Data mining examination of mitochondrial DNA variants in oxidative phosphorilation system diseases

The heteroplasmic ratio of various pathogenic alleles in the circular genome (mtDNA) of the mitochondria, determines the variant-burden and is responsible for the clinical heterogeneity of mitochondrial diseases. In the first step of the study, it was aimed to predict the efficacy of the sample used in diagnostic tests by making tissue variant-load comparisons. The mtDNA obtained simultaneously from different tissues of six volunteers was investigated using new generation sequencing technology. The number of variants detected in the muscle was higher than in blood. In terms of variant type, the number of SNVs was greater in muscle, while in indels the number was approximately. The number of indels was high in common variants. The frequency of common variants in the muscle was found to be 1.2 times higher, 3.7 times higher in common SNV and 1.1 times higher in common indel. It was determined that 28.6% of the variants in the blood could not be detected in the muscle, and 36.5% of the variants in the muscle could not be detected in the blood. The findings supported the view that mtDNAs with variants are eliminated in cells with high mitosis, such as blood, and that this advantage does not occur in non-dividing cells such as muscle. In addition, the results coincided with the view that indels in mitochondria are formed during the repair of newly formed SNVs, so indel-type variants can be expected more in blood. The number of volunteers was insufficient for three tissue comparisons. In the second step, it was aimed to establish a relationship between clinical diversity and muscle pathogenic variant-load and to create a multi-linear regression model that can predict the ratio of muscle heteroplasmy. A data set was created with the characteristics of patients with the m.3243A>G pathogenic variant by examining 445 publications. Because of the missing heteroplasmy ratio information in this data set, Model-1 was created using half of the data set (421 rows, 26 columns). Model-2 was created with the data set (744 rows, 26 columns) half completed with Model-1. The adjusted R-squared values of these models, which are the association success scale, are 0.47 and 0.72, respectively. The increase in this value was associated with the increase in the number of data. As a result, it was thought that the development of the regression model would be possible by increasing the number of rows and eliminating the missing values with optimized clinical forms. Key Words: Mitochondrial DNA, heteroplasmy, mitochondrial diseases, m.3243A>G, multi-linear regression