Mercurial > repos > iracooke > mascot
changeset 7:75184e45bf84 draft
Uploaded
| author | iracooke |
|---|---|
| date | Fri, 13 Jun 2014 18:26:31 -0400 |
| parents | f36b84d1ac09 |
| children | f567d1818b84 |
| files | README mascot.xml mascot_to_pepxml.xml repository_dependencies.xml tool-data/mascot_databases.loc.sample tool-data/mascot_mods.loc.sample tool-data/pepxml_databases.loc.sample |
| diffstat | 7 files changed, 403 insertions(+), 0 deletions(-) [+] |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/README Fri Jun 13 18:26:31 2014 -0400 @@ -0,0 +1,9 @@ +This includes tools for running Mascot MS/MS searches + +Running this tool requires a working installation of Mascot (commercial software). +Mascot is available from http://www.matrixscience.com/ + +Requirements: +This package uses protk which must be installed separately. + +For instructions please see: https://github.com/iracooke/protk/#galaxy-integration
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/mascot.xml Fri Jun 13 18:26:31 2014 -0400 @@ -0,0 +1,205 @@ +<tool id="proteomics_search_mascot_1" name="Mascot MSMS Search" version="1.0.2"> + + <requirements> + <requirement type="package" version="1.3">protk</requirement> + </requirements> + + <description>Mascot MS/MS Search</description> + + + <command>mascot_search.rb + + #if $database.source_select=="built_in": + -d $database.dbkey + #else + -d $database.custom_db + #end if + + -f $fragment_ion_tol + + -S $server $input_file + + -o $output + + -r + + ## Variable Mods + + --var-mods=' + $variable_mods + ' + + --fix-mods=' + $fixed_mods + ' + --allowed-charges=$allowed_charges + + --enzyme=$enzyme + + --instrument=$instrument + + --precursor-ion-tol-units=$precursor_tolu + + --email=$email + + -v $missed_cleavages + + #if $security.security_use + --use-security + --username $security.username + --password $security.password + #end if + + #if $proxy + --proxy $proxy + #end if + + </command> + + <inputs> + + <param name="input_file" type="data" format="mgf" multiple="false" label="MSMS File" help="A Mascot Generic Format file containing MSMS Spectra"/> + + <conditional name="database"> + <param name="source_select" type="select" label="Database Type"> + <option value="built_in">Built-In</option> + <option value="custom_defined">Custom</option> + </param> + <when value="built_in"> + <param name="dbkey" type="select" format="text" > + <label>Database</label> + <options from_file="mascot_databases.loc"> + <column name="name" index="0" /> + <column name="value" index="2" /> + </options> + </param> + </when> + <when value="custom_defined"> + <param name="custom_db" type="text" size="80" value="SwissProt" label="Database Name" help="Exact name of a database defined on the Mascot server (case-sensitive)"/> + </when> + </conditional> + + <param name="variable_mods" format="text" type="select" multiple="true" label="Variable Modifications" help="Multiple values allowed"> + <options from_file="mascot_mods.loc"> + <column name="name" index="0" /> + <column name="value" index="2" /> + </options> + </param> + + <param name="fixed_mods" format="text" type="select" multiple="true" label="Fixed Modifications" help="Multiple values allowed"> + <options from_file="mascot_mods.loc"> + <column name="name" index="0" /> + <column name="value" index="2" /> + </options> + </param> + + <param name="missed_cleavages" type="select" format="text"> + <label>Missed Cleavages Allowed</label> + <option value="0">0</option> + <option value="1">1</option> + <option value="2">2</option> + </param> + + <param name="enzyme" type="select" format="text"> + <label>Enzyme</label> + <option