Gene Length Bias in Human Papillomavirus Integration Sites: A Statistical Analysis Using KNIME

Abstract Integration of high-risk Human Papillomavirus (HPV) into the human genome represents a critical event in carcinogenesis. While recurrent integration hotspots are well documented, the structural determinants governing host gene susceptibility remain incompletely defined. Here, we evaluate whether human gene length influences HPV integration probability using VISDB integration datasets for HPV16 and HPV18 merged with Ensembl BioMart annotations. Statistical analysis performed in KNIME reveals that HPV-integrated genes exhibit significantly greater length than non-integrated genes (Two-sided Wilcoxon-Mann-Whitney, p = 0.0001). Distributional analysis demonstrates a consistent rightward shift in log-transformed gene length among integrated genes. Genomic validation using the UCSC Genome Browser confirms integration within established oncogenic regions, including 8q24 near MYC. These findings identify gene length as a measurable structural bias contributing to HPV integration susceptibility, while underscoring the role of genomic context in hotspot formation.