Explainable Machine Learning Identifies Dosage Compensation Factors in Aneuploid Human Cancer Cells

Aneuploidy, a hallmark of cancer, leads to widespread changes in chromosome copy number, altering the abundance of hundreds or thousands of proteins. How-ever, evidence suggests that levels of proteins encoded on affected chromosomes are often buffered toward their abundances observed in diploid cells. Despite its preval-ence, the molecular mechanisms driving this protein dosage compensation remain largely unknown. It is unclear whether all proteins are buffered to a similar degree, what factors determine buffering, and whether dosage compensation varies across different cell lines or tumor types. Moreover, its potential adaptive advantage and therapeutic relevance remain unexplored. Here, we established a novel approach to quantify protein dosage buffering in a gene copy number-dependent manner, show-ing that dosage compensation is widespread but variable in cancer cell lines and in vivo tumor samples. By developing multifactorial machine learning models, we identify mean gene dependency, protein complex participation, haploinsufficiency, and mRNA decay as key predictors of buffering. We also show that dosage com-pensation can affect oncogenic potential and that higher buffering correlates with reduced proteotoxic stress and increased drug resistance. These findings highlight protein dosage compensation as a crucial regulatory mechanism and a potential therapeutic target in aneuploid cancers.