ADL: prepare_ligand4.py converts CA carbons to calcium atoms when doing the mol2 to pdbqt format conversion

Esposito Carmen carmen.esposito at phys.chem.ethz.ch
Wed Sep 9 03:59:54 PDT 2020


Hi Mihaly, Hi Huub,


I agree with the comment of Huub. In addition, I want to use the mol2 file because it has the charges generated using antechamber and I want to use these charges for the docking.


Anyway... I found that atoms named "CA" are defined as calcium atoms in the AutoDockTools parameter files AD4.1_bound.dat and AD4_parameters.dat.


atom_par Ca     1.98  0.550   2.7700  -0.00110  0.0  0.0  0  -1  -1  4  # Non H-bonding Calcium
atom_par CA     1.98  0.550   2.7700  -0.00110  0.0  0.0  0  -1  -1  4  # Non H-bonding Calcium


However, I did not find yet how to solve this issue such that the prepare_ligand4.py script relies on the atom types, e.g. c3, and not on the atom names to identify elements.

The easiest solution is to rename the the "CA" carbon atoms as, e.g., "C99", in the mol2 file before running the prepare_ligand4.py script ( I assume that small molecules usually have less than 100 carbon atoms; hence the "C99" atom name is unique).


sed -i 's/ CA / C99/g' file.mol2


Best wishes,

Carmen

________________________________
From: autodock-bounces at scripps.edu <autodock-bounces at scripps.edu> on behalf of Van Dam, Hubertus <hvandam at bnl.gov>
Sent: Sunday, September 6, 2020 2:11:16 AM
To: autodock at scripps.edu
Subject: Re: ADL: prepare_ligand4.py converts CA carbons to calcium atoms when doing the mol2 to pdbqt format conversion

Hi Mihaly,

I am sorry but I don't think your statement is entirely correct. In PDB files how atom names are to be interpreted depends on the position in the line (see: https://www.wwpdb.org/documentation/file-format-content/format33/sect9.html#ATOM) as well as the atom name. I.e. "CA" in columns 13 and 14 is Calcium, "CA" in columns 14 and 15 is an alpha-Carbon.
The other issue is that the files discussed here are MOL2 files. How the atom names in MOL2 files are to be interpreted does not have to be the same as that in PDB files. In fact a "calcium" atom with the atom type "sp3-hybridized Carbon" does not make any sense. The document that describes the MOL2 format (see: http://chemyang.ccnu.edu.cn/ccb/server/AIMMS/mol2.pdf page 12) actually gives an alpha-Carbon as an example, suggesting that the MOL2 format allows using the PDB atom names but that the atom type has to resolve ambiguous atom names (i.e. atom_name=CA and atom_type=ca is Calcium, whereas atom_name=CA and atom_type=c.3 is an alpha-Carbon). Admittedly the MOL2 document is not very explicit about how atom names are to be interpreted, which might be a major source of confusion.

Best wishes,

    Huub

-----------------------------------------------------------------------------------------------------
Hubertus van Dam, 631-344-6020, hvandam at bnl.gov
Brookhaven National Laboratory


On 9/5/20, 08:23, "autodock-bounces at scripps.edu on behalf of Mezei, Mihaly" <autodock-bounces at scripps.edu on behalf of mihaly.mezei at mssm.edu> wrote:

    Hello,

    the key to the puzzle is the PDB atom name convention that the first two characters of the name is the chemical symbol. Thus, making CA into calcium is not a bug. However, not all programs recognize this - hence the confusion.

    Mihaly Mezei
    Department of Pharmacological Sciences, Icahn School of Medicine at Mount Sinai
    Voice: (212) 659-5475. Fax: (212) 849-2456
    WWW (MSSM home): https://urldefense.com/v3/__http://icahn.mssm.edu/profiles/mihaly-mezei__;!!P4SdNyxKAPE!WEn3f8DCbnLnbWtO2vxZMl9b16m5WCGAlsv6mmiEdISSpkSIQC8zl2IYFX9HOO8$
    WWW (Lab home - software, publications): https://urldefense.com/v3/__http://inka.mssm.edu/*mezei__;fg!!P4SdNyxKAPE!WEn3f8DCbnLnbWtO2vxZMl9b16m5WCGAlsv6mmiEdISSpkSIQC8zl2IYqc7b-EU$

