Confident Predictability: Identifying reliable gene expression patterns for individualized tumor classification using a local minimax kernel algorithm
2011

Predicting Cancer with Gene Expression Patterns

Sample size: 280 publication Evidence: moderate

Author Information

Author(s): Jones Lee K, Zou Fei, Kheifets Alexander, Rybnikov Konstantin, Berry Damon, Tan Aik Choon

Primary Institution: University of Massachusetts, Lowell

Hypothesis

Can a local minimax kernel algorithm improve the predictability of cancer diagnoses based on gene expression data?

Conclusion

The study demonstrates that the local minimax kernel algorithm can significantly reduce error rates in cancer predictions for a subset of patients.

Supporting Evidence

  • More than 50% of leukemia patients were identified as confidently predictable with reduced error rates.
  • 55% of patients in the global cancer map dataset had their error rates significantly lowered.
  • One-third of prostate cancer patients had their error rates cut by 70%.

Takeaway

This study shows a new way to predict if someone has cancer by looking at their genes, helping doctors make better decisions.

Methodology

The study used a local minimax kernel learning algorithm combined with feature selection via the k-TSP algorithm on three gene expression datasets.

Limitations

The method's effectiveness may vary with different datasets and the need for further validation in clinical settings.

Participant Demographics

The study included patients with various types of cancer, specifically leukemia, prostate cancer, and other tumors.

Statistical Information

Confidence Interval

90% CI

Digital Object Identifier (DOI)

10.1186/1755-8794-4-10

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication