Giselle Brown
MIT Department: Chemistry
Faculty Mentor: Masha Elkin
Undergraduate Institution: University of California, Los Angeles
Hometown: Solana Beach, California
Website: LinkedIn
Biography
Giselle Brown, a rising senior at the University of California, Los Angeles, is studying Chemistry with a concentration in Computing in the hope of pursuing a PhD in Organic Chemistry. She is currently a Clare Boothe Luce Scholar and Queens Road Foundation Fellow for her research in Abigail Doyle’s group. She joined this group after spending two summers designing and synthesizing novel drugs in immuno-oncology for a small-molecule biotech. Giselle’s research interests are in green chemistry-minded catalysis and its use in drug discovery. In the Doyle Group, she does research combining high-throughput experimentation and machine learning to develop generalized conditions for the Chan-Evans-Lam reaction and develop a method for enantioselective oxidative radical polar crossover. Giselle is involved in the UCLA community through her positions as Green Chemistry Chair for Student Members of the American Chemical Society, which works to connect chemistry and biochemistry students to each other and the department, and as the co-president of SAGE, a sustainability club on campus that does corporate consulting and environmental legislation outreach. She is committed to connecting women and LGBTQIA+ individuals to opportunities that will further their careers in STEM. Outside of the lab, Giselle enjoys climbing, exploring new coffee shops, and visiting museums.
Abstract
Mechanistic Insights into Amide Coupling Reactivity Cliffs
Giselle Brown1, Jihye Roh2, Babak Mahjour2, Connor Coley2, Masha Elkin3
1Department of Chemistry and Biochemistry, University of California, Los Angeles
2Department of Chemical Engineering, Massachusetts Institute of Technology
3Department of Chemistry, Massachusetts Institute of Technology
Amide couplings are the most common reactions in the synthesis of pharmaceutical compounds. However, their often-limited reactivity and wide range of conditions leads to poor atom economy, meaning wasted chemicals. Efforts to predict the yield of amide bond formation as a function of substrate structure and reaction conditions have failed to result in a predictive model. It is hypothesized that one of the major reasons for the difficulty of predicting the yield of amide couplings is the presence of “reactivity cliffs”—pairs of substrates that are structurally similar, but undergo amide coupling in very different yields. To go about elucidating these reactivity cliffs from the CAS database of 8 million reaction, Tanimoto and Cosine similarity models were used. These models calculate molecular similarity based off molecular fingerprints and were used to find the “reactivity cliffs”, along with filtering for a large yield difference and commercial availability. Mechanistic experiments, like doping in an incompatible functional group, can be used to investigate what long-range interactions might be causing the decreased reactivity. By quantifying these affects, models can be better fit for subtle differences in structure and be able to make more substantiated predictions to aid in smoother lead optimization campaigns. The solubility of substrates was one factor discovered to be an important in creating a model sensitive of “reactivity cliffs”.
