AutoPRM: Method Preparation and Data Analysis for Internal Standard Triggered Parallel Reaction Monitoring (IS-PRM).
Titus aims to simplify preparation and analysis of IS-PRM experiments [1]. Supports analysis of survey runs to characterize internal standard peptides, preparation of methods, peak area ratio estimation and protein-level quantitation.
- AutoPRM requires Julia 1.10. Download julia and add it to the PATH.
- Open an instance of the julia REPL
- Type ";" to activate the shell from within the REPL. Then, navigate to the desired directory and clone the AutoPRM.jl repository.
shell> git clone https://github.com/nwamsley1/AutoPRM.jl.git
and not move into the package directory
shell> cd AutoPRM.jl
- Return to julia by hitting the backspace key. Activate the julia package manager by typing "]" into the REPL and enter the following:
(@v1.10) pkg> activate
(@v1.10) pkg> develop ./
(@v1.10) pkg> add ./
AutoPRM exports three methods "SearchPRM", "SearchSurvey", and "BuildSurvey". Each takes a single argument, that is a file path to a .json parameters files (examples included). To access these methods import the AutoPRM package from within the Julia REPL.
julia> using AutoPRM
AutoPRM requires Thermo .raw files be converted to an Apache .arrow format with a specific column specification. Use PioneerConverter to convert .raw files.
BuildSurvey constructs a table of precursors, m/z ratios, charge states, and intensity thresholds needed to build a SurveyMethod to analyze a new plate or aliquot of SIL peptides. This table can be directly loaded into the Thernmo method editor.
BuildSurvey requires takes a tab-delimited petide list file with at least two columns, "PROTEIN" and "PRECURSOR". The PROTEIN column is the Gene name, Uniprot ID, etc. that you want to associate with the PRECURSOR.
The PRECURSOR is the peptide sequence using one of the 20 standard single-letter amino-acid abbreviasions. On Mac/Linux, you can inspect an example file "HUMAN_IMMUNOLOGY_PEPTIDE_LIST_SEP_2023.txt" with the head command:
╭─[email protected] ~/TEST_DATA/SUREQUANT_JUL24/HUMAN_IMMUNOLOGY
╰─➤ head HUMAN_IMMUNOLOGY_PEPTIDE_LIST_SEP_2023.txt
PROTEIN PEPTIDE VENDOR ID WELL
ARG1 GGVEEGPTVLR Thermo NR102342.1 A1
ARG1 VMEETLSYLLGR Thermo NR102342.2 A2
AXL APLQGTLLGYR Thermo NR102342.3 A3
AXL TATITVLPQQPR Thermo NR102342.4 A4
B2M VNHVTLSQPK Thermo NR102342.5 A5
B2M VEHSDLSFSK Thermo NR102342.6 A6
CCL20 QLANEGCDINAIIFHTK Thermo NR102342.7 A7
CCL20 LSVCANPK Thermo NR102342.8 A8
CD163 EAEFGQGTGPIWLNEVK Thermo NR102342.9 A9
The parameter file it a text file in the .json format. It contains instructions for building the survey run. It often makes sense to split a single peptide panel into multiple survey runs, for example, one per charge state. However, for a short list of peptides, we can prepare a single run.
In this cas all the SIL pepetides in the mix are heavy, so we specify K[Hlys] and R[Harg] as fixed modifications. The peptide_list_path argument is an absolute path to the peptide list file (above). Example:
╭─[email protected] ~/TEST_DATA/SUREQUANT_JUL24/HUMAN_IMMUNOLOGY
╰─➤ cat params/build_survey_params.json
{
"fixed_mods":[
["C","C[Carb]"],
["K$","K[Hlys]"],
["R$","R[Harg]"]
],
"modification_masses":
{
"Carb":57.021464,
"Harg":10.008269,
"Hlys":8.014199
},
"charges": [2, 3, 4],
"peptide_list_path":"/Users/n.t.wamsley/TEST_DATA/SUREQUANT_JUL24/HUMAN_IMMUNOLOGY/HUMAN_IMMUNOLOGY_PEPTIDE_LIST_SEP_2023.txt"
}
Using the list of SIL peptides, we will start Julia, import the AuroPRM library, and run the "BuildSurvey" method providing a path to the parameter file. This will build a folder SurveyMethod containing SurveyMethod.csv that we can import into the Survey method file in the Thermo method editor.
╭─[email protected] ~/TEST_DATA/SUREQUANT_JUL24/HUMAN_IMMUNOLOGY
╰─➤ julia
julia> using AutoPRM
[ Info: Precompiling AutoPRM [161ade51-a71e-4c5f-b262-f1262fd3217d]
[ Info: Skipping precompilation since __precompile__(false). Importing AutoPRM [161ade51-a71e-4c5f-b262-f1262fd3217d].
┌ Warning: Replacing docs for `AutoPRM.PSM :: Union{}` in module `AutoPRM`
└ @ Base.Docs docs/Docs.jl:243
julia> BuildSurvey("params/build_survey_params.json")
"/Users/n.t.wamsley/TEST_DATA/SUREQUANT_JUL24/HUMAN_IMMUNOLOGY/SurveyMethod/SurveyMethod.csv"
julia> exit()
╭─[email protected] ~/TEST_DATA/SUREQUANT_JUL24/HUMAN_IMMUNOLOGY
╰─➤ head SurveyMethod/SurveyMethod.csv
Compound,Formula,Adduct,m/z,z,intensity_threshold
ARG1_GGVEEGPTVLR[Harg],,,562.3026884000001,2,10000.0
ARG1_GGVEEGPTVLR[Harg],,,375.20421773333334,3,10000.0
ARG1_GGVEEGPTVLR[Harg],,,281.6549824,4,10000.0
ARG1_VMEETLSYLLGR[Harg],,,710.8726283999999,2,10000.0
ARG1_VMEETLSYLLGR[Harg],,,474.25084439999995,3,10000.0
ARG1_VMEETLSYLLGR[Harg],,,355.9399523999999,4,10000.0
AXL_APLQGTLLGYR[Harg],,,599.8445284000001,2,10000.0
AXL_APLQGTLLGYR[Harg],,,400.2321110666667,3,10000.0
AXL_APLQGTLLGYR[Harg],,,300.4259024,4,10000.0
Move all survey runs into the 'survey_runs' folder. From the PioneerConverter directory, convert the .raw survey runs to .arrow format
╭─[email protected] ~/TEST_DATA/SUREQUANT_JUL24/NRF2_HP/PioneerConverter ‹master*›
╰─➤ dotnet run ../../HUMAN_IMMUNOLOGY/survey_runs/IMMUNOSIL_80nMSIL_1ugPEP_METHTEST_08022024_01.raw
Converting: IMMUNOSIL_80nMSIL_1ugPEP_METHTEST_08022024_01
batchSize: 10000
n_threads: 2
Starting Conversion For: IMMUNOSIL_80nMSIL_1ugPEP_METHTEST_08022024_01
Execution Time: 6844 ms for IMMUNOSIL_80nMSIL_1ugPEP_METHTEST_08022024_01
The parameter file cat params/search_survey_params.json contains the instructions for how to search the survey run. Make sure to change the ms_data_dir to the directory containing the .arrow formatted survey run. Also set peptide_list_path to HUMAN_IMMUNOLOGY_PEPTIDE_LIST_SEP_2023.txt, the same precursor list used to build the survey run
╭─[email protected] ~/TEST_DATA/SUREQUANT_JUL24/HUMAN_IMMUNOLOGY
╰─➤ cat params/search_survey_params.json
{
"right_precursor_tolerance": 0.001,
"left_precursor_tolerance": 0.001,
"precursor_rt_tolerance": 0.3,
"b_ladder_start": 3,
"y_ladder_start": 4,
"precursor_charges": [2, 3, 4],
"precursor_isotopes": [0],
"transition_charges": [1, 2],
"transition_isotopes": [0],
"fragment_match_ppm": 40,
"minimum_fragment_count": 5,
"fragments_to_select": 5,
"precursor_rt_window": 0.3,
"max_variable_mods": 2,
"fixed_mods":[
["C","C[Carb]"],
["K$","K[Hlys]"],
["R$","R[Harg]"]
],
"variable_mods":
[],
"modification_masses":
{
"Carb":57.021464,
"Harg":10.008269,
"Hlys":8.014199
},
"ms_file_conditions":
{
"_35NCE_":"35NCE",
"_40NCE_":"40NCE",
"GAPDH":"GAPDH"
},
"ms_data_dir": "/Users/n.t.wamsley/TEST_DATA/SUREQUANT_JUL24/HUMAN_IMMUNOLOGY/survey_runs/arrow_out",
"peptide_list_path": "/Users/n.t.wamsley/TEST_DATA/SUREQUANT_JUL24/HUMAN_IMMUNOLOGY/HUMAN_IMMUNOLOGY_PEPTIDE_LIST_SEP_2023.txt"
}
Searching the survey generates output in the same folder as the .arrow raw files. Outputs are iapi_method.csv, transition_list.csv, precursors_summary.csv, and a folder figures. The iapi_method.csv is an input for the Thermo IAPI SureQuant method. It specifies
the precursor m/z ratios, expected fragment ions, etc. To search an IS-PRM run, transition_list.csv is required as an input. precursors_summary.csv reports the best psm for each precursor accross all the survey runs searched. This can be useful for
combining the results of many surveys at different NCE or FAIMS CV values. The figures folder contains chromatograms and annotated spectra for the SIL peptides from the survey run.
