∆9-Tetrahydrocannabinol (THC), the primary psychoactive compound in Cannabis, is responsible for the experience known colloquially as “being high.” Considering its alarmingly high rates of usage, THC’s effects on movement behavior are insufficiently studied. My project addresses this crucial gap in our knowledge by investigating the dose-dependent effects of THC on movement behavior using mouse models in tandem with novel behavioral neuroscience techniques. My research aims to establish a preclinical model for THC-induced impairment, focusing on studying its impact on locomotor control. My main experimental tool is a behavioral linear track, which is a clear glass corridor with a 45 degree-angled mirror placed beneath it. The linear track allows us to create a standardized multi-dimensional environment in which mice are recorded after they are treated with either a control or variable doses of THC. The videos taken of the mice are then analyzed using SLEAP. SLEAP is a machine-learning, pose-estimation algorithm that I helped train to track individual points of interest on the mice, such as the nose, paws, and tail. Behaviors of interest, such as walking, rearing, and grooming, are classified by a random forest algorithm that analyzes SLEAP label data to output identified behaviors. This data is then tabulated and graphed to reflect the dose-dependent changes in behavior elicited by THC. These classifications are also used to further analyze metrics during a represented behavior. For instance, for a walk event, we can utilize positional data from SLEAP to calculate and measure kinematic features such as stride length and limb speed, allowing us to distinguish between an unimpaired and an impaired walk. This computerized analysis approach minimizes human bias, reduces error, and produces exhaustive data that can characterize subtle differences in behavior, like when comparing mice exposed to low THC doses of 0.1mg/kg and 0.3 mg/kg.

VirScan, a revolutionary technology based on Phage Immunoprecipitation Sequencing (PhIP-Seq), allows the interrogation of antibody responses to all known human viruses using a small blood volume, providing information on an individual's previous viral exposures. This study aims to provide a comprehensive data quality assessment system for VirScan, which will improve its reliability and interpretability by routinely assessing VirScan data quality at both the sample, assay (N=96 samples in replicate), and sequencing batch levels (N=192 samples in replicate). The study focuses on creating standards and thresholds for data quality at all three levels, considering aspects such as aligned reads, read depth, percent of epitopes discovered, and correlation of sequence counts between replicates. The assay/batch-level analysis provides metrics like the mean, median, standard deviation, and range of mapped reads and correlations for count and peptide detection, evaluating consistency, accuracy, and comparability across assays and batches. Further, these criteria can effectively categorize sample quality into Good, Questionable, and Failed, identifying samples that may need to be repeated or excluded from analysis. These quality calls were all encoded within an R Shiny App, enabling a user-friendly and flexible interpretation of VirScan data. Implementing this systematic quality control strategy will considerably improve the usability of VirScan in research and clinical contexts, allowing for more trustworthy interpretations of an individual's viral exposure history while also contributing to a better knowledge of immune response dynamics.

The immune system and inflammatory responses to viral infections are regulated by molecular circadian rhythms in mouse models. Mice infected with influenza just prior to their active phase have a mortality rate four times higher than mice infected just prior to their rest phase. As a result, circadian rhythms are hypothesized to regulate viral replication and early immune responses in airway epithelia during viral infections. Prior work has shown circadian cycles regulate gene expression in human epithelial cells. However, the influence of time of infection on viral replication in human airway epithelia has not yet been explored. We hypothesized that circadian-synchronized human airway epithelial cells would demonstrate differential viral replication and immune responses when infected at two different times of day. To address this gap, we differentiated primary epithelial cells from healthy children at an air-liquid interface to create an ex vivo cellular model of the human airway. Airway epithelial cells underwent circadian synchronization using temperature cycled incubators and were exposed on the apical surface to human rhinovirus-16 at time 0 and 12 hours during a circadian cycle. The RNA from seven total cell lines was sequenced and viral genome copy number was quantified at hour 96 following infection using GeneSig qPCR. Infection at hour 12 led to two-fold higher viral genome copy number 96 hours after infection as compared to hour 0. Infection late in the circadian phase (time 12) leads to increased viral replication at the airway epithelium and may explain the difference in mortality in mouse models of viral infection. Ongoing work is investigating immune responses based on time of infection. In the future, we will investigate changes in circadian regulation of viral infection in airway epithelia from healthy children and children with airway diseases such as asthma.

Myotonic dystrophy type 1 (DM1) is a genetic disease that causes many serious health conditions in a variety of tissues including skeletal muscle stiffening and cardiac conduction disorders. This disease affects 1 in 2,300 people worldwide and is the most common form of muscular dystrophy. DM1 is caused by a CTG repeat expansion, which in lay terms means that in a gene, there's a sequence of 10 CTG DNA bases. However, in a specific part of the gene responsible for making messenger RNA (mRNA), the number of CTG repeats increases significantly. This unusual mRNA sequence is linked to the development of the disease. This mutated mRNA (messenger RNA) disables the splicing regulator muscle-blind-like 1 (MBNL1) gene and ultimately causes disease. It does this by sequestering and limiting the MBNL1s critical role in splicing mRNA (figure 1). In my proposed research project, I am focusing on cardiac function when testing adeno-associated viral vector (AAV)-mediated systemic delivery of the MBNL1 gene to increase MBNL1 protein expression in muscle. The lab found that body-wide delivery of AAV vectors with CK8-intron-MBNL1, which expressed MBNL1 only in striated muscle, was toxic in the hearts of mice and caused death (figure 2). Over the last few months, my mentor Matt Karolak and I have learned together methods such as echocardiography and tissue histological techniques to determine whether it is possible to prevent MBNL1 protein production and its damaging effects in the heart while still expressing MBNL1 protein in skeletal muscle for therapeutic disease benefits.

What is personal space? It describes the “invisible bubble” immediately around an individual, where intrusion by others may feel uncomfortable or even threatening. Despite personal space being important for determining the size and dynamics of large groups and how animals behave, it is unknown how animals generate and maintain their sense of personal space. I explore this question using Drosophila flies, a powerful model system for studying behavior. Individual flies communicate via a large set of behaviors (e.g. kicking, flicking the wings, flying, and running) that can shape group interactions. Between pairs of flies, there is a rich repertoire of interactions that cause them to either move away (e.g. appendage touching) or move closer (e.g.courtship behaviors). To capture these behaviors, I use high resolution cameras running at 150 fps to collect videos of flies interacting in a circular arena and deep-learning software for annotating body parts (SLEAP). I use this data to track the poses of multiple flies throughout time and quantify appendage touching behaviors in addition to an estimate of each fly’s personal space. Using optogenetics, I model a situation in which a wild-type fly responds and adapts to being surrounded by other flies that crowd tightly. One expected outcome is that the wild-type fly will increase appendage touching in response to the space around itself shrinking. My goal is to determine exactly how the dynamics of such inter-fly interactions form a sense of personal space for each fly. My work will uncover the patterns of interactions that develop personal space and how these patterns scale out to larger social networks.