A Comprehensive Analysis of Novel Variations Associated with Bile Duct Cancer: Insights into Expression, Methylation, and 3D Protein Structure

Abstract

Cholangiocarcinoma is a rare but highly lethal cancer of the biliary epithelium, marked by heterogeneous molecular subtypes, unclear etiology, and poor five-year survival, highlighting the need for new diagnostic and therapeutic strategies; therefore, this study integrates genomic, transcriptomic, single-cell, methylomic, and molecular-dynamics data to pinpoint pathogenic variants. We performed an integrative multi-omics analysis of publicly available datasets. Somatic variants from 23 tumor samples in The Cancer Genome Atlas were annotated with 11 pathogenicity tools (AUC >= 0.86 across EVE, REVEL, SIFT, AlphaMissense, DEOGEN2 were the most stringent). Differential gene expression was assessed in matched bulk RNA-seq (tumor vs. non-tumor) using DESeq2 with Benjamini-Hochberg FDR correction. A single-cell RNA-seq dataset comprising 23,782 cells from an intrahepatic cholangiocarcinoma was clustered with marker genes identified by Wilcoxon rank-sum tests. Illumina 450 K methylation arrays (52 tumors, 12 normal livers) were analyzed with limma and DMRcate to detect differentially methylated probes and regions. AlphaFold3 models of wild-type and MAP2K1R49C were subjected to 50 ns all-atom molecular-dynamics simulations in GROMACS; conformational shifts were quantified by RMSD/RMSF and stability tested with FoldX5. Twenty-four tumor-specific missense variants were detected. The four highest-confidence pathogenic substitutions (EVE, REVEL, SIFT, AlphaMissense, DEOGEN2) occurred in TUBB3, FLNC, ABCA1, and MAP2K1. Bulk RNA-seq confirmed significant dysregulation of these genes and enrichment of extracellular-matrix organization, cytoskeletal remodeling, MAPK signaling, and cholesterol-efflux pathways. Single-cell analysis resolved 23 transcriptionally distinct clusters; proliferative malignant cholangiocytes selectively over-expressed ABCA1 and MAP2K1, indicating tumor-cell specificity. Methylome profiling identified 148,928 DMPs and 7040 DMRs; promoter hypomethylation of TUBB3 and ABCA1 correlated with their transcriptional activation. Substituting Arg-49 with Cys in MAP2K1 dismantles the Arg-centred hydrogen-bond/salt-bridge cluster, reduces hydrophobic packing, and, corroborated by 50 ns MD (Welch's t = -58.06, p = 3.17 x 10-230) and FoldX5 (Delta Delta G = +2.3 kcal mol-1), significantly destabilises the protein, manifesting as higher backbone RMSD and increased local flexibility relative to wild type. This multi-omics, public data-driven synthesis delineates a coherent network of genomic, epigenomic, transcriptomic, and structural vulnerabilities, offering a rational framework for therapeutic targeting of cholangiocarcinoma. This study reveals novel bile duct-associated variations that expand our understanding of cholangiocarcinoma pathogenesis and provide potential targets for precision medicine approaches.