Deciphering individual triticale grain weight patterns: A gaussian mixture model approach.

Area of Science:

• Agricultural Science
• Biometrics
• Plant Physiology

Background:

• Grain weight is a critical factor influencing crop yield.
• The underlying distribution structure of grain weight is often oversimplified or overlooked in crop studies.

Purpose of the Study:

• To analyze the individual grain weight distribution of triticale using a Gaussian Mixture Model (GMM).
• To investigate if grain weight distribution follows a single or multiple normal distributions.
• To explore the implications of grain weight distribution for understanding crop physiology.

Main Methods:

• Individual grain weights of three triticale cultivars were analyzed over time post-heading.
• The Gaussian Mixture Model (GMM) was employed to fit the grain weight distributions.
• Model fit was evaluated using the Corrected Akaike Information Criterion (AICc) and Bayesian Information Criterion (BIC).

Main Results:

• The grain weight distribution of triticale was better represented by a sum of two normal distributions, not a single one.
• This bimodal distribution pattern was observed across different triticale cultivars and seeding rates.
• The findings suggest a link between grain weight distribution and the physiological characteristics of triticale spikelets.

Conclusions:

• Grain weight distribution in triticale is complex and better modeled by multiple normal distributions.
• Understanding the nuanced distribution structure is crucial for accurately assessing crop traits and yield.
• This approach offers a more refined perspective on crop phenotyping and physiological analysis.