value="Trypsin">Trypsin</option> + <option value="Trypsin/P">Trypsin/P</option> + <option value="Arg-C">Arg-C</option> + <option value="Asp-N">Asp-N</option> + <option value="Asp-N_ambic">Asp-N_ambic</option> + <option value="Chymotrypsin">Chymotrypsin</option> + <option value="CNBr">CNBr</option> + <option value="CNBr+Trypsin">CNBr+Trypsin</option> + <option value="Formic_acid">Formic_acid</option> + <option value="Lys-C">Lys-C</option> + <option value="Lys-C/P">Lys-C/P</option> + <option value="LysC+AspN">LysC+AspN</option> + <option value="Lys-N">Lys-N</option> + <option value="PepsinA">PepsinA</option> + <option value="semiTrypsin">semiTrypsin</option> + <option value="TrypChymo">TrypChymo</option> + <option value="TrypsinMSIPI">TrypsinMSIPI</option> + <option value="TrypsinMSIPI/P">TrypsinMSIPI/P</option> + <option value="V8-DE">V8-DE</option> + <option value="V8-E">V8-E</option> + <option value="none">none</option> + </param> + + <param name="allowed_charges" type="select" format="text"> + <label>Peptide Charge</label> + <option value="8-">8-</option> + <option value="7-">7-</option> + <option value="6-">6-</option> + <option value="5-">5-</option> + <option value="4-">4-</option> + <option value="3-">3-</option> + <option value="2-,3- and 4-">2-,3- and 4-</option> + <option value="2- and 3-">2- and 3-</option> + <option value="2-">2-</option> + <option value="1-,2- and 3-">1-,2- and 3-</option> + <option value="1-">1-</option> + <option value="Mr">Mr</option> + <option value="1+">1+</option> + <option value="1+, 2+ and 3+">1+,2+ and 3+</option> + <option value="2+">2+</option> + <option value="2+ and 3+">2+ and 3+</option> + <option value="2+,3+ and 4+">2+,3+ and 4+</option> + <option value="3+">3+</option> + <option value="4+">4+</option> + <option value="5+">5+</option> + <option value="6+">6+</option> + <option value="7+">7+</option> + <option value="8+">8+</option> + </param> + + <param name="instrument" type="select" format="text"> + <label>Instrument</label> + <option value="ESI-QUAD-TOF">ESI-QUAD-TOF</option> + <option value="MALDI-TOF-PSD">MALDI-TOF-PSD</option> + <option value="ESI-TRAP">ESI-TRAP</option> + <option value="ESI-QUAD">ESI-QUAD</option> + <option value="ESI-FTICR">ESI-FTICR</option> + <option value="MALDI-TOF-TOF">MALDI-TOF-TOF</option> + <option value="ESI-4SECTOR">ESI-4SECTOR</option> + <option value="FTMS-ECD">FTMS-ECD</option> + <option value="ETD-TRAP">ETD-TRAP</option> + <option value="MALDI-QUAD-TOF">MALDI-QUAD-TOF</option> + <option value="MALDI-QIT-TOF">MALDI-QIT-TOF</option> + <option value="MALDI-ISD">MALDI-ISD</option> + <option value="CID+ETD">CID+ETD</option> + </param> + + <param name="fragment_ion_tol" label="Fragment Ion Tolerance" type="float" value="0.65" min="0" max="10000" help="Fragment ion tolerance in Daltons"/> + + <param name="precursor_ion_tol" label="Precursor Ion Tolerance (Da or ppm)" type="float" value="100" min="0" max="10000" help="Enter a value in Daltons or ppm depending on the units chosen below"/> + <param name="precursor_tolu" type="select" format="text"> + <label>Precursor Ion Tolerance Units</label> + <option value="ppm">ppm</option> + <option value="Da">Da</option> + </param> + + <param name="server" type="text" label="URL to the cgi directory on the Mascot Server " size="60" value="http://www.exampleserver.com/mascot/cgi/"/> + <param name="proxy" type="text" label="Proxy Server URL including proxy port" size="60" value="" help="eg http://proxy.latrobe.edu.au:8080"/> + + <conditional name="security"> + <param name="security_use" type="boolean" label="Use Mascot Security?" help="Select