MitoIMP: A Computational Framework for Imputation of Missing Data in Low-Coverage Human Mitochondrial Genome

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Incomplete human mitochondrial genome sequences hinder accurate comparisons across different data resources.
• Existing partial sequences present challenges for phylogenetic and population genetic analyses.

Purpose of the Study:

• To develop a computational framework for accurately imputing missing nucleotides in human mitochondrial genome sequences.
• To provide a practical solution for compensating low genome coverage in mitochondrial DNA data.
• To assess the performance of the imputation framework across diverse human mitochondrial haplogroup lineages.

Main Methods:

• Development of a computational framework for nucleotide imputation in mitochondrial genomes.
• Application of the framework to worldwide human mitochondrial haplogroup lineages.
• Performance assessment using precision metrics and multidimensional scaling analysis.

Main Results:

• The imputation framework achieves a precision of 0.99 or higher for most human mitochondrial DNA lineages.
• The approach corrects blurred relationships in multidimensional scaling analysis caused by low-coverage sequences.
• An open-source program, MitoIMP, was developed to implement the imputation procedure.

Conclusions:

• The developed framework offers a reliable method for addressing missing data in human mitochondrial genome sequences.
• MitoIMP provides a practical solution to improve the utility of partial and fragmented mitochondrial DNA datasets.
• This tool will facilitate more robust comparisons and analyses of human mitochondrial genome diversity.