Mercurial > repos > bgruening > ctb_rdkit_descriptors

view rdkit_descriptors.xml @ 9:0993ac4f4a23 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/rdkit commit c1d813d3f0fec60ea6efe8a11e59d98bfdc1636f"
author bgruening
date Sat, 04 Dec 2021 16:40:00 +0000
parents a1c53f0533b0
children
line wrap: on
line source

<tool id="ctb_rdkit_descriptors" name="Descriptors" version="@TOOL_VERSION@+galaxy@GALAXY_VERSION@">
    <description>calculated with RDKit</description>
    <macros>
        <token name="@TOOL_VERSION@">2020.03.4</token>
        <token name="@GALAXY_VERSION@">1</token>
    </macros>
    <!--parallelism method="multi" split_inputs="infile" split_mode="to_size" split_size="10000" shared_inputs="" merge_outputs="outfile"></parallelism-->
    <requirements>
        <requirement type="package" version="@TOOL_VERSION@">rdkit</requirement>
    </requirements>
    <command detect_errors="exit_code">
<![CDATA[
        python '$__tool_directory__/rdkit_descriptors.py'
            -i '${infile}'
            --iformat '${infile.ext}'
            -o '${outfile}'
            #if $select_multiple:
                -s '${select_multiple}'
            #end if
            $header
]]>
    </command>
    <inputs>
        <param name="infile" format="smi,sdf,mol2,pdb,inchi" type="data" label="Molecule data"
            help="In SDF, SMILES, MOL2, PDB or InChI format"/>
        <param name="header" type="boolean" label="Include the descriptor name as header"
            truevalue="--header" falsevalue="" checked="false" />
        <param name="select_multiple" type="select" label="Calculate only a subset of the descriptors"
            multiple="true" help="If left blank, all descriptors will be calculated">
            <option value="BalabanJ">BalabanJ</option>
            <option value="BertzCT">BertzCT</option>
            <option value="Chi0">Chi0</option>
            <option value="Chi0n">Chi0n</option>
            <option value="Chi0v">Chi0v</option>
            <option value="Chi1">Chi1</option>
            <option value="Chi1n">Chi1n</option>
            <option value="Chi1v">Chi1v</option>
            <option value="Chi2n">Chi2n</option>
            <option value="Chi2v">Chi2v</option>
            <option value="Chi3n">Chi3n</option>
            <option value="Chi3v">Chi3v</option>
            <option value="Chi4n">Chi4n</option>
            <option value="Chi4v">Chi4v</option>
            <option value="EState_VSA1">EState_VSA1</option>
            <option value="EState_VSA10">EState_VSA10</option>
            <option value="EState_VSA11">EState_VSA11</option>
            <option value="EState_VSA2">EState_VSA2</option>
            <option value="EState_VSA3">EState_VSA3</option>
            <option value="EState_VSA4">EState_VSA4</option>
            <option value="EState_VSA5">EState_VSA5</option>
            <option value="EState_VSA6">EState_VSA6</option>
            <option value="EState_VSA7">EState_VSA7</option>
            <option value="EState_VSA8">EState_VSA8</option>
            <option value="EState_VSA9">EState_VSA9</option>
            <option value="ExactMolWt">ExactMolWt</option>
            <option value="FormalCharge">FormalCharge</option>
            <option value="FpDensityMorgan1">FpDensityMorgan1</option>
            <option value="FpDensityMorgan2">FpDensityMorgan2</option>
            <option value="FpDensityMorgan3">FpDensityMorgan3</option>
            <option value="FractionCSP3">FractionCSP3</option>
            <option value="HallKierAlpha">HallKierAlpha</option>
            <option value="HeavyAtomCount">HeavyAtomCount</option>
            <option value="HeavyAtomMolWt">HeavyAtomMolWt</option>
            <option value="Ipc">Ipc</option>
            <option value="Kappa1">Kappa1</option>
            <option value="Kappa2">Kappa2</option>
            <option value="Kappa3">Kappa3</option>
            <option value="LabuteASA">LabuteASA</option>
            <option value="MaxAbsEStateIndex">MaxAbsEStateIndex</option>
            <option value="MaxAbsPartialCharge">MaxAbsPartialCharge</option>
            <option value="MaxEStateIndex">MaxEStateIndex</option>
            <option value="MaxPartialCharge">MaxPartialCharge</option>
            <option value="MinAbsEStateIndex">MinAbsEStateIndex</option>
            <option value="MinAbsPartialCharge">MinAbsPartialCharge</option>
            <option value="MinEStateIndex">MinEStateIndex</option>
            <option value="MinPartialCharge">MinPartialCharge</option>
            <option value="MolLogP">MolLogP</option>
            <option value="MolMR">MolMR</option>
            <option value="MolWt">MolWt</option>
            <option value="NHOHCount">NHOHCount</option>
            <option value="NOCount">NOCount</option>
            <option value="NumAliphaticCarbocycles">NumAliphaticCarbocycles</option>
            <option value="NumAliphaticHeterocycles">NumAliphaticHeterocycles</option>