this if you need to enter a username and password" truevalue="true" falsevalue="false" /> + <when value="false" /> + <when value="true"> + <param name="username" type="text" label="Username" size="60" value="" help="Username on Mascot Server"/> + <param name="password" type="text" label="Password" size="60" value="" help="Mascot Password. Password is encrypted when over the internet but is stored in clear-text on the Galaxy server"/> + </when> + </conditional> + <param name="email" type="text" label="Email " size="60" value="" help=""/> + + + + </inputs> + <outputs> + <data format="mascotdat" name="output" metadata_source="input_file" label="mascot_vs_${database.dbkey if $database.has_key('dbkey') else $database.custom_db}.${input_file.display_name}.mascotdat"/> + </outputs> + + <help> + Run a Mascot MS/MS Ion Search + </help> + +</tool>
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/mascot_to_pepxml.xml Fri Jun 13 18:26:31 2014 -0400 @@ -0,0 +1,82 @@ +<tool id="mascot_to_pepxml_1" name="Mascot to pepXML" version="1.0.2"> + <requirements> + <requirement type="package" version="1.3">protk</requirement> + <requirement type="package" version="4.6.3">trans_proteomic_pipeline</requirement> + </requirements> + + <description>Converts a mascot results file to pepXML</description> + +<command>mascot_to_pepxml.rb + $input_file + + -o $output + + #if $database.source_select=="built_in": + -d $database.dbkey + #else + -d $database.fasta_file + #end if + + #if $explicit_enzyme.explicit_enzyme_use + --enzyme $explicit_enzyme.enzyme + #end if + + $shortid + +</command> +<inputs> + + + <param name="input_file" type="data" format="mascotdat" multiple="false" label="Input File" help="Mascot results file"/> + + <conditional name="database"> + <param name="source_select" type="select" label="Database source" help="A local copy of the database used in the Mascot search"> + <option value="built_in">Built-In</option> + <option value="input_ref">Uploaded fasta file</option> + </param> + <when value="built_in"> + <param name="dbkey" type="select" format="text" > + <label>Database</label> + <options from_file="pepxml_databases.loc"> + <column name="name" index="0" /> + <column name="value" index="2" /> + </options> + </param> + </when> + <when value="input_ref"> + <param name="fasta_file" type="data" format="fasta" label="Uploaded FASTA file" /> + </when> + </conditional> + + <param name="shortid" type="boolean" label="Short ID" help="Use protein id from the Mascot result file instead reading from the fasta database." truevalue="--shortid" falsevalue="" /> + + <conditional name="explicit_enzyme"> + <param name="explicit_enzyme_use" type="boolean" label="Specify Enzyme" help="If left unchecked the Enzyme will be read from the input file" truevalue="true" falsevalue="false" /> + <when value="false" /> + <when value="true"> + <param name="enzyme" type="text" label="Enzyme" size="80" value="trypsin" help="Semi-cleavage can be specified as semisample_enyzme eg semitrypsin"/> + </when> + </conditional> + + + +</inputs> +<outputs> + <data format="raw_pepxml" metadata_source="input_file" name="output" label="${input_file.display_name}.pepXML" /> +</outputs> + + <tests> + <test> + <param name="source_select" value="input_ref"/> + <param name="fasta_file" value="bsa.fasta"/> + <param name="input_file" value="F002832.dat"/> + <output name="output" file="bsa_mascot2xml.pep.xml" compare="sim_size" delta="600" /> + </test> + + </tests> + +<help> + Convert mascot results from mascotdat to pepXML +</help> + +</tool>
