Isovar uORF evidence catalogue, dataset version 2026-10-08.2.

The Isovar API and builder are Apache-2.0. Source data retain their own licenses.

GENCODE Phase I models and legacy MS annotations, plus HLA mapping/count data:
Copyright (c) 2025 Leron Kok. MIT license, included as AUTHOR-LICENSE.txt.
VanHeeschLab/deutsch_kok_et_al_2024, commit
b65ae0c09a53e9efe083d806d617581e6139bdde.
https://github.com/VanHeeschLab/deutsch_kok_et_al_2024
Associated studies: Mudge et al. 2022, DOI 10.1038/s41587-022-01369-0;
Deutsch et al. 2026, DOI 10.1038/s41586-026-10459-x.

Community spectrum review input:
Aaron Wacholder and the community assessment authors, archived dataset version 1.
https://doi.org/10.6084/m9.figshare.30131869.v1
License: Creative Commons Attribution 4.0 International (CC BY 4.0).
https://creativecommons.org/licenses/by/4.0/
Associated study: Wacholder et al. 2026, DOI 10.1038/s41467-025-68002-x.

Changes: selected uORF/uoORF models; normalized coordinates and source metadata;
joined sequence-compatible peptide observations; retained individual reviews,
source attribution and unknown mapping counts; calculated peptide intervals and
conservative review labels. Full original article text and supplementary
curation tables are not redistributed. Source URLs, SHA256 pins, evidence scope
and exclusions are in catalog metadata and manifest.json.

Initiation-context sequence: Ensembl release 101, human GRCh38 GTF/cDNA.
Ensembl-generated data are available without restriction:
https://www.ensembl.org/info/about/legal/disclaimer.html
Changes: exact ORF path/protein validation; extracted transcript-oriented start
flanks and genomic placements; added partial-context/positional Kozak priors.
Sequence priors are not measurements of translation or protein abundance.