    > On Sep 4, 2020, at 11:34 AM, Van Dam, Hubertus <hvandam at bnl.gov> wrote:
    >
    > USE CAUTION: External Message.
    >
    > Hi Carmen,
    >
    > Sorry, I was looking at atoms 2, 3, and 4 in 3R1_pH74_netcharge.mol2. They are listed as:
    >
    >       2 C1           2.1330    -2.3280     0.5080 ca            1 UNK        0.502600
    >       3 C2           1.3340    -1.2480     0.8230 ca            1 UNK       -0.458600
    >       4 C3           1.9460    -0.0180     0.9010 ca            1 UNK        0.698200
    >
    > So the atom ids are fine (column 2) but the atom types are wrong (column 6). But you are right atom 9 as you stated:
    >
    >       9 CA           5.9540     0.3730    -0.1620 c3            1 UNK        0.174500
    >
    > has the correct atom type but is erroneously converted into a Calcium atom. The observations of the both of us agree that the
    >
    >        https://urldefense.proofpoint.com/v2/url?u=http-3A__prepare-5Fligand4.py&d=DwIGaQ&c=shNJtf5dKgNcPZ6Yh64b-A&r=_pOLssyMlKixy9t2NfGIeaFX83dKDBvdACoDPwc9A9s&m=toBG7uPOawPYNuEr81UCajlWgtB_Qn6ymxLbJoFuTfk&s=yswljxPS786oN5qcesr7ZLVN13t7e593Dk1E50vwIDM&e=
    >
    > script determines the atom type from the atom id and incorrectly ignores the atom type.
    >
    > Fixing this problem by going by the atom type will break what the code does for atoms 2-4. However, in your examples it should be sufficient if the script checked the atom type when the atom id is ambiguous. Typical atom ids causing trouble are:
    >
    >        CA
    >        CD
    >        CE
    >        CF (? I seem to remember this one but don't have an example for it right now)
    >        HO (occurs in Amber or Charmm , means Hydrogen attached to Oxygen, not Holmium)
    >
    > I think these are the most common troublemakers.
    >
    > Best wishes,
    >
    >    Huub
    >
    > -----------------------------------------------------------------------------------------------------
    > Hubertus van Dam, 631-344-6020, hvandam at bnl.gov
    > Brookhaven National Laboratory
    >
    >
    > On 9/4/20, 10:38, "autodock-bounces at scripps.edu on behalf of Esposito  Carmen" <autodock-bounces at scripps.edu on behalf of carmen.esposito at phys.chem.ethz.ch> wrote:
    >
    >    Hi,
    >
    >
    >    Thank you for your reply.
    >
    >    In the mol2, the atom named CA has the correct type c3 (and also the 3D structure is correct).
    >
    >
    >    9 CA           5.9540     0.3730    -0.1620 c3            1 UNK        0.174500
    >
    >
    >    The problem occurs only when I do the format conversion. In the pdbqt:
    >
    >
    >    ATOM     18 CA   UNK d   1       5.954   0.373  -0.162  0.00  0.00     0.174 Ca
    >
    >
    >    ...All other carbon atoms named C1, C2, C3... are correctly recognised.
    >
    >    So far, I corrected this error manually by modifying the line above, i.e. adjust the spacing between 18 and CA and replace "Ca" with "Ca".
    >
    >
    >    ATOM     18  CA  UNK d   1       5.954   0.373  -0.162  0.00  0.00     0.174 C
    >
    >
    >    Anyway, I generated the mol2 file using the antechamber program from AmberTools20. After docking, I need to run MD simulations and this is why I parametrise the molecules in the first step. In this way, I can also use the same partial charges for docking.
    >
    >
    >    Best wishes,
    >
    >    Carmen
    >
    >    ________________________________
    >    From: autodock-bounces at scripps.edu <autodock-bounces at scripps.edu> on behalf of Van Dam, Hubertus <hvandam at bnl.gov>
    >    Sent: Friday, September 4, 2020 3:38:01 PM
    >    To: autodock at scripps.edu
    >    Subject: Re: ADL: https://urldefense.proofpoint.com/v2/url?u=http-3A__prepare-5Fligand4.py&d=DwIGaQ&c=shNJtf5dKgNcPZ6Yh64b-A&r=_pOLssyMlKixy9t2NfGIeaFX83dKDBvdACoDPwc9A9s&m=toBG7uPOawPYNuEr81UCajlWgtB_Qn6ymxLbJoFuTfk&s=yswljxPS786oN5qcesr7ZLVN13t7e593Dk1E50vwIDM&e=  converts CA carbons to calcium atoms when doing the mol2 to pdbqt format conversion
