Multivalent interactions are a hallmark of biological liquid-liquid phase separation (LLPS). However, many molecular systems with multivalent interactions do not undergo phase separation, suggesting that multivalency alone is insufficient to drive the process. The goal of my project is to determine the physical determinants of LLPS of multidomain proteins and to use these principles to directly interrogate the function of biomolecular condensates in signaling pathways. Empirical and theoretical studies from synthetic polymers have shown that weak self-self interactions are a driving force for LLPS, which is reflected in the experimental thermodynamic parameter; the second osmotic (or scattering) virial coefficient. I hypothesize that these weak self-self interactions, which are also present in structured biomolecules, drive LLPS and can be systematically tuned to promote or inhibit the formation of biomolecular condensates. I am testing these principles on an engineered system composed of several Small Ubiquitin-like Modifier proteins (SUMO), which phase separates readily with the addition of SUMO interaction motifs (SIMs). I will apply principles learned from the engineered system to a natural signaling pathway that is activated upon antigen recognition by T cells. Components of this pathway, including the adaptor protein, LAT, and its various ligands assemble into phase separated clusters upon T cell activation. I will manipulate weak self-self interactions of two critical signaling molecules in this system, Grb2 and Nck, in order to favor or disfavor phase separation without changing interactions known to be essential for signaling (e.g. SH2-pTyr and SH3-proline). I will then use these molecules to examine the specific role of phase separation in T cell signaling. Overall, I hope to provide insights into the physical mechanisms that drive formation of biomolecular condensates and explore whether these mechanisms can be used to manipulate cell function in novel ways.