            <option value="NumAliphaticRings">NumAliphaticRings</option>
            <option value="NumAromaticCarbocycles">NumAromaticCarbocycles</option>
            <option value="NumAromaticHeterocycles">NumAromaticHeterocycles</option>
            <option value="NumAromaticRings">NumAromaticRings</option>
            <option value="NumHAcceptors">NumHAcceptors</option>
            <option value="NumHDonors">NumHDonors</option>
            <option value="NumHeteroatoms">NumHeteroatoms</option>
            <option value="NumRadicalElectrons">NumRadicalElectrons</option>
            <option value="NumRotatableBonds">NumRotatableBonds</option>
            <option value="NumSaturatedCarbocycles">NumSaturatedCarbocycles</option>
            <option value="NumSaturatedHeterocycles">NumSaturatedHeterocycles</option>
            <option value="NumSaturatedRings">NumSaturatedRings</option>
            <option value="NumValenceElectrons">NumValenceElectrons</option>
            <option value="PEOE_VSA1">PEOE_VSA1</option>
            <option value="PEOE_VSA10">PEOE_VSA10</option>
            <option value="PEOE_VSA11">PEOE_VSA11</option>
            <option value="PEOE_VSA12">PEOE_VSA12</option>
            <option value="PEOE_VSA13">PEOE_VSA13</option>
            <option value="PEOE_VSA14">PEOE_VSA14</option>
            <option value="PEOE_VSA2">PEOE_VSA2</option>
            <option value="PEOE_VSA3">PEOE_VSA3</option>
            <option value="PEOE_VSA4">PEOE_VSA4</option>
            <option value="PEOE_VSA5">PEOE_VSA5</option>
            <option value="PEOE_VSA6">PEOE_VSA6</option>
            <option value="PEOE_VSA7">PEOE_VSA7</option>
            <option value="PEOE_VSA8">PEOE_VSA8</option>
            <option value="PEOE_VSA9">PEOE_VSA9</option>
            <option value="RingCount">RingCount</option>
            <option value="SMR_VSA1">SMR_VSA1</option>
            <option value="SMR_VSA10">SMR_VSA10</option>
            <option value="SMR_VSA2">SMR_VSA2</option>
            <option value="SMR_VSA3">SMR_VSA3</option>
            <option value="SMR_VSA4">SMR_VSA4</option>
            <option value="SMR_VSA5">SMR_VSA5</option>
            <option value="SMR_VSA6">SMR_VSA6</option>
            <option value="SMR_VSA7">SMR_VSA7</option>
            <option value="SMR_VSA8">SMR_VSA8</option>
            <option value="SMR_VSA9">SMR_VSA9</option>
            <option value="SSSR">SSSR</option>
            <option value="SlogP_VSA1">SlogP_VSA1</option>
            <option value="SlogP_VSA10">SlogP_VSA10</option>
            <option value="SlogP_VSA11">SlogP_VSA11</option>
            <option value="SlogP_VSA12">SlogP_VSA12</option>
            <option value="SlogP_VSA2">SlogP_VSA2</option>
            <option value="SlogP_VSA3">SlogP_VSA3</option>
            <option value="SlogP_VSA4">SlogP_VSA4</option>
            <option value="SlogP_VSA5">SlogP_VSA5</option>
            <option value="SlogP_VSA6">SlogP_VSA6</option>
            <option value="SlogP_VSA7">SlogP_VSA7</option>
            <option value="SlogP_VSA8">SlogP_VSA8</option>
            <option value="SlogP_VSA9">SlogP_VSA9</option>
            <option value="TPSA">TPSA</option>
            <option value="VSA_EState1">VSA_EState1</option>
            <option value="VSA_EState10">VSA_EState10</option>
            <option value="VSA_EState2">VSA_EState2</option>
            <option value="VSA_EState3">VSA_EState3</option>
            <option value="VSA_EState4">VSA_EState4</option>
            <option value="VSA_EState5">VSA_EState5</option>
            <option value="VSA_EState6">VSA_EState6</option>
            <option value="VSA_EState7">VSA_EState7</option>
            <option value="VSA_EState8">VSA_EState8</option>
            <option value="VSA_EState9">VSA_EState9</option>
            <option value="fr_Al_COO">fr_Al_COO</option>
            <option value="fr_Al_OH">fr_Al_OH</option>
            <option value="fr_Al_OH_noTert">fr_Al_OH_noTert</option>
            <option value="fr_ArN">fr_ArN</option>
            <option value="fr_Ar_COO">fr_Ar_COO</option>
            <option value="fr_Ar_N">fr_Ar_N</option>
            <option value="fr_Ar_NH">fr_Ar_NH</option>
            <option value="fr_Ar_OH">fr_Ar_OH</option>
            <option value="fr_COO">fr_COO</option>
            <option value="fr_COO2">fr_COO2</option>
            <option value="fr_C_O">fr_C_O</option>
            <option value="fr_C_O_noCOO">fr_C_O_noCOO</option>
            <option value="fr_C_S">fr_C_S</option>
            <option value="fr_HOCCN">fr_HOCCN</option>
            <option value="fr_Imine">fr_Imine</option>
            <option value="fr_NH0">fr_NH0</option>
            <option value="fr_NH1">fr_NH1</option>
            <option value="fr_NH2">fr_NH2</option>
            <option value="fr_N_O">fr_N_O</option>
            <option value="fr_Ndealkylation1">fr_Ndealkylation1</option>
            <option value="fr_Ndealkylation2">fr_Ndealkylation2</option>
