Eastern Caucasus is home to more than 30 peoples speaking Caucasian, Iranian and Turkic languages. Fusion of multiple migration flows together with the complex population structure of the Eastern Caucasus make it more difficult to analyze its gene pool: this is the most poorly studied one among all regions of the Caucasus. The study is aimed to identify the main patterns of the autosomal gene pool variation in this region. A total of 356 genomes of 29 ethnic groups were studied using the large panels of SNP markers: 243 genomes of 22 peoples of the Eastern Caucasus and 113 genomes of 7 peoples living in adjacent regions. The bioinformatics analysis involved the use of the ADMIXTURE ancestral component method and the gene pool variability principal component analysis (РСА). The hypothesis of three genetic strata, the interaction of which forms the structure of gene pool of the Eastern Caucasus, was put forward. The “Dagestan” stratum carries information about the gene pool of the ancient autochthonous population of the Eastern Caucasus. The “Iranian” stratum represents the legacy of ancient and middle-aged migrations surges of the Iranian-speaking population: it constitutes three quarters of the gene pool of modern Azerbaijan and about one third of the Dagestan peoples' gene pool. The “Steppe” stratum represents a negligible influence of the Eurasian steppe. Interaction of three genetic strata is only indirectly related to the peoples' linguistic affiliation, however, the association with linguistics is more obvious in the Caucasian-speaking peoples. Four genetically distinct groups of indigenous population of the Eastern Caucasus have been identified, the combination of which should be included in the characteristics of its autosomal gene pool: 1) Dargins, Laks; 2) Avars, Lezghins, Tabasarans, Aghuls, Rutul people, Tsakhur people; 3) Kumyks, Tat people and Azerbaijanis living in Dagestan; 4) Azerbaijanis and Talysh living in Azerbaijan. The directions of further research have been defined.