Kristina Ushakova1, Viktor Shamansky2, Alina A Mikhailova3, Alina G Mikhailova4, Evgeniy Tretiakov5, Ilia Mazunin6, Konstantin Popadin7, Konstantin Gunbin81ITMO University, IKBFU Center for Mitochondrial Functional Genomics, kristina.ushakova@outlook.com2IKBFU Center for Mitochondrial Functional Genomics, v.a.shamanskiy@gmail.com3IKBFU Center for Mitochondrial Functional Genomics, mihailovaalina777@yandex.ru4IKBFU Center for Mitochondrial Functional Genomics, polarsong4@gmail.com5Medizinische UniversitГ¤t Wien, Evgenii.O.Tretiakov@gmail.com6Skoltech Center of Life Sciences, ilya.mazunin@yandex.ru7Ecole polytechnique Federale de Lausanne, IKBFU Center for Mitochondrial Functional Genomics, konstantinpopadin@gmail.com8Institute of Cytology and Genetics SB RAS, IKBFU Center for Mitochondrial Functional Genomics, genkvg@gmail.com The mitochondrial DNA (mtDNA) is a highly streamlined genome with essential role for cellular metabolism. To create a resource for investigation of various aspects of human mtDNA evolution we reconstructed global human mtDNA tree using 40000 complete human mtDNAs. To make our reconstructions robust we used several alternative models of nucleotide substitutions (TN92, GTR); several alternative approaches to reconstruct ancestral states for each nucleotide in each node (parsimony, empirical Bayesian approach and ML marginal reconstruction) and tested statistical significance of each internal node (SH-like approximate likelihood ratio test, approximate Bayes test and ultrafast bootstrap approximation). As a result we obtained a detailed phylogenetic tree of human mitochondrial genomes with more than one million reconstructed single-nucleotide substitutions. We confirmed high quality of our reconstructions, demonstrating: (i) an absence of stop codons on internal nodes, (ii) strong expected excess of C>T and A>G transitions (heavy chain notation) among synonymous substitutions etc. As a first task demonstrating utility of our tree, we reconstructed detailed mtDNA germline mutational spectrum (global, gene-specific and with nucleotide context) based on a collection of more than 300’000 synonymous substitutions. Next we compared it with somatic mtDNA mutational spectrum, derived from cancer data. Globally, two mutational spectra were similar to each other, while local shifts in the frequency of several types of substitutions may be explained by different mutagenic environment between female gamete line and various somatic tissues. We hope that this resource will help to address numerous questions of human mtDNA evolution in the future.В