Abstract:

Quantum sensing leverages the extreme sensitivity of quantum systems like ultracold atoms and molecules to probe minute changes in their environment beyond classical limits. This sensitivity becomes especially pronounced near exceptional points (EPs)–critical parameter regimes where eigenstates coalesce to produce nonlinear responses, nonreciprocal dynamics, and phase transitions. For instance, quantum scattering EPs correspond to invisibility points where particles pass through barriers unimpeded. We aim to advance high-precision quantum molecular sensors utilizing these properties by establishing experimentally realizable EPs for quantum particles interacting with one-dimensional potential barriers. We analyze the triple-Gaussian potential due to its parity-time reversal-symmetry phase transition and simple optical implementation. Particle data is computed by numerically solving Schrödingerʼs equation and benchmarked with the triple-rectangular potential using the analytic Transfer Matrix Method. We demonstrate theoretically the successful realization of EP invisibility effects with a triple-Gaussian potential, verifiable with an ultracold atom experiment using rubidium-87 Bose-Einstein condensates and spatial light modulators. A key finding is that raising the central Gaussian barrier out of the three merges EPs, acting as a quantum filter by selectively permitting certain energies to pass. Additionally, we discover novel EPs for the triple-delta potential, a theoretical system consisting of three infinitesimally thin barriers acting as a direct atomic or molecular analog of optical systems currently under investigation for high-precision quantum sensors. Understanding EPs has broad applications, including sensitive quantum sensors for early earthquake detection, noninvasive biosensing through “invisible” nanoparticles for effective diagnostics and treatments, and enantio-sensitive quantum control for chemical reactivity. This project lays the foundation for further EP studies, with future work incorporating thermal noise to support experimental testing and modeling three-dimensional systems to develop practical atomic and molecular quantum technology.

Speaker:

Shrikar Dulam, University of Illinois, Urbana-Champaign

Shrikar Dulam is an undergraduate student at the University of Illinois Urbana-Champaign, pursuing a degree in Computer Science + Physics. His research interests lie at the intersection of quantum physics and computational science, with a focus on quantum computing and quantum sensing.

Abstract:

Respiratory viruses such as human rhinovirus (HRV), influenza, and respiratory syncytial virus (RSV) are leading causes of global illness, particularly affecting vulnerable populations. Despite their impact on morbidity and mortality, antiviral treatment options remain limited and increasing drug-resistant viral strains emerge. Studies suggest that L-type calcium channel blockers (CCBs), commonly prescribed for cardiovascular conditions, may inhibit viral replication by disrupting calcium ion homeostasis, a process critical to the viral replication cycle. This undergraduate thesis investigates the repurposing potential of CCBs, particularly amlodipine and its derivatives (“AMP” compounds), as antiviral agents through structure-based virtual screening (SBVS).

The initial phase (Jan-Nov 2024) focused on HRV-16, targeting the VP1 capsid protein (PDB: 1ND3) using Schrödinger software. A receptor grid was developed based on the known HRV inhibitor pleconaril. CCBs and AMP compounds docked with scores comparable to pleconaril, and seven top candidates were identified for potential therapeutic development.

Building on these findings, the second phase (Spring 2025) extended the screening to the neuraminidase (NA) protein of influenza (PDB: 2HT7, 4KS5) and the fusion (F) protein of RSV (PDB: 6VKD, 7KQD). AMP compounds docked with stronger scores than known active ligands at the NA active site, and docking scores significantly correlated with in vitro inhibition for the 2HT7 grid (p=0.0011). QSAR analysis predicted favorable drug-like properties of AMP compounds. Enrichment performance was assessed using active compounds and DUD-E generated decoys on the 2HT7 grid, demonstrating modest discrimination between actives and decoys (ROC AUC=0.61), and limited early enrichment (RIE=0.54). Docking to RSV F protein models showed no significant correlation with biological activity, however, AMP ligands docked with scores rivaling those of known active ligands.

This research highlights the potential of repurposing L-type CCBs as broad-spectrum antivirals. These findings support a novel antiviral strategy leveraging existing drug scaffolds to address critical gaps in respiratory virus therapeutics.