julia> SearchSurvey("params/search_survey_params.json")
shell> head survey_runs/arrow_out/iapi_method.csv
protein_name,sequence,precursor_mz,precursor_charge,retention_time,precursor_intensity,hyperscore,NthIntensity,sumTopN,file_name,condition,transition_mz
ARG1,GGVEEGPTVLR[Harg],562.3026884000001,2,39.35995,7.8305648e7,51.076622009277344,9.583797e6,3.9467732e7,IMMUNOSIL_80nMSIL_1ugPEP_METHTEST_08022024_01.arrow,NONE,652.3972;781.4386;910.4792;595.376;214.11655
ARG1,VMEETLSYLLGR[Harg],710.8726283999999,2,61.726166,1.5517284e7,54.347801208496094,1.9420418e6,8.438351e6,IMMUNOSIL_80nMSIL_1ugPEP_METHTEST_08022024_01.arrow,NONE,718.4056;932.5331;1061.574;831.4879;631.37427
AXL,APLQGTLLGYR[Harg],599.8445284000001,2,51.279736,4.1923796e7,54.990875244140625,4.6274795e6,1.9375808e7,IMMUNOSIL_80nMSIL_1ugPEP_METHTEST_08022024_01.arrow,NONE,789.456;917.5154;1030.5983;518.30206;732.4348
AXL,TATITVLPQQPR[Harg],667.8869284000001,2,45.642937,2.6542056e7,56.771541595458984,5.0537445e6,1.9140718e7,IMMUNOSIL_80nMSIL_1ugPEP_METHTEST_08022024_01.arrow,NONE,635.3557;948.5567;748.43976;847.5089;274.14142
...
shell> head survey_runs/arrow_out/transition_list.csv
protein_name,sequence,precursor_charge,precursor_isotope,transition_names
ARG1,GGVEEGPTVLR[Harg],2,0,y6+1;y7+1;y8+1;y5+1;b3+1
ARG1,VMEETLSYLLGR[Harg],2,0,y6+1;y8+1;y9+1;y7+1;y5+1
AXL,APLQGTLLGYR[Harg],2,0,y7+1;y8+1;y9+1;y4+1;y6+1
AXL,TATITVLPQQPR[Harg],2,0,y5+1;y8+1;y6+1;y7+1;b3+1
...
shell> head survey_runs/arrow_out/precursors_summary.csv
protein_name,sequence,precursor_mz,precursor_charge,retention_time,precursor_intensity,hyperscore,NthIntensity,sumTopN,file_name,condition,transition_mz
FOXP3,HNLSLHK[Hlys],428.7475883999999,2,24.704,1.0979155e6,25.879701614379883,40209.586,186015.69,IMMUNOSIL_80nMSIL_1ugPEP_METHTEST_08022024_01.arrow,NONE,719.4211;565.30206;702.3848;365.18796;605.3863
CD8A,AAEGLDTQR[Harg],485.74501339999995,2,27.910803,2.7493786e7,44.79069137573242,3.184588e6,1.2822804e7,IMMUNOSIL_80nMSIL_1ugPEP_METHTEST_08022024_01.arrow,NONE,699.37195;828.4157;529.2649;642.3484;272.124
...
Just as with "SearchSurvey", move all survey runs into the 'survey_runs' folder. From the PioneerConverter directory, convert the .raw survey runs to .arrow format
The parameter file cat params/search_prm_params.json contains the instructions for how to search the survey run. Make sure to change the ms_data_dir to the directory containing the .arrow formatted survey run. You need to provide the transition_list.csv generated by SearchSurvey.
