The molecular mechanism unraveling why a particular type of pediatric brain tumor (pBT) behaves so differently from child to child or genetic/epigenetic changes in the mitochondrial genome vary from tumor to tumor is not clearly understood. Despite the identification of mitochondrial DNA (mtDNA) mutations in different types of pBT, the contribution of mitochondrial dysfunction-related genes or proteins that are selectively up- or down-regulated in pBT of different types has not been comprehensively examined. In the present study, we combined a 2D DIGE approach with protein identification using MALDI-TOF MS and LC-MS/MS, coupled with mtDNA genomics to screen brain samples for discovering changes in protein expression, and mtDNA sequence variation and mtDNA copy number in the disease states. Two-dimensional gel electrophoresis-based differential proteomic analysis of the brain tumors showed that 116 proteins were found to be up- or down-regulated in brain tumors. Some of the proteins up-regulated in tumors compared to controls were dihydropyrimidinase-like 2; glial fibrillary acidic protein isoform 2; phosphoserine aminotransferase isoform 1; Sirt2 histone deacetylase; and C10orf2 protein, mitochondrial DNA helicase. Proteins down-regulated in brain tumors compared to controls were heat shock protein 90 kDa beta, BiP; guanine nucleotide binding protein (G protein), beta polypeptide 2-like 1, isoform CRA_d; histone H2B.1; neurofilament, light polypeptide 68 kDa; Annexin I; and RAN. These differentially expressed proteins may provide useful information for developing molecular markers of diagnostic or prognostic value. To investigate further the role of mitochondrial dysfunction, we examined the effects of mtDNA copy number, oxidative damage, and mtDNA variants as independent or combined risk factors for the development of pBTs. Bayesian network and mechanistic hierarchical structure Markov Chain Monte Carlo (MCMC) modeling were used to analyze the relationship between these variables. The combined effects of G3196, 9952A, 10006G, 100398G, oxidative mtDNA damages, and mtDNA copy number increased the probability of developing brain tumors in female children by 51 times more when compared to normal incidence of pediatric brain tumors. Comparison of mechanistic structure models also supported the finding that female children who have the wild type allele G3196, variant allele 9952A, variant allele 10006G, variant allele10398A, and high mtDNA copy number had increased probability of developing pediatric brain tumors. Estimation of nuclear genes controlling mitochondrial biogenesis and development of brain, cortical dysplasia, and the effect of the environment using MCMC method showed that these latent variables had a very significant contribution in the development of pediatric brain tumors. Together, these results suggest that mitochondrial genome and tumor proteome are important contributors to brain tumor risk in children, and findings from this study may guide the prospects for targeting mitochondria for therapeutic treatment of childhood brain tumor.