            <option value="fr_Nhpyrrole">fr_Nhpyrrole</option>
            <option value="fr_SH">fr_SH</option>
            <option value="fr_aldehyde">fr_aldehyde</option>
            <option value="fr_alkyl_carbamate">fr_alkyl_carbamate</option>
            <option value="fr_alkyl_halide">fr_alkyl_halide</option>
            <option value="fr_allylic_oxid">fr_allylic_oxid</option>
            <option value="fr_amide">fr_amide</option>
            <option value="fr_amidine">fr_amidine</option>
            <option value="fr_aniline">fr_aniline</option>
            <option value="fr_aryl_methyl">fr_aryl_methyl</option>
            <option value="fr_azide">fr_azide</option>
            <option value="fr_azo">fr_azo</option>
            <option value="fr_barbitur">fr_barbitur</option>
            <option value="fr_benzene">fr_benzene</option>
            <option value="fr_benzodiazepine">fr_benzodiazepine</option>
            <option value="fr_bicyclic">fr_bicyclic</option>
            <option value="fr_diazo">fr_diazo</option>
            <option value="fr_dihydropyridine">fr_dihydropyridine</option>
            <option value="fr_epoxide">fr_epoxide</option>
            <option value="fr_ester">fr_ester</option>
            <option value="fr_ether">fr_ether</option>
            <option value="fr_furan">fr_furan</option>
            <option value="fr_guanido">fr_guanido</option>
            <option value="fr_halogen">fr_halogen</option>
            <option value="fr_hdrzine">fr_hdrzine</option>
            <option value="fr_hdrzone">fr_hdrzone</option>
            <option value="fr_imidazole">fr_imidazole</option>
            <option value="fr_imide">fr_imide</option>
            <option value="fr_isocyan">fr_isocyan</option>
            <option value="fr_isothiocyan">fr_isothiocyan</option>
            <option value="fr_ketone">fr_ketone</option>
            <option value="fr_ketone_Topliss">fr_ketone_Topliss</option>
            <option value="fr_lactam">fr_lactam</option>
            <option value="fr_lactone">fr_lactone</option>
            <option value="fr_methoxy">fr_methoxy</option>
            <option value="fr_morpholine">fr_morpholine</option>
            <option value="fr_nitrile">fr_nitrile</option>
            <option value="fr_nitro">fr_nitro</option>
            <option value="fr_nitro_arom">fr_nitro_arom</option>
            <option value="fr_nitro_arom_nonortho">fr_nitro_arom_nonortho</option>
            <option value="fr_nitroso">fr_nitroso</option>
            <option value="fr_oxazole">fr_oxazole</option>
            <option value="fr_oxime">fr_oxime</option>
            <option value="fr_para_hydroxylation">fr_para_hydroxylation</option>
            <option value="fr_phenol">fr_phenol</option>
            <option value="fr_phenol_noOrthoHbond">fr_phenol_noOrthoHbond</option>
            <option value="fr_phos_acid">fr_phos_acid</option>
            <option value="fr_phos_ester">fr_phos_ester</option>
            <option value="fr_piperdine">fr_piperdine</option>
            <option value="fr_piperzine">fr_piperzine</option>
            <option value="fr_priamide">fr_priamide</option>
            <option value="fr_prisulfonamd">fr_prisulfonamd</option>
            <option value="fr_pyridine">fr_pyridine</option>
            <option value="fr_quatN">fr_quatN</option>
            <option value="fr_sulfide">fr_sulfide</option>
            <option value="fr_sulfonamd">fr_sulfonamd</option>
            <option value="fr_sulfone">fr_sulfone</option>
            <option value="fr_term_acetylene">fr_term_acetylene</option>
            <option value="fr_tetrazole">fr_tetrazole</option>
            <option value="fr_thiazole">fr_thiazole</option>
            <option value="fr_thiocyan">fr_thiocyan</option>
            <option value="fr_thiophene">fr_thiophene</option>
            <option value="fr_unbrch_alkane">fr_unbrch_alkane</option>
            <option value="fr_urea">fr_urea</option>
            <option value="qed">qed</option>
        </param>
    </inputs>
    <outputs>
        <data format="tabular" name="outfile" />
    </outputs>
    <tests>
        <test>
            <param name="infile" ftype='sdf' value="CID_3037.sdf" />
            <param name="header" value="True" />
            <output name="outfile" ftype='tabular' file="rdkit_descriptors_result1.tab" />
        </test>
        <test>
            <param name="infile" ftype='sdf' value="CID_3037.sdf" />
            <param name="header" value="True" />
            <param name="select_multiple" value="qed,FormalCharge,MolWt" />
            <output name="outfile" ftype='tabular' file="rdkit_descriptors_subset.tab" />
        </test>
        <test>
            <param name="infile" ftype='pdb' value="mol.pdb" />
            <param name="header" value="False" />
            <param name="select_multiple" value="FormalCharge" />
            <output name="outfile" ftype='tabular' file="mol_pdb_charges.tab" />
        </test>
    </tests>
    <help>
<![CDATA[

