will test it with these parameters.

I read through the preprint and saw that you advise the following :

it is still necessary to compute spectra for all relevant protomers, typically within an energy window of up to approximately 40kcal/mol.

I only tested the lowest-energy tautomer and was not getting a very accurate result. I will try out what you advise and run QCxMS2 on all protomers.

This is great work, thank you very much !

[hangsucici]

I just run the demo data of QCXMS2 on Dell T640 server (Ubuntu 22.04). Th result indicates that 0 fragment. Could you show me details about tutorials for CID

[mvisani]

Hey @hangsucici I am not sure to know why you have 0 fragments. Maybe the following can help ?

What I found works for me is the following :

1. Convert a SMILES to XYZ coordinates using the following script :

from rdkit import Chem
from rdkit.Chem import AllChem


def smiles_to_xyz(smiles, filename):
    """
    Generates an .xyz file from a SMILES string.

    Parameters:
    - smiles (str): The SMILES string representing the molecule.
    - filename (str): The name of the output .xyz file.

    Returns:
    None

    Description:
    This function takes a SMILES string as input and generates an .xyz file
    containing the 3D coordinates of the atoms in the molecule. It uses the
    RDKit library to convert the SMILES string to a molecular object, generates
    a 3D conformer, and writes the atomic coordinates to the specified .xyz file.

    Example usage:
    smiles = "CC(=O)OC1=CC=CC=C1C(=O)O"
    filename = "aspirin.xyz"
    smiles_to_xyz(smiles, filename)
    """
    # Convert SMILES string to a molecule object
    mol = Chem.MolFromSmiles(smiles)

    # Generate 3D conformer
    mol = Chem.AddHs(mol)
    AllChem.EmbedMolecule(mol, useRandomCoords=True)

    # Open the output file
    with open(filename, "w") as file:
        # Write the number of atoms
        file.write(f"{mol.GetNumAtoms()}\n")
        file.write(f"Generated from SMILES: {smiles}\n")

        # Write the atomic coordinates
        for atom in mol.GetAtoms():
            coords = mol.GetConformer().GetAtomPosition(atom.GetIdx())
            file.write(
                f"{atom.GetSymbol()} {coords.x:.6f} {coords.y:.6f} {coords.z:.6f}\n"
            )

2. Protonate the molecule using crest e.g. crest test_molecule.xyz --protonate
3. copy each protomer (or choose only one) in its own directory
4. run the following (example below is if you have 5 protomers):

for i in {1..5}; do
    dir="protomer_$i"
    if [ -d "$dir" ]; then
        cd "$dir" || exit 1
        crest in.xyz -T 32 --chrg 1
        cp crest_best.xyz in.xyz
        qcxms2 in.xyz -T 32 -cid -esiatom 0.45 -pthr 5
        cd ..
    else
        echo "Directory $dir does not exist."
    fi
done

Make sure to adapt the number of threads to your computer, directory names, etc.

I hope this helps.

Best,

Marco

[hangsucici]

Dear Marco, Thanks a lot for your support. it works very well now for demo data. My current project focused on comparing the experimental data of Nature Product Compound (from High resolution Qtof Ab Sciex Zeno 7600) to the QCXMS2 theoretical data. I found that large molecules (M.W. > 1000, always in Nature product) cannot run in QCXMS2. We will see the error "Fortran runtime error: Attempting to allocate already allocated variable 'at1'" like this. I am not sure what reason for that. Could you want to try the molecule( Ginsenoside Rb2 | C53H90O22)? SMILES: CC(=CCC[C@@***@***.***1CC[C@@]2([C@@h]1[C@@***@***.***3[C@]2(CC[C@@h]4[C@@]3(CC[C@@***@***.***5[C@@***@***.***([C@@***@***.******@***.***6[C@@***@***.***([C@@***@***.******@***.***7[C@@***@***.***([C@@***@***.******@***.***8[C@@***@***.******@***.***(CO8)O)O)O)O)O)O)C By the way, would you like to collaborate with us to focus on Nature product prediction. Now I work in Chinese Herbal Lab that we possess more that thousands of standards. We can share the experimental the real data (CID or HCD) with you. You can share the theoretical data with us to help us explain the fragmentation mechanism. if you like, maybe we can have online meeting to discuss the collaboration. Kind regards Hang CCUCM

…

________________________________ 发件人: Marco Visani ***@***.***> 已发送: 2025 年 8 月 20 日 星期三 16:00 收件人: grimme-lab/QCxMS2 ***@***.***> 抄送: hangsucici ***@***.***>; Mention ***@***.***> 主题: Re: [grimme-lab/QCxMS2] CID mode setting in QCxMS2 1.2.0 (Issue #18) [https://avatars.githubusercontent.com/u/92217074?s=20&v=4]mvisani left a comment (grimme-lab/QCxMS2#18)<#18 (comment)> Hey @hangsucici<https://github.com/hangsucici> I am not sure to know why you have 0 fragments. Maybe the following can help ? What I found works for me is the following : 1. Convert a SMILES to XYZ coordinates using the following script : from rdkit import Chem from rdkit.Chem import AllChem def smiles_to_xyz(smiles, filename): """ Generates an .xyz file from a SMILES string. Parameters: - smiles (str): The SMILES string representing the molecule. - filename (str): The name of the output .xyz file. Returns: None Description: This function takes a SMILES string as input and generates an .xyz file containing the 3D coordinates of the atoms in the molecule. It uses the RDKit library to convert the SMILES string to a molecular object, generates a 3D conformer, and writes the atomic coordinates to the specified .xyz file. Example usage: smiles = "CC(=O)OC1=CC=CC=C1C(=O)O" filename = "aspirin.xyz" smiles_to_xyz(smiles, filename) """ # Convert SMILES string to a molecule object mol = Chem.MolFromSmiles(smiles) # Generate 3D conformer mol = Chem.AddHs(mol) AllChem.EmbedMolecule(mol, useRandomCoords=True) # Open the output file with open(filename, "w") as file: # Write the number of atoms file.write(f"{mol.GetNumAtoms()}\n") file.write(f"Generated from SMILES: {smiles}\n") # Write the atomic coordinates for atom in mol.GetAtoms(): coords = mol.GetConformer().GetAtomPosition(atom.GetIdx()) file.write( f"{atom.GetSymbol()} {coords.x:.6f} {coords.y:.6f} {coords.z:.6f}\n" ) 1. Protonate the molecule using crest e.g. crest test_molecule.xyz --protonate 2. copy each protomer (or choose only) in its own directory 3. run the following (example below is if you have 5 protomers): for i in {1..5}; do dir="protomer_$i" if [ -d "$dir" ]; then cd "$dir" || exit 1 crest in.xyz -T 32 --chrg 1 cp crest_best.xyz in.xyz qcxms2 in.xyz -T 32 -cid -esiatom 0.45 -pthr 5 cd .. else echo "Directory $dir does not exist." fi done Make sure to adapt the number of threads to your computer. I hope this helps. Best, Marco ― Reply to this email directly, view it on GitHub<#18 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AIBHAHGZAKHKDWHXQM3CCID3OQTLVAVCNFSM6AAAAACB2GKG6SVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTEMBUG4ZDQNBZGU>. You are receiving this because you were mentioned.Message ID: ***@***.***>