Analysis of high-throughput biological data using their rank values

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• High-throughput technologies generate vast gene expression and cytogenetics data.
• Existing analysis methods often require specialized expertise and significant computational resources.

Purpose of the Study:

• To introduce a versatile, computationally efficient method for analyzing gene expression and cytogenetics data.
• To detect recurrent copy number aberrations and identify statistically significant gene subsets.

Main Methods:

• Utilizes data-ordering rank values, linear algebra, and the Perron-Frobenius theorem.
• Extends existing methods for differential gene expression analysis to copy number aberration detection.
• Develops a rank-based one-sample Student's t-test and introduces stability scores for quality assessment.

Main Results:

• The proposed method is applicable to both gene expression profiling and cytogenetics datasets.
• Demonstrates a fast, deterministic analysis with low computational requirements.
• Enables statistically significant subset selection through associated gene probabilities and provides quality control via stability scores.

Conclusions:

• Presents a unique, unified method for analyzing diverse high-throughput biological data.
• Offers a computationally inexpensive and statistically robust alternative for detecting genomic alterations.
• The method is accessible via the R package 'fcros' on CRAN.