Data Analysis Modules: panoply_mimp

panoply_mimp

Description

This module utilizes the MIMP algorithm for predicting the impact of missense mutations on kinase-substrate phosphorylation (Wagih et al. 2015). This module takes in mutation, phosphoproteomic, and protein sequence data and prepares them as input to the rmimp R package. For each sample individually, missense mutations in close proximity to phosphosites (called phosphorylation-related SNV, or "pSNV") are first identified, and then kinase specificity models are used to predict if each pSNV disrupts an existing phosphosite or creates a new phosphosite. These changes to expected phosphosites are referred to as kinase rewiring events as they can potentially alter known kinase networks. This module returns the results from individual samples as well as a summary of results across the whole input dataset, including heatmaps to visualize results. If a given sample does not have any MIMP results, this means that either there were no pSNVs identified or none of the pSNVs gave rise to kinase rewiring events.

Input

Required inputs:

• mutation_file: (.maf file) mutation data table, must contain columns describing the mutation type (e.g. missense), the amino acid change and position, the sample identifier, and transcript identifier.
• phospho_file: (.gct file) phosphoproteomic data table, row metadata must include a columns for phosphosite and protein identifiers.
• fasta_file: (.fasta file) list of protein sequences.
• ids_file: (.Rdata file) mapping file between transcript and protein identifiers.
• master_yaml: (.yaml file) master parameters file.
• output_prefix: (String) prefix for naming the output tar file.

Optional inputs:

• groups_file_path: (.csv file, default = NULL) path to subset of sample annotations (groups file); if provided these annotations will be added as annotation tracks to output heatmaps.
• search_engine: (String, default = "SpectrumMill") name of search engine used to process phosphoproteomic data. Options are "SpectrumMill" or "other." If "SpectrumMill," the user does not have to specify values for phosphosite_col and protein_id_col below; if "other," the user must specify values for phosphosite_col and protein_id_col .
• phosphosite_col: (String, default = NULL) if search_engine = "other," provide name of column in phospho row metadata that indicates phosphosite position.
• protein_id_col: (String, default = NULL) if search_engine = "other," provide name of column in phospho row metadata that indicates protein identifier/accession number.
• mutation_AA_change_colname: (String, default = "Protein_Change") name of column in mutation MAF file that indicates amino acid change due to mutation.
• mutation_type_col: (String, default = "Variant_Classification") name of column in mutation MAF file that indicates type of mutation (e.g. missense).
• sample_id_col: (String, default = "Tumor_Sample_Barcode") name of column in mutation MAF file that indicates sample identifier.
• transcript_id_col: (String, default = "refseq_mrna_id") name of column in mutation MAF file that indicates transcript identifier.

Output

An output tar file called output_prefix_mimp_output.tar contains all results and summaries in a directory called "mimp_results_dir":

Results for individual samples are included in the "results_by_sample" sub-directory. Each sample has a further sub-directory labeled by sample identifier. These results include:

• In the "mutation_info" directory:
  • If the sample had mutations, full mutation data for the given sample (before any formatting): "sampleID_mutation_data_all.csv"
  • If any protein identifiers appeared in the phospho data that were not also in the fasta file, they were removed from analysis and listed here: "sampleID_phospho_identifiers_not_in_fasta.csv"
  • If there were any discrepancies between the reference amino acid in the mutation data and the amino acid at that position in the fasta file, they are listed here: "sampleID_mismatchedAA_mutation_vs_fasta.csv"
  • If there were any mutations that caused a gain of an S, T, or Y amino acid, they are listed here: "sampleID_STY_gain_mutation.csv"
  • If there were any mutations that caused a loss of an S, T, or Y amino acid, they are listed here: "sampleID_STY_loss_mutation.csv"
• In the "mimp_input" directory:
  • Formatted mutation data file used as input to the MIMP algorithm: "sampleID_mutation_mimp_input.csv"
  • Formatted phospho data file used as input to the MIMP algorithm: 01BR026_phospho_mimp_input.csv
• In the "mimp_results" directory:
  • If MIMP predicted any kinase rewiring events in the sample, full MIMP results for the sample: "sampleID_mimp_output_kinase_rewiring_events.csv"
  • If mimp predicted any kinase rewiring events in the sample, HTML report summarizing full MIMP results for the sample: "html/MIMP_results.html"

MIMP results for the whole dataset, if any, are included in the "mimp_results_dir" home directory:

• Concatenated list of full MIMP results from all samples: "mimp_output_kinase_rewiring_events_all.csv"
• Matrix of log-ratio results for all predicted kinase rewiring events by MIMP, indicating for each sample the likelihood and direction of altered kinase activity due to a specific point mutation: "kinase_rewiring_events_matrix_kinase_gene_mut_level.csv" (table); "mimp_results_predicted_kinase_rewiring_kinase_gene_mut_level_heatmap.pdf" and ".png" (heatmap)
• Matrix of log-ratio results summarizing overall altered kinase activity predicted by MIMP, indicating the maximum absolute value likelihood and direction of a kinase's altered activity in each sample: "kinase_rewiring_events_matrix_kinase_level.csv" (table); "mimp_results_predicted_kinase_rewiring_kinase_level_heatmap.pdf" and ".png" (heatmap)

Other files with mutation info for the full dataset in the "mimp_results_dir" home directory, if applicable:

• Concatenated list of all mutations causing a gain of an S, T, or Y amino acid in all samples: "all_STY_gain_mutation.csv"
• Concatenated list of all mutations causing a loss of an S, T, or Y amino acid in all samples: "all_STY_loss_mutation.csv"
• Concatenated list of all discrepancies between the reference amino acid in the mutation data and the amino acid at that position in the fasta file in all samples: "all_mismatchedAA_mutation_vs_fasta.csv"