panoply_clumps_ptm

Description

This module runs Clumps-PTM, a top-down spatial-proteomics analysis tool that identifies proteins with nearby-clusters of differentially-regulated PTM-sites (phosphorylation, acetylation, and/or ubiquitination). The algorithm was adapted from the CLUMPS method for detecting clusters of mutations in 3D protein structures; it calculates a weighted average proximity score across all differentially-modified residue pairs in a given protein, with weights given according to logFC and significance. An empirical p-value is calculated by permuting across the possible PTM-sites within the protein, before correction for multiple-testing. A full description of the algorithm can be found in the Method Details of Geffen et al. 2023.

Input

Required inputs:

• diff_exp_file: (.tsv file) results file from panoply_clumps_ptm_diffexp, containing differential expression results for all PTM -omes, for a given annotation

• var_sites_file: (.tsv file) filtered mapping file (filt_results) from panoply_clumps_ptm_mapping, containing all varaible sites with valid PDB coordinates

• PDB_ref_bucket: (String) Google-Cloud Bucket containing a tarred copy of the PDB structural archive (i.e. https://files.wwpdb.org/pub/pdb/data/structures/divided/pdb/). A public bucket, pulled from a frozen 2025 snapshot, can be found at: "gs://fc-385e9b4e-43ff-44b3-8cf7-036a2a96d102/pdbs_2025_tars/"

  • PDB_DIR: Internal parameter listing the files to import from PDB_ref_bucket

• output_prefix: (String, default="results") prefix used to name the output tar file

• yaml_file: (.yaml file) master-parameters.yaml

Optional inputs:

• run_combined: (Boolean, default=true) if TRUE analysis will be run on all PTM datasets combined, in addition to each -ome separately

• weight_col: (String, default="logFC") column from differential-expression dataset to use as weights in ClumpsPTM

• accession_col: (String, default="description") GCT rdesc column with protein accession IDs; must use the same ID type as the provided FASTA_ref_file file.

• variable_sites_col: (String, default="variableSites") GCT rdesc column with PTM variable site(s) (e.g. 'T527t')

• DEBUG_MODE: (Boolean, default=false) Debugging toggle; if true, a small subset of proteins will be analyzed. Should be turned off for analysis.

Output

• results: (.tar file)

References

1. Geffen, Y. et al. Pan-cancer analysis of post-translational modifications reveals shared patterns of protein regulation. Cell 186, 3945-3967.e26 (2023).