╭─[email protected] ~/TEST_DATA/SUREQUANT_JUL24/PYR_IMMUNO
╰─➤ cat params/search_prm_params.json
{
"right_precursor_tolerance": 0.5,
"left_precursor_tolerance": 0.5,
"precursor_rt_tolerance": 0.3,
"b_ladder_start": 3,
"y_ladder_start": 4,
"precursor_charges": [2, 3, 4],
"precursor_isotopes": [0],
"transition_charges": [1, 2],
"transition_isotopes": [0],
"fragment_match_ppm": 40,
"minimum_fragment_count": 5,
"fragments_to_select": 5,
"precursor_rt_window": 0.3,
"max_variable_mods": 2,
"fixed_mods":[
["C","C[Carb]"],
["K$","K[Hlys]"],
["R$","R[Harg]"]
],
"variable_mods":
[],
"modification_masses":
{
"Carb":57.021464,
"Harg":10.008269,
"Hlys":8.014199
},
"ms_file_conditions":
{
"_35NCE_":"35NCE",
"_40NCE_":"40NCE",
"GAPDH":"GAPDH"
},
"ms_data_dir":"/Users/n.t.wamsley/TEST_DATA/SUREQUANT_JUL24/PYR_IMMUNO/raw/arrow_out",
"transition_list_path":"/Users/n.t.wamsley/TEST_DATA/SUREQUANT_JUL24/PYR_IMMUNO/transition_list.csv"
}
Running SearchPRM generates two folders in the directory of the input raw files. These are tables and figures. figures includes chromatogram mirror plots of the light and heavy peptides. tables includes peptide level peak area ratios (PARs) and protien level abundances. The protein abundances are given on the base-2 log scale and determined from the peptide PAR values by the MaxLFQ algorithm.
julia> SearchPRM("params/search_prm.json")
peptide.csv
shell> head raw/arrow_out/tables/peptide.csv
ms_file_idx,sequence,protein_names,par,isotope,goodness_of_fit,file_name
7,HSHTLQEVK,GZMB,0.004318168575910173,light,0.05777005467620414,PYR_id90_IMMUNO_081024.arrow
4,HSHTLQEVK,GZMB,0.018577864396230517,light,0.2539605437419735,PYR_id72_IMMUNO_081024.arrow
3,HSHTLQEVK,GZMB,0.009270848940970923,light,0.2749798547215026,PYR_id68_IMMUNO_081024.arrow
6,HSHTLQEVK,GZMB,,light,,PYR_id86_IMMUNO_081024.arrow
5,HSHTLQEVK,GZMB,,light,,PYR_id82_IMMUNO_081024.arrow
protein.csv
head raw/arrow_out/tables/protein.csv 127 ↵
experiments,log2_abundance,peptides,protein,file_name
7,-7.491780529820246,HSHTLQEVK;VAQGIVSYGR,GZMB,PYR_id90_IMMUNO_081024.arrow
3,-7.634289477594166,VAQGIVSYGR,GZMB,PYR_id68_IMMUNO_081024.arrow
2,-7.534940994159977,VAQGIVSYGR,GZMB,PYR_id108_IMMUNO_081024.arrow
4,-8.252236268017503,AAEGLDTQR;TWNLGETVELK,CD8A,PYR_id72_IMMUNO_081024.arrow
3,-6.840801610882583,AAEGLDTQR;TWNLGETVELK,CD8A,PYR_id68_IMMUNO_081024.arrow
2,-5.923361587587841,AAEGLDTQR;TWNLGETVELK,CD8A,PYR_id108_IMMUNO_081024.arrow
5,-8.55081342862816,AAEGLDTQR;TWNLGETVELK,CD8A,PYR_id82_IMMUNO_081024.arrow
[1]
Gallien S, Kim SY, Domon B. Large-Scale Targeted Proteomics Using Internal Standard Triggered-Parallel Reaction Monitoring (IS-PRM). Mol Cell Proteomics. 2015 Jun;14(6):1630-44. doi: 10.1074/mcp.O114.043968. Epub 2015 Mar 9. PMID: 25755295; PMCID: PMC4458725.
[2]
Cox J, Hein MY, Luber CA, Paron I, Nagaraj N, Mann M. Accurate proteome-wide label-free quantification by delayed normalization and maximal peptide ratio extraction, termed MaxLFQ. Mol Cell Proteomics. 2014 Sep;13(9):2513-26. doi: 10.1074/mcp.M113.031591. Epub 2014 Jun 17. PMID: 24942700; PMCID: PMC4159666
[3]
Stopfer LE, Flower CT, Gajadhar AS, Patel B, Gallien S, Lopez-Ferrer D, White FM. High-Density, Targeted Monitoring of Tyrosine Phosphorylation Reveals Activated Signaling Networks in Human Tumors. Cancer Res. 2021 May 1;81(9):2495-2509. doi: 10.1158/0008-5472.CAN-20-3804. Epub 2021 Jan 28. PMID: 33509940; PMCID: PMC8137532.
[4]
Wamsley et al. Targeted proteomic quantitation of NRF2 signaling and predictive biomarkers in HNSCC. https://doi.org/10.1101/2023.03.13.532474