A searching procedure for transformations and models in a classical Mendelian cross breeding study

Area of Science:

• Genetics
• Biometry
• Quantitative Genetics

Background:

• Classical Mendelian crosses generate data requiring appropriate statistical analysis.
• Selecting the best-fitting genetic model is crucial for accurate interpretation.
• Wright's criteria provide a framework for evaluating measurement scales.

Purpose of the Study:

• To develop a systematic procedure for data transformation in Mendelian crosses.
• To identify the most parsimonious and best-fitting genetic model.
• To evaluate model invariance under data transformations.

Main Methods:

• Application of Wright's four criteria for measurement scales.
• Utilizing Cavalli's least-squares fitting procedure for model comparison.
• Systematic testing of all possible genetical (sub)models.

Main Results:

• A procedure was established to find adequate data transformations.
• The best-fitting genetic model was selected based on parsimony and statistical significance.
• Model invariance was observed when the homogeneity of variance criterion was met.

Conclusions:

• The developed method provides a robust approach for analyzing Mendelian cross data.
• Data transformation is essential for meeting measurement scale requirements.
• Parsimonious and statistically significant models are preferred in genetic analysis.