I developed a strategy for creating 3D models of non-canonical DNA and RNA structures of any topology and complexity. This strategy allowed me to analyze and systematize all possible geometric variations of non-canonical DNA and RNA structures at contacts between duplexes containing guanine or cytosine repeats. Using this strategy, I discovered new combinations of non-canonical DNA shapes significantly more complex than the Holliday structure, including a new type of i-motif (with a twist). I predicted the existence of new combinations of shapes from non-complementary single-stranded DNA: dimer and trimer, and quadruplex layers. The shapes I discovered provide new insights into the mechanisms of homologous recombination and repair. The developed strategy also allowed me to create a model of polytymine hairpin DNA. I implemented flexible docking of ligands with high conformational capacity: synthetic and bio-oligomers, both with and without 3D structure due to non-covalent interactions (quadruplexes, i-motifs, various hairpins) (peptides, polysaccharides, phospholipids), and antibiotics (olivomycin and oligomycin derivatives). I hypothesized that during polymer coil formation, nuclei (loops, hairpins) are formed first, then interacting with each other, leading to coil formation. I verified this hypothesis using MD on several synthetic and biopolymers.