Two sample comparisons including zero-inflated continuous data: A parametric approach with applications to microarray...

Area of Science:

• Molecular Biology
• Biostatistics

Background:

• Micro-array experiments frequently yield mixed distributions with a zero clump.
• Challenges include skewed continuous data and small sample sizes, limiting traditional statistical methods.

Purpose of the Study:

• To develop novel statistical tests for analyzing micro-array data with zero-inflated distributions.
• To address limitations of existing methods, particularly under small sample conditions.

Main Methods:

• Proposed an exact test for location-scale family distributions.
• Developed Generalized Pivot Quantity (GPQ) based parametric tests for non-location-scale distributions.
• Conducted simulation studies to assess performance.

Main Results:

• The new exact and GPQ tests demonstrated superior size and power compared to popular two-part tests.
• Performance advantages were most pronounced in small sample size scenarios.
• Outperformed Wilcoxon rank sum, t-test, Kolmogorov-Smirnov, Ansari-Bradley, and Siegel-Tukey tests.

Conclusions:

• The proposed exact and GPQ tests offer a robust solution for analyzing zero-inflated data in micro-array studies.
• These methods enhance statistical inference quality, especially when dealing with limited sample sizes.
• Provides more reliable comparisons between biological groups.