Human mitochondrial genome compression using machine learning techniques.

Area of Science:

• Genomics
• Bioinformatics
• Machine Learning

Background:

• High-throughput sequencing generates vast amounts of genomic data, leading to concerns about storage and transmission costs.
• Effective compression of genome sequences remains a significant challenge in bioinformatics.

Purpose of the Study:

• To propose and evaluate a novel machine learning-based method for compressing human mitochondrial genome data.
• To address the growing need for efficient genome data compression solutions.

Main Methods:

• Development of DeepDNA, a machine learning technique specifically designed for genome sequence compression.
• Application of DeepDNA to human mitochondrial genome datasets.

Main Results:

• DeepDNA demonstrates effective compression of human mitochondrial genome data.
• The method's compression performance is comparable to established reference-based compression techniques.
• Achieves good compression results for both population and single genomes, regardless of redundancy levels.

Conclusions:

• DeepDNA offers a competitive non-reference-based compression method for genomic data.
• The approach is versatile, performing well across different genome types and redundancy levels.
• The developed codes are publicly available for further research and application.