A Dual-Background Statistical Framework for Phosphoproteomics Highlights Intrinsic, High-Confidence Phosphorylation...

Distinguishing genuine kinase-substrate motifs from background noise is a growing challenge, as mass spectrometry (MS)-based global phosphoproteomics identifies a rapidly expanding set of phosphorylation sites. One of the major limitations is selecting an appropriate background model that systematically controls both technical and biological sources of bias. Although using the entire proteome as a background in a FASTA format considers the overall amino acid composition, it is still prone to biases from protein abundance and the uneven distribution of sequence space (particularly around low-abundance proteins). By contrast, internal background methods can control experiment-specific detection biases, but they may not fully capture residue-specific compositions or general trends in phosphorylation.

I develop a Dual-Background Enrichment (DBE) framework with a position-specific enrichment (PSE) strategy, which involves analyzing motif enrichment against two distinct background models: (1) A residue-heterogeneous internal background composed of phospho-motifs centered on the residue; e.g., phosphoserine (pS) motifs are tested relative to the pool of all detected phosphothreonine (pT) and phosphotyrosine (pY) motifs from the same experiment. (2) A FASTA background that includes all S, T, and Y residues in the UniProtKB proteome sequences.

Motifs are classified as high confidence if they meet statistical significance (q ≤ 0.05, fold enrichment > 1.5) against both background models.

By applying the DBE strategy to a large-scale phosphoproteomics dataset, we distinguish motifs driven by amino acid composition (enriched in FASTA background only) from those reflecting kinase substrate specificity (enriched in both backgrounds). This dual-reference approach reduces false positives arising from sequence composition bias and enriches high-confidence candidate kinase recognition motifs.