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/repository_dependencies.xml Fri Jun 13 18:26:31 2014 -0400 @@ -0,0 +1,4 @@ +<?xml version="1.0"?> +<repositories description="Proteomics datatypes"> + <repository changeset_revision="f66f8ca7b7b9" name="proteomics_datatypes" owner="iracooke" toolshed="http://toolshed.g2.bx.psu.edu" /> + </repositories>
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/tool-data/mascot_databases.loc.sample Fri Jun 13 18:26:31 2014 -0400 @@ -0,0 +1,13 @@ +#This file lists the names of protein databases installed on Mascot +# +#In order to use interprophet to combine results from different search engines +#it is important that all searches are performed on the same database +#you should therefore ensure that each database installed on mascot has an equivalent +#database installed in the Protk databases directory (databases used by omssa and x!tandem) +#the mascot_to_pepxml tool will ask for this database when performing the conversion. +# +# Entries should follow the be structured as follows +# Display_name dbkey dbNameOnMascot dbkey +# +Swissprot spall_ SwissProt spall_ +Swissprot Human sphuman_ SPHuman sphuman_ \ No newline at end of file
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/tool-data/mascot_mods.loc.sample Fri Jun 13 18:26:31 2014 -0400 @@ -0,0 +1,77 @@ +#This file lists the names of chemical modifications acceptable for proteomics search engines +# +# +Acetyl (K) acetyl_k_ Acetyl (K) acetyl_k_ +Acetyl (N-term) acetyl_n-term_ Acetyl (N-term) acetyl_n-term_ +Acetyl (Protein N-term) acetyl_proteinn-term_ Acetyl (Protein N-term) acetyl_proteinn-term_ +Amidated (C-term) amidated_c-term_ Amidated (C-term) amidated_c-term_ +Amidated (Protein C-term) amidated_proteinc-term_ Amidated (Protein C-term) amidated_proteinc-term_ +Ammonia-loss (N-term C) ammonia-loss_n-termc_ Ammonia-loss (N-term C) ammonia-loss_n-termc_ +Biotin (K) biotin_k_ Biotin (K) biotin_k_ +Biotin (N-term) biotin_n-term_ Biotin (N-term) biotin_n-term_ +Carbamidomethyl (C) carbamidomethyl_c_ Carbamidomethyl (C) carbamidomethyl_c_ +Carbamyl (K) carbamyl_k_ Carbamyl (K) carbamyl_k_ +Carbamyl (N-term) carbamyl_n-term_ Carbamyl (N-term) carbamyl_n-term_ +Carboxymethyl (C) carboxymethyl_c_ Carboxymethyl (C) carboxymethyl_c_ +Cation:Na (C-term) cation_na_c-term_ Cation:Na (C-term) cation_na_c-term_ +Cation:Na (DE) cation_na_de_ Cation:Na (DE) cation_na_de_ +Deamidated (NQ) deamidated_nq_ Deamidated (NQ) deamidated_nq_ +Deamidated-N (N) deamidated-n_n_ Deamidated-N (N) deamidated-n_n_ +Dehydrated (N-term C) dehydrated_n-termc_ Dehydrated (N-term C) dehydrated_n-termc_ +Dehydro (C) dehydro_c_ Dehydro (C) dehydro_c_ +Dioxidation (M) dioxidation_m_ Dioxidation (M) dioxidation_m_ +Ethanolyl (C) ethanolyl_c_ Ethanolyl (C) ethanolyl_c_ +ExacTagAmine (K) exactagamine_k_ ExacTagAmine (K) exactagamine_k_ +ExacTagThiol (C) exactagthiol_c_ ExacTagThiol (C) exactagthiol_c_ +Formyl (N-term) formyl_n-term_ Formyl (N-term) formyl_n-term_ +Formyl (Protein N-term) formyl_proteinn-term_ Formyl (Protein N-term) formyl_proteinn-term_ +Gln->pyro-Glu (N-term Q) gln_pyro-glu_n-termq_ Gln->pyro-Glu (N-term Q) gln_pyro-glu_n-termq_ +Glu->pyro-Glu (N-term E) glu_pyro-glu_n-terme_ Glu->pyro-Glu (N-term E) glu_pyro-glu_n-terme_ +Guanidinyl (K) guanidinyl_k_ Guanidinyl (K) guanidinyl_k_ +ICAT-C (C) icat-c_c_ ICAT-C (C) icat-c_c_ +ICAT-C:13C(9) (C) icat-c_13c_9__c_ ICAT-C:13C(9) (C) icat-c_13c_9__c_ +ICPL (K) icpl_k_ ICPL (K) icpl_k_ +ICPL (Protein N-term) icpl_proteinn-term_ ICPL (Protein N-term) icpl_proteinn-term_ +ICPL:13C(6) (K) icpl_13c_6__k_ ICPL:13C(6) (K) icpl_13c_6__k_ +ICPL:13C(6) (Protein N-term) icpl_13c_6__proteinn-term_ ICPL:13C(6) (Protein