Cnngeno: A high-precision deep learning based strategy for the calling of structural variation genotype

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Genotype determination is crucial for organism characteristics, accelerated by sequencing technologies.
• Parametric statistical models struggle with genotype calling due to structural variations and coverage fluctuations in sequencing data.

Purpose of the Study:

• To develop a novel method, Cnngeno, for accurate deletion genotype calling from next-generation sequencing data.
• To address limitations of existing methods in handling complex sequencing data characteristics.

Main Methods:

• Cnngeno converts sequencing data into images.
• It utilizes a convolutional neural network (CNN) for genotype classification.
• A convolutional bootstrapping strategy enhances robustness against noisy labels.

Main Results:

• Cnngeno demonstrates superior precision in genotype calling compared to existing methods.
• The method effectively handles challenges posed by structural variations and coverage fluctuations.

Conclusions:

• Cnngeno offers an advanced, open-source solution for deletion genotype calling.
• Its CNN-based approach provides improved accuracy and robustness in genomic data analysis.