SeekNSee: A Rule-Based Framework for Prioritizing Proteotypic Peptides

SeekNSee is a locally executed R/Shiny application designed to enhance the reproducibility and auditability of peptide-candidate screening. This framework addresses the challenge of selecting peptide candidates for target proteomics, especially when users lack prior experimental data, programming skills, or stable access to external repositories. SeekNSee integrates a multicriteria problem to improve the screening process through enzyme-aware in silico digestion.

The application begins with a user-provided reference proteome and target sequences. It then generates peptides-candidates using nine digestion configurations and evaluates them with 12 rules, which are configured as ignore, apply, or block. SeekNSee stores numerical scores and the retained or blocking status of each peptide candidate independently. Every candidate-rule outcome is preserved in a rule-level trace, ensuring comprehensive data integrity and decision traceability.

Comprehensive computational validation confirmed the specified behavior in controlled cases, with 198 passing automated assertions. The application demonstrated successful operational and processing tests on both Windows and Ubuntu operating systems. A proteome-wide application covered all 20,416 reviewed human proteins, generating a substantial number of peptide-candidate and unique peptide sequences (733,864 and 581,026, respectively) while maintaining data integrity and preserving candidate-specific context.

Benchmarking against dia, dda, and deep-fractionated proteomics showed strong associations between SeekNSee score categories, peptide observations, and multiple analytical measurements. These results demonstrate that SeekNSee provides a transparent and traceable screening layer for generating peptide-candidates, paving the way for subsequent assay and experimental validation.