Intratumoral heterogeneity and clonal variability are among the central problems of clinical oncology, leading to resistance to therapy, relapse, and metastasis. High-throughput sequencing of the tumor exome makes it possible to investigate the subclonal tumor organization. Target panel, clinical exome, and complete exome sequencing data were compared in tumors with different mutational burden, acute myeloid leukemia (AML) in children and acral melanoma. Targeted sequencing of AML samples detected more than one potential driver mutation in the signaling pathway genes KIT, NRAS, KRAS, CBL, and FLT3 in one patient, reflecting the complex clonal structure of the tumor substrate. Clusters of mutant alleles corresponding to different populations of leukemic cells in a sample were isolated based on exome sequencing data from the same AML patients. A comparison of the mutation profile for a primary AML sample and samples obtained in remission and relapse made it possible to trace the dynamic changes in the clonal composition of the tumor. The subclonal tumor structure was investigated in an acral melanoma case as an example. Mutant alleles present in the sample with close frequencies were clustered using the SciClone and ClonEvol packages. The results were used to predict the intratumoral clonal composition and to assume a clonal evolution model, which described the changes in the clonal composition of the tumor during metastasis, including the appearance of new mutations that might be associated with further disease progression. The approach used in the work is suitable for identifying the mutations that cause the formation of new tumor clones, which may have a proliferative advantage, in particular, in conditions of antitumor therapy.