N-term) icpl_13c_6__proteinn-term_ +ICPL:13C(6)2H(4) (K) icpl_13c_6_2h_4__k_ ICPL:13C(6)2H(4) (K) icpl_13c_6_2h_4__k_ +ICPL:13C(6)2H(4) (N-term) icpl_13c_6_2h_4__n-term_ ICPL:13C(6)2H(4) (N-term) icpl_13c_6_2h_4__n-term_ +ICPL:13C(6)2H(4) (Protein N-term) icpl_13c_6_2h_4__proteinn-term_ ICPL:13C(6)2H(4) (Protein N-term) icpl_13c_6_2h_4__proteinn-term_ +ICPL:2H(4) (K) icpl_2h_4__k_ ICPL:2H(4) (K) icpl_2h_4__k_ +ICPL:2H(4) (Protein N-term) icpl_2h_4__proteinn-term_ ICPL:2H(4) (Protein N-term) icpl_2h_4__proteinn-term_ +iTRAQ4plex (K) itraq4plex_k_ iTRAQ4plex (K) itraq4plex_k_ +iTRAQ4plex (N-term) itraq4plex_n-term_ iTRAQ4plex (N-term) itraq4plex_n-term_ +iTRAQ4plex (Y) itraq4plex_y_ iTRAQ4plex (Y) itraq4plex_y_ +iTRAQ8plex (K) itraq8plex_k_ iTRAQ8plex (K) itraq8plex_k_ +iTRAQ8plex (N-term) itraq8plex_n-term_ iTRAQ8plex (N-term) itraq8plex_n-term_ +iTRAQ8plex (Y) itraq8plex_y_ iTRAQ8plex (Y) itraq8plex_y_ +Label:18O(1) (C-term) label_18o_1__c-term_ Label:18O(1) (C-term) label_18o_1__c-term_ +Label:18O(2) (C-term) label_18o_2__c-term_ Label:18O(2) (C-term) label_18o_2__c-term_ +Met->Hse (C-term M) met_hse_c-termm_ Met->Hse (C-term M) met_hse_c-termm_ +Met->Hsl (C-term M) met_hsl_c-termm_ Met->Hsl (C-term M) met_hsl_c-termm_ +Methyl (C-term) methyl_c-term_ Methyl (C-term) methyl_c-term_ +Methyl (DE) methyl_de_ Methyl (DE) methyl_de_ +Methylthio (C) methylthio_c_ Methylthio (C) methylthio_c_ +mTRAQ (K) mtraq_k_ mTRAQ (K) mtraq_k_ +mTRAQ (N-term) mtraq_n-term_ mTRAQ (N-term) mtraq_n-term_ +mTRAQ (Y) mtraq_y_ mTRAQ (Y) mtraq_y_ +mTRAQ:13C(3)15N(1) (K) mtraq_13c_3_15n_1__k_ mTRAQ:13C(3)15N(1) (K) mtraq_13c_3_15n_1__k_ +mTRAQ:13C(3)15N(1) (N-term) mtraq_13c_3_15n_1__n-term_ mTRAQ:13C(3)15N(1) (N-term) mtraq_13c_3_15n_1__n-term_ +mTRAQ:13C(3)15N(1) (Y) mtraq_13c_3_15n_1__y_ mTRAQ:13C(3)15N(1) (Y) mtraq_13c_3_15n_1__y_ +NIPCAM (C) nipcam_c_ NIPCAM (C) nipcam_c_ +Oxidation (HW) oxidation_hw_ Oxidation (HW) oxidation_hw_ +Oxidation (M) oxidation_m_ Oxidation (M) oxidation_m_ +Phospho (ST) phospho_st_ Phospho (ST) phospho_st_ +Phospho (Y) phospho_y_ Phospho (Y) phospho_y_ +Propionamide (C) propionamide_c_ Propionamide (C) propionamide_c_ +Pyridylethyl (C) pyridylethyl_c_ Pyridylethyl (C) pyridylethyl_c_ +Pyro-carbamidomethyl (N-term C) pyro-carbamidomethyl_n-termc_ Pyro-carbamidomethyl (N-term C) pyro-carbamidomethyl_n-termc_ +Sulfo (S) sulfo_s_ Sulfo (S) sulfo_s_ +Sulfo (T) sulfo_t_ Sulfo (T) sulfo_t_ +Sulfo (Y) sulfo_y_ Sulfo (Y) sulfo_y_ +TMT (K) tmt_k_ TMT (K) tmt_k_ +TMT (N-term) tmt_n-term_ TMT (N-term) tmt_n-term_ +TMT2plex (K) tmt2plex_k_ TMT2plex (K) tmt2plex_k_ +TMT2plex (N-term) tmt2plex_n-term_ TMT2plex (N-term) tmt2plex_n-term_ +TMT6plex (K) tmt6plex_k_ TMT6plex (K) tmt6plex_k_ +TMT6plex (N-term) tmt6plex_n-term_ TMT6plex (N-term) tmt6plex_n-term_
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/tool-data/pepxml_databases.loc.sample Fri Jun 13 18:26:31 2014 -0400 @@ -0,0 +1,13 @@ +#This file lists the names of protein databases installed locally in protk. +# These are used by omssa and x!tandem as well as the "mascot to pepxml" tool +# In order to combine search results with Interprophet searches must be run against an identical database +# +# Entries should follow the be structured as follows +# Display_name omssa_tandem_dbname dbkey +# +# +Swissprot spall_ SwissProt spall_ +Combined PlasmboDB (falciparum) and Swissprot Human plasmodb_pfalciparum_sphuman_ plasmodb_pfalciparum_sphuman plasmodb_pfalciparum_sphuman_ +Swissprot Human sphuman_ sphuman sphuman_ +Combined Swissprot/TRembl Human sptrhuman_ sptrhuman sptrhuman_ +Swissprot Mouse spmouse_ spmouse spmouse_