Our laboratory has utilized total synthesis, semi-synthesis, and fermentation of bacterial species to produce natural product analogs to study the relationship between conformational preferences and biological activity. Our goal is to show the importance of the conformational characteristics of natural products in the development of potential therapeutics, features that are commonly disregarded in structure-activity relationship studies.A new linear analog for zampanolide, a macrolide extracted from the marine sponge Faciospongia rimosa; and new diene analogs for GEX1A, a compound isolated from Streptomyces chromofuscus, were synthesized. Their conformational preferences were predicted with computational methods and corroborated by nuclear magnetic resonance experiments. The biological activity of the new compounds was tested on cancer cell lines, which allowed us to assess if the initial hypotheses were valid. In the case of zampanolide, it was demonstrated that the linear analog exhibited potent biological activity due to the conservation of an allylic strain, which restricted the conformational flexibility of the molecule. For GEX1A, it was shown that an increase in the population of the main conformational family did not lead to enhanced bioactivity.Additionally, an introduction to machine learning and some examples of its use in organic chemistry is given. This is in light of the work done in iridium-catalyzed hydrogen isotope exchange reactions, that was performed during a six-month internship at the Catalysis group in Merck. High throughput experimentation was used to create a large dataset to train a machine learning model, that predicted the deuterium incorporation on drug-like molecules. This model is now being used and optimized on pipeline projects at Merck.