Reliable detection of Host-Microbe Signatures in cancer using PRISM

Researchers led by Ghaddar et al. have developed PRISM, a new computational framework designed to accurately identify microbial organisms and remove contamination artifacts from human cancer genomic data. By applying PRISM to two of the largest cancer genomics repositories — The Cancer Genome Atlas (TCGA) and the Clinical Proteomic Tumor Analysis Consortium (CPTAC) — the team uncovered robust host-microbe signatures in several tumor types. These microbial signatures were further linked to specific molecular characteristics and clinical features of the tumors, suggesting functional relevance of the tumor microbiome. The work, published in Cancer Cell, addresses a critical methodological gap in the field, where contamination has historically undermined confidence in tumor-associated microbiome studies. PRISM could serve as a standardized tool to accelerate reliable research into how microorganisms influence cancer biology and patient outcomes.