EDISON-WMW: Exact Dynamic Programing Solution of the Wilcoxon-Mann-Whitney Test

Area of Science:

• Biostatistics
• Bioinformatics
• Computational Biology

Background:

• Hypothesis testing is crucial for determining statistical significance in research.
• The Wilcoxon-Mann-Whitney (WMW) test is widely used in medicine and life sciences for comparing two sample groups.
• Exact WMW test computation is challenging for large datasets with ties, often leading to approximations.

Purpose of the Study:

• To develop an efficient method for calculating the exact P value of the two-tailed unpaired WMW test, accommodating ties.
• To provide a computationally feasible solution for large sample cohorts in statistical analysis.

Main Methods:

• Developed EDISON-WMW, a novel approach utilizing dynamic programming to efficiently solve the combinatorial problem of the WMW test.
• Implemented optimization strategies and a parallel solution for enhanced performance.
• Validated the method on random data, benchmarked against 13 existing approaches, and applied it to molecular biomarker evaluation.

Main Results:

• EDISON-WMW accurately calculates exact P values for large cohorts (>1000 samples) with ties within minutes.
• Approximated P values were generally higher than the exact values obtained from EDISON-WMW.
• The algorithm demonstrated superior performance and is applicable to high-throughput omics data.

Conclusions:

• EDISON-WMW offers an efficient and accurate solution for exact WMW testing, overcoming limitations of traditional approximation methods.
• The tool is valuable for biomedical research, particularly in molecular diagnosis and omics data analysis.
• A web-based, multi-threaded version of EDISON-WMW is freely available for broader accessibility.