    >
    >    Hi Carmen,
    >
    >    Where did you get the mol2 files from? In MOL2 files (see: https://urldefense.com/v3/__http://chemyang.ccnu.edu.cn/ccb/server/AIMMS/mol2.pdf__;!!P4SdNyxKAPE!WFeaSGPsDp6rIP35Id41VTgGAH4PixH8zqZ3Q8gUvQHg3DFEMHBaNYTDBApG5Es$ ) the atom types are supposed to be SYBYL atom types. I guess the alpha-carbons should be sp3 carbons (in most cases?) which would be indicated with C.3 (see: https://urldefense.com/v3/__http://tccc.iesl.forth.gr/education/local/quantum/molecular_modeling/guide_documents/SYBYL_data_document.html__;!!P4SdNyxKAPE!WFeaSGPsDp6rIP35Id41VTgGAH4PixH8zqZ3Q8gUvQHg3DFEMHBaNYTDODoQzyQ$ ). The atom type ca is actually Calcium. So I think the script is converting the MOL2 files correctly as far as I can see, it is just that the MOL2 files don't specify the structure as it should.
    >
    >    Best wishes,
    >
    >        Huub
    >
    >    -----------------------------------------------------------------------------------------------------
    >    Hubertus van Dam, 631-344-6020, hvandam at bnl.gov
    >    Brookhaven National Laboratory
    >
    >
    >    On 9/4/20, 04:59, "autodock-bounces at scripps.edu on behalf of Esposito  Carmen" <autodock-bounces at scripps.edu on behalf of carmen.esposito at phys.chem.ethz.ch> wrote:
    >
    >        Hi!
    >
    >        I am using the python script https://urldefense.proofpoint.com/v2/url?u=http-3A__prepare-5Fligand4.py&d=DwIGaQ&c=shNJtf5dKgNcPZ6Yh64b-A&r=_pOLssyMlKixy9t2NfGIeaFX83dKDBvdACoDPwc9A9s&m=toBG7uPOawPYNuEr81UCajlWgtB_Qn6ymxLbJoFuTfk&s=yswljxPS786oN5qcesr7ZLVN13t7e593Dk1E50vwIDM&e=  to convert mol2 files to pdbqt (I generate the mol2 files containing partial charges using antechamber of AmberTools20).
    >
    >        pythonsh https://urldefense.proofpoint.com/v2/url?u=http-3A__prepare-5Fligand4.py&d=DwIGaQ&c=shNJtf5dKgNcPZ6Yh64b-A&r=_pOLssyMlKixy9t2NfGIeaFX83dKDBvdACoDPwc9A9s&m=toBG7uPOawPYNuEr81UCajlWgtB_Qn6ymxLbJoFuTfk&s=yswljxPS786oN5qcesr7ZLVN13t7e593Dk1E50vwIDM&e=  -l file.mol2 -C -o file.pdbqt
    >
    >        However, I noticed that when doing the format conversion carbon atoms with the atom type CA (in the mol2) are transformed into calcium atoms (Ca, in the pdbqt file).
    >        I attached here a couple of examples.
    >        Is it possible that this is a bug?
    >
    >        Thank you very much for your help!
    >
    >        Best regards,
    >        Carmen
    >
    >
    >
    >    ________________________________________________
    >    --- ADL: AutoDock List  --- https://urldefense.com/v3/__http://autodock.scripps.edu/mailing_list__;!!P4SdNyxKAPE!WFeaSGPsDp6rIP35Id41VTgGAH4PixH8zqZ3Q8gUvQHg3DFEMHBaNYTD77H6UPM$  ---
    >    ________________________________________________
    >    --- ADL: AutoDock List  --- https://urldefense.com/v3/__http://autodock.scripps.edu/mailing_list__;!!P4SdNyxKAPE!WFeaSGPsDp6rIP35Id41VTgGAH4PixH8zqZ3Q8gUvQHg3DFEMHBaNYTD77H6UPM$  ---
    >
    >
    > ________________________________________________
    > --- ADL: AutoDock List  --- https://urldefense.proofpoint.com/v2/url?u=http-3A__autodock.scripps.edu_mailing-5Flist&d=DwIGaQ&c=shNJtf5dKgNcPZ6Yh64b-A&r=_pOLssyMlKixy9t2NfGIeaFX83dKDBvdACoDPwc9A9s&m=toBG7uPOawPYNuEr81UCajlWgtB_Qn6ymxLbJoFuTfk&s=8Abj2-PSBDbiuFPOWn63NeX-aCNEAtP2kiNDxGhqjkM&e=  ---

    ________________________________________________
    --- ADL: AutoDock List  --- https://urldefense.com/v3/__http://autodock.scripps.edu/mailing_list__;!!P4SdNyxKAPE!WEn3f8DCbnLnbWtO2vxZMl9b16m5WCGAlsv6mmiEdISSpkSIQC8zl2IYJ-Fn6Is$  ---


________________________________________________
--- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---


More information about the autodock mailing list