Synergistic approach to probing the dynamics and mechanics of patchy soft matter

Tailoring microscopic details to tune bulk rheology is a key paradigm in soft matter physics, yet the vast parameter space associated with constituent interactions precludes a fully systematic approach. To address this, we have designed a synergistic strategy to explore the parameter space that comprises simulations, experimental rheology, and machine learning. As a case study, we choose DNA-based self-assembled fluids whose viscoelastic response can be fine-tuned by manipulating the base sequencing of the constituent nucleic acid nanostars. We use coarse-grained simulations, benchmarked against experimental data, to obtain the rheology of the DNA fluids, which feeds forward to a framework of Gaussian Process Regression and active learning. The latter is then used to explore the rheological design space with high predictive precision. The pipeline is designed to be deployed iteratively for the rational design and accelerated discovery of generic soft matter suspensions.