.. class:: infomark

**What this tool does**

| RDKit is an open source toolkit for cheminformatics and machine learning.
| This tool calculates various molecular descriptors for the input data. The table below shows a brief overview of the descriptors (or families of related descriptors) which are calculated.
|
| - Gasteiger/Marsili Partial Charges
| - BalabanJ
| - BertzCT
| - Ipc
| - HallKierAlpha
| - Kappa1 - Kappa3
| - Chi0, Chi1
| - Chi0n - Chi4n
| - Chi0v - Chi4v
| - FormalCharge
| - MolLogP
| - MolMR
| - MolWt
| - HeavyAtomCount
| - HeavyAtomMolWt
| - NHOHCount
| - NOCount
| - NumHAcceptors
| - NumHDonors
| - NumHeteroatoms
| - NumRotatableBonds
| - NumValenceElectrons
| - RingCount
| - SSSR
| - TPSA
| - LabuteASA
| - PEOE_VSA1 - PEOE_VSA14
| - SMR_VSA1 - SMR_VSA10
| - SlogP_VSA1 - SlogP_VSA12
| - EState_VSA1 - EState_VSA11
| - VSA_EState1 - VSA_EState10
| - Topliss fragments
|
| A full list of the descriptors can be obtained here_.

.. _here: https://rdkit.readthedocs.org/en/latest/GettingStartedInPython.html#list-of-available-descriptors

-----

.. class:: warningmark

**Hint**

Use the **cut columns from a table** tool to select just the desired descriptors. Alternatively, if you just want to calculate a subset of the descriptors, there is an option available to do so.

-----

.. class:: infomark

**Input**

| - `SDF Format`_
| - `SMILES Format`_
| - `Corina MOL2`_
| - PDB
| - InChi

.. _SDF Format: http://en.wikipedia.org/wiki/Chemical_table_file
.. _SMILES Format: http://en.wikipedia.org/wiki/Simplified_molecular_input_line_entry_specification
.. _Corina MOL2: http://www.molecular-networks.com/products/corina

-----

.. class:: infomark

 **Output**

Tabular file, where each descriptor (value) is shown in a separate column.


]]>
    </help>
    <citations>
        <citation type="bibtex">
            @article{rdkit,
                author = {Greg Landrum and others},
                title = {RDKit: Open-source cheminformatics},
                url ={http://www.rdkit.org}
            }</citation>
    </citations>
</tool>