<i>Raw Data and Curated Datasets for NeoBridge: A Generalizable Neoantigen Prioritization Framework</i>

<p dir="ltr">This repository contains the raw and curated datasets required to reproduce the analyses and train the NeoBridge predictive framework.</p><p dir="ltr">The <code>.zip</code> file contains the <code>data/raw/</code> directory structure expected by the NeoBridge source code, including:</p><ul><li><b>Core Neoantigen Database:</b></li><li><ul><li><code>NeoPepBase.csv</code>: The harmonized, experimentally validated pan-cancer neoantigen dataset used for model training and benchmarking.</li></ul></li><li><b>ICGC ARGO PDAC Cohort (Clinical & Phylogeny):</b></li><li><ul><li><code>clinical_data_APGI/</code> & <code>clinical_data_PACA/</code>: Detailed clinical metadata (follow-up, primary diagnosis, therapy) for the 180 PDAC patients analyzed.</li><li><code>Master_Neoantigenos_Filogenia_PhyClone.csv</code>: The integrated neoantigen candidate list mapped to the PhyClone-derived clonal architectures (clonal/subclonal stratification).</li><li><code>dataset_pancreas_completo.csv</code> & <code>todos_los_peptidos_unidos.tsv</code>: The complete pool of extracted neo-peptides derived from the PDAC cohort.</li></ul></li><li><b>Multi-Omic Expression & Proteomics Reference Data:</b></li><li><ul><li><code>GTEx/</code>: Baseline healthy tissue expression profiles (<code>GTEx_all_mean.csv</code>, <code>GTEx_all_median.csv</code>) used to filter non-expressed genomic regions.</li><li><code>PAAD.rnaseqv2__illuminahiseq_...txt</code>: TCGA-PAAD normalized RNA-seq expression data.</li><li><code>mRNA_RSEM_UQ_log2_Tumor.cct</code>, <code>proteomics_gene_level_MD_abundance_tumor.cct</code> & <code>abundance_peptide_MD.tsv</code>: CPTAC matched transcriptomic and proteomic data used for the multi-omic intersection analysis and validation.</li></ul></li><li><b>Genomic & Mutation Alignments:</b></li><li><ul><li><code>TCGA_PAAD_Mutation.maf</code> & <code>Mutation_results.maf</code>: Mutation Annotation Format files for cross-referencing variant frequencies.</li><li><code>sequence_proteins.fasta</code>: Reference protein sequences used during feature engineering.</li></ul></li><li><b>MHC Binding Motifs (Feature Engineering):</b></li><li><ul><li><code>Mix_motif/</code>: Directory containing over 700 Position Weight Matrices (PWMs) for various HLA class I alleles across peptide lengths of 8 to 12 amino acids.

These are essential for calculating the biophysical determinants and differential agretopicity features.</li></ul></li></ul><p dir="ltr"><b>Instructions:</b> Extract the contents of this <code>.zip</code> file directly in the data directory of your cloned NeoBridge code repository. This will correctly populate the <code>data/raw/</code> path required by the Python scripts to run the pipeline seamlessly.</p>