Glossary¶
Key terms in pan-cancer phosphoproteomics, organized alphabetically.
Activity inference — Estimating the activity level of a kinase from the phosphorylation states of its known substrates. Distinct from substrate prediction.
Annotation bottleneck — The gap between the number of observed phosphosites (~240K) and those with known kinase or function (<5%). The central limitation of computational phosphoproteomics.
CPTAC — Clinical Proteomic Tumor Analysis Consortium. NCI-funded program producing matched proteomics, phosphoproteomics, and genomics across cancer types.
DDA (Data-Dependent Acquisition) — MS acquisition mode selecting the top-N most abundant precursor ions per cycle for fragmentation. Deeper per-run but less reproducible across samples than DIA.
DIA (Data-Independent Acquisition) — MS acquisition mode fragmenting all precursors within defined m/z windows. Better quantitative reproducibility but computationally demanding.
decoupleR — Unified framework for inferring kinase, transcription factor, and pathway activities from omics data. Wraps multiple statistical methods (ULM, VIPER, GSVA) with a consistent API.
Enrichment analysis — Statistical test asking whether substrates of a kinase are over-represented among differentially phosphorylated sites. The basis of KSEAapp and KEA3.
funscoR — Random forest classifier predicting functional relevance of phosphosites from evolutionary, structural, and regulatory features.
IMAC — Immobilized Metal Affinity Chromatography. Phosphopeptide enrichment using Fe3+ or Ti4+ metal ions. The standard enrichment in CPTAC workflows.
ILP (Integer Linear Programming) — Optimization framework used by CARNIVAL and PHONEMeS to find the most parsimonious signaling network consistent with observed phosphosite changes.
Kinase — Enzyme that transfers a phosphate group from ATP to a substrate protein. The human genome encodes 518 kinases (the "kinome").
Kinase-substrate relationship — A validated connection between a kinase and a specific phosphosite it phosphorylates. Fewer than 10,000 such pairs are curated in PhosphoSitePlus.
KinaseLibrary — Positional scanning peptide array data for 303 Ser/Thr kinases. The most comprehensive motif reference for kinase specificity (Johnson et al., Nature 2023).
Kinome — The complete set of protein kinases encoded in a genome. The human kinome contains 518 kinases.
Localization probability — Confidence score for assigning a phospho group to a specific residue within a peptide. Sites with probability >0.75 are typically considered localized.
Motif — The amino acid sequence pattern surrounding a phosphorylation site that determines kinase specificity. Example: CDK substrates prefer [S/T]-P-x-[K/R].
OmniPath — Meta-resource integrating 100+ databases of signaling interactions. The prior knowledge backbone for pathway reconstruction tools.
Phosphatase — Enzyme that removes phosphate groups from proteins. The counterpart to kinases in reversible phosphorylation signaling.
Phosphoproteomics — Mass spectrometry-based measurement of phosphorylation states across thousands of proteins simultaneously.
Phosphosite — A specific amino acid residue (Ser, Thr, or Tyr) on a protein that is phosphorylated. Identified by protein name and residue number (e.g., AKT1-S473).
PhosphoSitePlus — The primary curated database of phosphorylation sites and kinase-substrate relationships. Contains ~240,000 human phosphosites.
Prior knowledge network — A graph of known protein-protein interactions and regulatory relationships used as input for pathway reconstruction. OmniPath is the standard source.
Ratio compression — Systematic underestimation of fold changes in TMT-based quantification due to co-isolation interference. Affects pan-cancer phosphoproteomics datasets.
Stoichiometry — The fraction of a protein population that is phosphorylated at a given site. Rarely measured in standard phosphoproteomics; confounds abundance-based analyses.
Substrate — A protein that is phosphorylated by a kinase. One kinase may have hundreds of substrates; one substrate may be phosphorylated by multiple kinases.
TMT (Tandem Mass Tag) — Isobaric labeling reagent enabling multiplexed quantification of 11-18 samples per MS run. The backbone of CPTAC datasets.
TiO2 — Titanium dioxide. Used for phosphopeptide enrichment; favors peptides with acidic residues near the phosphosite.