Zero-inflated generalized Poisson regression mixture model for mapping quantitative trait loci underlying count trait...

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Count phenotypes are prevalent in biological studies.
• Existing quantitative trait loci (QTL) mapping methods often assume Poisson distribution, struggling with excess zeros.
• Misinterpreting zero counts can lead to inaccurate genetic inferences.

Purpose of the Study:

• To develop a novel interval mapping approach for QTL detection in count traits with a high proportion of zeros.
• To model the genetic effects on zero-inflated count phenotypes using a robust statistical framework.

Main Methods:

• Proposed a zero-inflated generalized Poisson regression mixture model to simultaneously address zero inflation and overdispersion.
• Implemented the approach using the Expectation-Maximization (EM) algorithm with embedded Newton-Raphson optimization.
• Conducted genome-wide scans for QTL detection and effect estimation.

Main Results:

• Extensive simulation studies validated the performance of the proposed method.
• The approach successfully identified significant QTLs controlling cholesterol gallstone formation in a mouse F(2) intercross dataset.
• Demonstrated the utility of the developed method in a real-world biological context.

Conclusions:

• The developed interval mapping approach effectively handles zero-inflated count data for QTL analysis.
• This method provides a more accurate inference of genetic architecture for complex count traits.
• The findings have implications for understanding the genetic basis of diseases and other biological phenomena characterized by count data.