Rutgers University Develops PRISM Tool to Analyze Cancer Genomic Data
Addressing Contamination in Cancer Genomic Studies
Researchers at Rutgers Cancer Institute have created a computational method called PRISM (Precise Identification of Species of the Microbiome) to resolve a critical challenge in cancer genomics. When analyzing tumor DNA sequences, scientists frequently detect genetic material from bacteria, viruses, and fungi. However, distinguishing whether these microorganisms genuinely reside in tumor tissues or represent environmental contamination has remained contentious.
The Contamination Conundrum in Genomic Analysis
Microbial DNA can enter samples through laboratory reagents, airborne particles, or skin contact during processing. Previous studies analyzing identical genomic datasets produced conflicting conclusions about microbial presence in tumors. “There could be DNA particles floating in the air. How do you know what you’re finding came from the tissue you were interested in?” explains Subhajyoti De, PhD, senior author of the study published in Cancer Cell.
How PRISM Enhances Microbial Detection Accuracy
PRISM employs a multi-stage analytical approach:
- Rapid initial screening of sequencing data
- Stringent removal of residual human DNA sequences
- Full-length alignment against microbial reference databases
- Machine learning classification to differentiate true microbial signals from contaminants
Validation Against Known Microbial Profiles
When tested on 833 samples with established microbial compositions, PRISM demonstrated over 90% sensitivity and specificity, outperforming five existing analysis methods. This validation confirms its reliability for research applications.
Key Findings From Cancer Genome Analysis
Applying PRISM to 4,400 tumor samples across 25 cancer types revealed:
- Strong microbial signals in cancers originating from microbe-rich environments (head/neck, gastrointestinal, cervical)
- Minimal microbial presence in tumors from internal organs without environmental exposure
- Contamination patterns explaining previous discrepancies in microbiome studies
Case Study: Microbial Links in Pancreatic Cancer
PRISM identified a subset of pancreatic tumors containing E. coli bacteria capable of producing colibactin—a DNA-damaging toxin. These tumors showed molecular changes associated with fibrotic tissue development, which can impede drug penetration. Patients with smoking histories demonstrated higher microbial abundance, suggesting potential environmental influences on tumor microbiomes.
Research Implications for Cancer Treatment
PRISM enables researchers to:
- Reanalyze existing genomic datasets without costly new sequencing
- Identify patients who may benefit from microbiome-targeted therapies
- Investigate microbial influences on treatment response variability
“We can generate hypotheses about which patients might respond to certain therapies or which microbial metabolites might be druggable targets,” notes lead author Bassel Ghaddar.
Accessibility and Future Applications
The tool is freely available to academic researchers through GitHub, with commercial applications protected by Rutgers’ intellectual property filings. Beyond oncology, PRISM shows promise for studying gastrointestinal diseases and other conditions where microbiomes influence pathogenesis.
Next Steps for Researchers and Clinicians
This breakthrough creates opportunities to:
- Validate microbial-tumor interaction hypotheses through targeted studies
- Develop clinical trials incorporating microbiome analysis
- Refine cancer subtyping based on microbial signatures
Rutgers Cancer Institute continues to investigate how tumor microbiomes influence cancer progression and treatment outcomes. Researchers worldwide can now apply PRISM to advance precision oncology initiatives.
For researchers interested in implementing this methodology, Rutgers provides detailed documentation through their computational biology resources. Institutions may contact the Office of Research Commercialization for licensing inquiries.
