DECODE
Summary¶
DECODE is a reproducible proteogenomic workflow for identifying amino acid-substituted peptides that are not explained by matched genomic or transcriptomic sequence. It builds sample-specific protein databases, searches mass-spectrometry data for substitution-like mass shifts, applies orthogonal validation filters and quantifies substituted proteoforms with the ratio of amino acid substitution [1].
Workflow¶
DECODE integrates matched RNA sequencing, genomic sequencing and liquid chromatography-mass spectrometry data:
- Construct sample-specific protein databases: assemble transcripts, call sequence variants and translate coding sequences according to the genetic code.
- Discover candidate substitutions: search unmatched peptide spectra for systematic precursor and fragment-ion mass shifts consistent with a single amino acid substitution.
- Remove alternative explanations: exclude candidates explained by measured genomic variants, known post-translational modifications, alternative reading frames or protease-cleavage artefacts.
- Validate peptide sequences: require supporting fragment ions and a substitution-specific false discovery rate, then compare observed retention times and fragmentation patterns with independent predictions and searches.
- Quantify proteoforms: pair each substituted amino acid peptide (SAAP) with its canonical base peptide (BP) and calculate the ratio of amino acid substitution.[1]
Application¶
The published analysis applied DECODE to 1,094 human samples from six cancer types and 26 healthy tissues, plus additional primary-cell and mouse tissue datasets. It identified 8,746 unique substituted peptide/base-peptide pairs and supported the finding that alternate RNA decoding can produce stable, abundant and tissue-associated proteoforms.[1]
Validation Strategy¶
The workflow prioritizes specificity over sensitivity. Its validation includes site-localizing fragment ions, independent database searches, predicted fragmentation and retention-time agreement, a false discovery rate calculated specifically across substituted peptides, multiple-protease agreement and targeted peptide-centric searches. Candidates at problematic cleavage sites or with plausible competing explanations are removed.[1]
Availability¶
- Documentation and supporting data: decode.slavovlab.net
- Source code and reproducibility materials: SlavovLab/decode