Hecaton: reliably detecting copy number variation in plant genomes using short read sequencing data

Area of Science:

• Genomics
• Plant Science
• Computational Biology

Background:

• Copy number variation (CNV) plays a role in plant adaptive evolution.
• Existing computational methods for CNV detection often produce false positives in complex plant genomes.

Purpose of the Study:

• To develop a robust computational workflow for accurate and comprehensive CNV detection in plant genomes.
• To address the limitations of current CNV detection tools in plant data.

Main Methods:

• Developed Hecaton, a novel machine-learning-based workflow integrating multiple CNV detection algorithms.
• Tailored the workflow specifically for the complexities of plant genomes.
• Evaluated Hecaton on short-read sequencing data from *Arabidopsis thaliana*, rice, maize, and tomato.

Main Results:

• Hecaton outperforms existing methods in CNV detection accuracy for plant genomes.
• Successfully identifies dispersed duplications, a common CNV type in plants, unlike other tools.
• Demonstrates efficient memory usage and running time on diverse tomato accessions.

Conclusions:

• Hecaton offers a reliable and robust method for detecting CNVs in plants.
• The workflow is expected to benefit both fundamental and applied research linking plant genotype to phenotype.