Speaker:

Aiden T. Day, Saint Joseph’s College of Maine

Aiden T. Day is a recent graduate of Saint Joseph’s College of Maine, where he earned a Bachelor of Science in Medical Biology/Pre-Medicine with a minor in Chemistry. He is an aspiring surgeon-scientist with a deep commitment to advancing patient care through translational research. His academic and research background spans medicinal chemistry, public health, and neurosurgery.

Abstract:

Polyethylene terephthalate (PET) is one of the world’s most common thermoplastics. Its accumulation has detrimental impacts on the environment and human health. Currently, mechanical recycling is the best way to reduce PET in the environment. However, it consumes significant energy while emitting greenhouse gases, and it can only be performed a limited number of times. Biodegradation techniques are being explored to find more sustainable, efficient PET recycling methods. One such technique, enzymatic recycling, will only be feasible at an industrial scale once effective crystalline PET-degrading enzymes are developed. This project aims to design and express a PET hydrolase that exhibits the characteristics associated with stronger crystalline degradation activity — a hydrophobic, linear, and wide active site cleft, and increased thermostability. Amino acid residue mutations were introduced to the sequence of an existing hydrolase (PDB ID 7QJP) to enhance the desired characteristics. The resulting mutants were modeled using AlphaFold2 (a neural network for protein structure prediction) on Google Colab. In Schrödinger Maestro, they were then evaluated using the following features: protein structure alignment, protein preparation, docking (to a PET fragment), and ligand interaction diagrams; additionally, the cysteine mutation feature was used to identify potential disulfide bonds. Finally, DeepSTABp (deep learning software) was used to predict the melting temperatures. This resulted in 80 promising mutants with the desired characteristics. A preliminary transformation (heat shock) and expression were conducted for the top-performing mutant; it was successfully expressed, indicating that it’s a viable enzyme. Optimized induction conditions will yield enough for it (and similar mutants) to be harvested and used in an assay to confirm its degradation efficiency. This has the potential to allow large-scale implementation of enzymatic recycling, offering a more sustainable way to recycle PET plastic.

Speaker:

Mena Boggs, NCSSM/NC State University

Mena Boggs is an incoming freshman at North Carolina State University, where she plans to major in Chemical Engineering, Computer Science, or Bioinformatics. She is passionate about environmental protection, which inspired her research on plastic-degrading enzymes.

Abstract:

The widespread use of Machine Learning (ML) techniques in chemical applications has come with the pressing need to analyze extremely large molecular libraries. In particular, clustering remains one of the most common tools to dissect the chemical space. Unfortunately, most current approaches present unfavorable time and memory scaling, which makes them unsuitable to handle million- and billion-sized sets. Here, we propose to bypass these problems with a time- and memory-efficient clustering algorithm, BitBIRCH. This method uses a tree structure similar to the one found in the Balanced Iterative Reducing and Clustering using Hierarchies (BIRCH) algorithm to ensure O(N) time scaling. BitBIRCH leverages the instant similarity (iSIM) formalism to process binary fingerprints, allowing the use of Tanimoto similarity and reducing memory requirements. Our tests show that BitBIRCH is already > 1,000 times faster than standard implementations of the Taylor-Butina clustering for libraries with 1,500,000 molecules. Furthermore, we explore strategies to handle tremendously larger sets, which we applied in the clustering of one billion molecules under 5 hours using a parallel/iterative BitBIRCH approximation. To the best of our knowledge, this is the first time a billion molecules have been clustered. BitBIRCH increases efficiency without compromising the quality of the resulting clusters, which we were able to conclude by assessing for the Calinski-Harabasz and Davies-Bouldin indices. We also explore further applications of the BitBIRCH algorithm in chemical space exploration, segmentation, and in drug-design pipelines. The problem of clustering billions of molecules was thought to be unsurmountable with current algorithms, but BitBIRCH not only makes this possible, but accomplishes this task in just a couple of hours.

Speaker:

Vicky Jung, University of Florida

Vicky Jung is an undergraduate student at the University of Florida majoring in Data Science with a minor in Bioinformatics. She works in the Miranda-Quintana Lab where she mainly assisted with BitBIRCH, a novel approach to clustering in cheminformatics.