Project Summary Biological macromolecules exhibit an amazing degree of conformational heterogeneity as required for their various functions. This heterogeneity is inherent in almost all biological macromolecules and is influenced by interactions with ions, solutes, and other macromolecules. An understanding of the conformational properties of macromolecules and their structure-function relationships is essential to facilitate drug discovery. To this end, our laboratory has focused on a comprehensive research program that includes optimizing and extending empirical atomic force fields for biological and drug-like molecules, developing novel conformational and solute sampling methods and applying those tools in collaborative studies on systems of biological and therapeutic relevance. Biological systems to be studied include the role of Mg2+ ions in RNA folding and on conformational ensembles of RNA, glycan conformational properties and the role of glycoproteins in immune response involving antibodies (Ab) and Fcγ receptors such as FcγRIIIa and FcεRIγ. Interactions between Ab Fc and FcγRIIIa and interactions between Fcγ receptors will be investigated to reveal their contribution towards intracellular phosphorylation and downstream signaling. Methods developments will further optimize and extend the additive (nonpolarizable) CHARMM36 and polarizable classical Drude oscillator force fields (FF). Work on the additive FF will involve the explicit optimization of the off-diagonal Lennard-Jones (LJ) terms to overcome limitations in the use of combining rules. Drude FF work will involve extending the coverage of chemical space including non-standard amino acids, nucleic acids, and carbohydrates, as well as drug-like molecules in the context of the Drude General Force Field (DGenFF). In addition, advances in the accuracy of the Drude FF will include improved treatment of off-diagonal LJ and non-bonded Thole shielding interactions, improved treatment of ions in biological systems, and additional optimization of the protein and nucleic acid parameters of the FF targeting the equilibrium between folded and unfolded states of peptides and short oligonucleotides. This optimization will use enhanced sampling to facilitate generation of conformational ensembles required to apply parameter reweighting approaches. Improved conformational sampling of macromolecules will be achieved using generative deep neural nets (DNN) for the identification of reaction coordinates in conjunction with metadynamics as well as non-equilibrium simulation methods targeting specific degrees of freedom with emphasis on oligonucleotides and polysaccharides. Our solute sampling approach based on the oscillating chemical potential μex Grand-Canonical Monte Carlo/Molecular Dynamics method will be extended to large systems (> 1 million atoms) through efficient use of GPUs, improved treatment of long-range interactions, and the use of the polarizable Drude FF. These developments will be implem...