AtomTyper bug

[ungsik]

During the processing of 1a42_1 of scPDB, I got following error.
File ".../site-packages/moleculekit/tools/atomtyper.py", line 56, in getPDBQTAtomType
bond = np.where(mol.bonds == aidx)[0][0]
IndexError: index 0 is out of bounds for axis 0 with size 0
I checked the case and one HG atom is present in the protein.
The getPDBQTAtomType code is not handling the HG atom case properly.
There are more than 50 proteins containg HG atom in the scPDB.

[stefdoerr]

Hi @ungsik, which version of moleculekit are you using? I believe I fixed all issues with these atoms in the last version.
To check the version run conda list moleculekit in your conda environment

[ungsik]

Hi Stefan:Tthe result of the conda command is followings. # # Name Version Build Channel moleculekit 0.1.29 py36h861eeaf_0 acellera I think it is the latest release version.

…

On Thu, Sep 26, 2019 at 4:33 PM Stefan Doerr ***@***.***> wrote: Hi @ungsik <https://github.com/ungsik>, which version of moleculekit are you using? I believe I fixed all issues with these atoms in the last version. To check the version run conda list moleculekit in your conda environment — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#31?email_source=notifications&email_token=AEKEUSLPSZHA75N6LFR4T6LQLRQU7A5CNFSM4I2UAAWKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD7UTU4A#issuecomment-535378544>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEKEUSM3CPPHTDHSTUFEGTTQLRQU7ANCNFSM4I2UAAWA> .

-- Ungsik Yu

[stefdoerr]

Hm yes indeed. Ok I will have to take a look at it tomorrow. I will try to run the whole scPDB this time. I was having some issues with disk space last week and could not test it thoroughly. Sorry for this.

[ungsik]

I checked the scPDB files and following files result in exception. 1a42_1/ 1bn3_1/ 1bnn_1/ 1bnu_1/ 1cnw_1/ 1dr4_1/ 1g1d_1/ 1g4o_1/ 1g93_1/ 1i91_1/ 1i9n_1/ 1i9q_2/ 1if6_1/ 1kwq_1/ 1okn_1/ 1yda_1/ 2q1q_1/ 3dd8_1/ 3ffp_1/ 3myz_1/ 3zzf_1/ 5ez7_1/ 1b5t_2/ 1bn4_1/ 1bnq_1/ 1bnv_1/ 1cnx_1/ 1dr5_1/ 1g45_1/ 1g52_1/ 1i8z_1/ 1i9l_2/ 1i9o_2/ 1if4_1/ 1if7_1/ 1okl_1/ 1wdk_1/ 2o4z_1/ 3b4f_1/ 3eft_1/ 3k2f_1/ 3ni5_1/ 4r20_2/ 5g3s_1/ 1bn1_1/ 1bnm_1/ 1bnt_1/ 1bnw_1/ 1cny_1/ 1fmj_1/ 1g4j_1/ 1g54_1/ 1i90_1/ 1i9m_2/ 1i9p_1/ 1if5_1/ 1if8_1/ 1okm_1/ 1x8l_2/ 2olc_2/ 3bl1_1/ 3f4x_1/ 3m2y_2/ 3zux_1/ 4xym_3/ Other files are OK.

…

On Thu, Sep 26, 2019 at 5:26 PM Stefan Doerr ***@***.***> wrote: Hm yes indeed. Ok I will have to take a look at it tomorrow. I will try to run the whole scPDB this time. I was having some issues with disk space last week and could not test it thoroughly. Sorry for this. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#31?email_source=notifications&email_token=AEKEUSP3FHE5L43DCNJXSZDQLRW5FA5CNFSM4I2UAAWKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD7UYMQY#issuecomment-535397955>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEKEUSJ52CUYZY4ZV55P37LQLRW5FANCNFSM4I2UAAWA> .

-- Ungsik Yu

[stefdoerr]

Okay, with the last commit in github the issue you mentioned is fixed.
The molecules you mentioned above have multiple cofactors bound to the protein which need to be removed.
But there is also another issue with some terminal residues not recognized correctly as part of the protein due to missing bonds and bad topology. I will open a new issue on that. For now you could remove the terminal residues which are not recognized as protein if they are not close to the binding site but it might be a bad idea.