TAL-SRX: an intelligent typing evaluation method for KASP primers based on multi-model fusion

Area of Science:

• Agricultural Science
• Bioinformatics
• Genetics

Background:

• Accurate KASP primer typing is vital for efficient marker-assisted breeding.
• Current manual and algorithmic methods for KASP typing lack efficiency and scalability.

Purpose of the Study:

• To develop an intelligent and accurate method for evaluating KASP primer typing effects.
• To improve the efficiency and stability of marker screening in breeding programs.

Main Methods:

• Proposed TAL-SRX method integrating deep learning (ANN, LSTM, Transformer) and traditional machine learning in a Stacking framework.
• Utilized five-fold cross-validation for model stability and a soft voting strategy for fusing algorithms.
• Tested on 3399 KASP typing results from cotton variety resources.

Main Results:

• Achieved 92.83% accuracy and an AUC value of 0.9905 on cotton variety data.
• Demonstrated high accuracy, consistency, and stability, outperforming single models.
• TAL-SRX showed superior performance compared to other integrated combinations.

Conclusions:

• The TAL-SRX model offers robust evaluation of KASP marker typing effects.
• This method provides essential technical support for molecular marker-assisted breeding.
• The integrated approach enhances the speed and accuracy of selecting excellent molecular markers.