Selecting MYB–NFIB and MYBL1–NFIB junction peptides for experimental testing in adenoid cystic carcinoma

Version 14.3: data and code for “Selecting MYB–NFIB and MYBL1–NFIB junction peptides for experimental testing in adenoid cystic carcinoma.” The resource links read-reconstructed fusion junctions to same-library HLA class I assignments, peptide predictions, specimen-source annotations and normal-protein sequence comparisons across 284 public RNA-sequencing libraries from six BioProjects. The primary shortlist contains 57 peptide–HLA pairs comprising 34 sequences. Thirty-six pairs have support from 19 libraries with documented tumor-source assignments; 21 pairs depend on eight libraries whose individual tumor/normal roles remain unresolved. Selection requires MHCflurry affinity percentile and NetMHCpan no-context eluted-ligand rank both ≤2% in the same observation. An eight-pair fixed testing panel is included. The repository includes the manuscript and supplement, the complete shortlist, accession crosswalks, current and subset-analysis workbooks, figures, reference sequences, saved predictions and reproduction code. Additional analyses assess HLA eligibility, alignment-start support, sequence coverage, coordinate mapping, prediction rules, study dependence and normal-protein neighbours. A full-cohort HLA suffix-relaxation analysis supplies 111 qualifying pairs, including 38 MYBL1–NFIB pairs compared with three under the primary rule, and retains all 57 primary pairs. Normal-sequence comparisons report predicted affinity percentiles and eluted-ligand ranks, not measured binding. Public HLA eligibility flags and anonymized scored observations support replay of filtering and selection. Unlinked donor-relationship matrices and aggregate metrics support replay of conditional donor calculations without distributing individual fingerprints or accession-linked relationship classifications. Bulk sequencing reads and third-party predictor executables/model weights are not included; accessions, software versions and reference identifiers are supplied.