{
  "abstract": "Background: Non-canonical open reading frames (ncORFs) encode many microproteins that are rarely detected by mass spectrometry (MS), hindering their annotation as proteins. The properties that make them detectable in public MS repositories are poorly understood.\n\nResults: We analysed 7097 Ribo-seq-supported human ncORFs, 1725 of which have peptide evidence in PeptideAtlas builds, using 18 biological features and embeddings from the ESM2 protein language model (PLM) in logistic regression and ensemble classifiers (ROC-AUC 0.768 ± 0.006 over five runs). Translation uniformity across the ORF (Cumulative In-Frame Ribosome Footprint Uniformity, CIF_HBP), mammalian conservation of ORF structure (ORBLv_mammal) and predicted HLA class I-binding peptide density were top-ranked by both SHAP and permutation importance. Adding ESM2 embeddings to the three top biological features did not change ROC-AUC (0.689 in both cases; ΔAUC = 0.000 ± 0.003), possibly because ESM2 was pre-trained on much longer canonical proteins. About 90% of positives are supported by HLA immunopeptidomics alone and 99% by HLA data alone or together with non-HLA data, so the model largely reflects immunopeptidome visibility; because all ncORFs have Ribo-seq support, it predicts MS detectability conditional on translation; when whole biotypes were held out, AUC fell from 0.706 to 0.659–0.669. Of the 200 top-ranked ncORFs, 27 were identified in a Ribo-seq-guided proteogenomic study (5.26-fold over chance, p = 7.2 × 10⁻ ¹³ ) and 28 in a second, more recent dataset (6.9-fold over chance, p = 1.9 × 10⁻ ¹⁶ ; significant after Bonferroni or Holm correction for two tests).\n\nConclusions: Sustained translation across the ORF and evolutionary conservation of ORF structure are statistically associated with the cumulative MS detection of ncORF microproteins, but label-source bias, the absence of protein abundance as a feature and biotype-specific patterns limit generality. These associations can help prioritise candidates for targeted proteomic discovery.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": null,
  "author_total": 1,
  "authors": [
    {
      "name": "Ivo Fierro-Monti"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/21qv92mk6dbff.bib",
  "cited_by": 0,
  "doi": "10.1186/s12915-026-02777-3",
  "fulltext_url": null,
  "id": "21qv92mk6dbff",
  "issns": [
    "1741-7007"
  ],
  "landing_url": "https://link.springer.com/article/10.1186/s12915-026-02777-3",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "journal",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [],
  "openalex": null,
  "pdf_url": null,
  "pmcid": null,
  "pmid": null,
  "published": "2026-11-09",
  "publisher": "Springer Science and Business Media LLC",
  "ris": "https://research.draconitic.com/paper/21qv92mk6dbff.ris",
  "sources": [
    {
      "first_seen": "2026-11-09T12:22:47Z",
      "native_id": "10.1186/s12915-026-02777-3",
      "source": "crossref",
      "url": "https://link.springer.com/article/10.1186/s12915-026-02777-3"
    }
  ],
  "subjects": [
    {
      "field": "biochemistry",
      "origin": "journal-openalex:1312|Genomics and Phylogenetic Studies",
      "subfield": "molecular-biology"
    },
    {
      "field": "biology",
      "origin": "issn:1741-7007",
      "subfield": "general"
    }
  ],
  "title": "Translation uniformity and evolutionary conservation are associated with microprotein mass spectrometry detectability",
  "title_text": "Translation uniformity and evolutionary conservation are associated with microprotein mass spectrometry detectability",
  "type": "peer-reviewed",
  "updated_at": "2026-11-09T12:22:47Z",
  "url": "https://research.draconitic.com/paper/21qv92mk6dbff",
  "venue": "BMC Biology",
  "versions": []
}