(2fx) Design and Engineering of Molecules Using Molecular Simulations and Machine Learning

The central theme of my research interests is in the development and application of inverse computational methods to solve problems in biophysics, chemistry, and materials science. To this end, I envisage in developing a cross-disciplinary computational design & engineering group working in close collaboration with experimentalists and theorists to address impactful societal problems. My current interests lie in problems requiring innovative and automated solutions for enhancing health and reducing illness, and sustainable materials development. Specifically, my projects are designed to tackle the challenges in engineering biosensors, materials for purifying biologics, therapeutics utilizing cellular trash collector, biopolymer properties and biomaterials. This utilizes my combined expertise in molecular dynamics (MD) simulations, enhanced sampling methods, and machine learning developed during my doctoral and postdoctoral training with Prof. Sapna Sarupria at Clemson University and Prof. Andrew L. Ferguson at The University of Chicago, respectively.

Teaching Interests

My interest in teaching stems from my own passion for learning and through the inspiration for teaching and research conveyed by all my teachers and mentors. Furthermore, I enjoy teaching because it presents an opportunity to inspire and interact with a new generation of engineers and scientists. I am interested in teaching both undergraduate and graduate level transport phenomena, applied numerical methods, chemical engineering thermodynamics, statistical mechanics, molecular simulations, and data science applied to chemical engineering classes. I am well prepared to teach molecular simulations related courses with a focus on its practical applications in chemical engineering, materials science, and biophysics. These classes will include topics such as — atomic-scale modeling and simulations, coarse-grained modeling methods, developing force fields, integrator algorithms, thermostat and barostat methods, advanced sampling methods, Markov state models, and recent applications of machine learning in molecular simulations. Furthermore, topics related to numerical methods, scientific computing, and best practices in developing robust and scalable scientific software will be covered. These courses will prepare students for a variety of computational science careers.

Abstract

Molecular simulations can provide detailed insights into the governing forces and mechanisms of various complex systems. When coupled with machine learning, molecular simulations can enable high throughput discovery of optimal molecules while providing insights into their design rules for various applications such as development of effective therapeutics, engineering robust catalysts, and discovery of novel sustainable materials. In our presentation, we will discuss the development and application of one such strategy named computational active learning comprising molecular dynamics (MD) simulations, enhanced sampling methods and machine learning (ML) for (a) designing switchable materials and (b) engineering probes for detecting forever chemicals in water.

Polarizable nanoparticles (NPs) have rich and complex phase behavior that can be used for designing switchable self-assembling materials for applications in energy storage, optoelectronic devices, and drug delivery. To realize such applications, understanding the design space of NPs is necessary. While conventional computational methods are suitable, inordinate computational resources are required for (i) simulating polarization forces and (ii) exploring the vast design space. Our goal is to address these problems by using image method to efficiently and accurately simulate polarization forces and developing a computational active learning framework comprising coarse-grained MD simulations, umbrella sampling method, and ML. We demonstrate our approach to understand the regions of spontaneous self-assembly between two NPs along the experimental design space. These maps quantitatively predict the multiparticle NP self-assembly and are in good agreement with experimental scattering measurements. We use the computed maps to engineer self-assembly of polarizable NPs that are capable of triggered assembly or disassembly via temperature and solvent quality with potential applications in sensors, smart windows, and drug delivery.

As another example, we will show the development and application of a computational active learning approach for efficiently engineering probes to detect forever chemicals in water. Because the molecular design space of probes to selectively bind with a given target forever chemical is astronomical, conventional computational techniques are hardly intractable. To address this problem, we develop an active learning framework involving all-atom MD simulations, parallel-bias metadynamics, deep representational learning of probes and multi-objective Bayesian optimization. We demonstrate our approach to engineer linear probes for selectively binding with a harmful forever chemical named perfluorooctanesulfonic acid (PFOS) in the presence of a common interferent sodium dodecyl sulfate (SDS). By searching just ~6% of the search space, our results reveal the linear probes with optimal sensitivity and selectivity to PFOS. We found that the sensitivity of linear probes increases by ~0.5 k_BT with an additional methylene group in the probe. However, the linear probes have weak selectivity (~1 k_BT), thereby revealing their limitations as practical probes to selectively detect PFOS in the presence of SDS. In our presentation, we will discuss the development of the active learning framework and present the discovered optimal linear probes and their design rules. Finally, we will discuss the applications of the discovered probes for engineering more complex non-linear probes that can be used for developing efficient water treatment systems.