Vascular malformations (VM) are congenital malformations of veins, arteries, capillaries, and/or lymphatic vessels. They are primarily caused by post-zygotic activating mutations within oncogenes of the PI3K-MTOR-RAS-MAPK pathways. The major types of VM include capillary malformations (CM), lymphatic malformations (LM), venous malformation (VeM), and arteriovenous malformations (AVM). Identifying the molecular etiology of VM is important, as there are increasingly targeted medical therapies. However, current molecular diagnosis requires surgically excised tissues, as the causative mutations are somatic and not present in blood cells. Cell-free DNA (cfDNA) is an emerging diagnostic analyte that is already in wide use in cancer diagnosis and prenatal genetic screening.  Given the role of mutant endothelial cells in the development of VM and their proximity to blood and lymphatic fluid, we hypothesized that mutant cfDNA will be detectable in plasma or cyst fluid from individuals with VMs. This project aims to develop a non-invasive methods to identify genetic causes of VM without surgery to reduce the costs and risks relating to surgical biopsy. Utilizing digital droplet polymerase chain reaction (ddPCR) technology, we are able to amplify cfDNA fragments to detect specific well known “hotspot” DNA mutations in PIK3CA, BRAF, TEK, and others. We continue to determine the sensitivity of allele-specific cfDNA analysis in individuals with known, tissue-based mutations from a retrospective cohort, and test prospective cfDNA from VM patients with unknown mutations using both ddPCR screening and/or a high sensitivity gene sequencing panel. Ultimately, our hope is to bring these techniques into the clinical arena to increase the possibility of medical therapies for individuals with vascular malformations.

Platelets aggregate at the site of injury to stop bleeding, but disruptions to hemostasis can cause bleeding or thrombosis. Studying platelet-plug area and contractile force can predict whether bleeding or thrombosis is likely to occur. Microfluidic devices, composed of polydimethylsiloxane (PDMS), are used to study these biomechanics by inducing aggregation through shear flow. As blood flows through the device, the platelets pass over a rigid block in the channel which causes platelets to activate, deflecting a flexible post within the channel. This deflection is used to calculate the platelet forces based on the material properties of the PDMS. These microfluidic devices are single use and require a fabrication process that spans multiple days. Additionally, creating duplicate silicon master molds is a laborious and expensive process that necessitates cleanroom training. I am engineering and implementing a more efficient process for the fabrication of these devices, while limiting the use of the master mold that undergoes long-term wear from repeated uses. My focus is on improving the efficiency of the initial negative mold creation process by using a different material, urethane resin, to replicate the master, which allows us to make more negatives simultaneously without needing to fabricate another silicon master. To compare the devices produced using the onyx master with the silicon master, I am running three blood experiments, each with varying levels of antibodies, on two devices fabricated by the onyx and silicon masters. The aggregation sizes and forces are being observed between each of the experiments. I expect the results to be similar within a degree of certainty, proving the onyx master is equivalent to the silicon master and can be used to increase microfluidic device production and increase the availability of platelet biomechanics studies.

Morphine and other mu-opioid receptor (MOR) ligands are commonly used treatments for pain management. Long-term administration of opiates results in receptor desensitization and tolerance, inhibiting treatment efficacy and contributing to the current opioid epidemic. Following opioid administration, peroxiredoxin 6 (PRDX6) is recruited to the opioid receptor complex through JNK (cJun N-terminal Kinase)-dependent signaling. PRDX6 activation generates reactive oxygen species (ROS), resulting in opioid receptor desensitization and one form of opioid tolerance. Recently, cannabidiol (CBD) has been implicated in decreasing acute morphine tolerance. The present study will establish a connection between morphine and CBD cotreatment with the hypothesis that cotreatment will decrease ROS production. To measure ROS, I will employ the genetically encoded ROS sensor HRM63, which fluoresces proportionally to ROS production. Because ROS is a product of JNK/PRDX6 signaling, measuring the strength of ROS-dependent fluorescence is a simple way to visualize my hypothesis using in vitro cell models. In each experiment, treatments of morphine, CBD, or combined morphine and CBD will be delivered to plated HEK293 cells stably expressing both MOR and HRM63. I will then image the cells using fluorescence microscopy to quantify ROS response to treatment. I predict that cotreatment of CBD with morphine will result in lower ROS production compared to morphine treatment alone. These results will be crucial in the ongoing characterization of CBD's role in opioid tolerance.

Autosomal dominant polycystic kidney disease (ADPKD) is a genetic disorder that affects people who have mutations in the PKD1 or PKD2 genes. The prominent feature of this disease is the formation of cysts in the kidneys. Despite ADPKD being genetically heterozygous, cystic tissue from patients with ADPKD also contains cells with homozygous mutations. This causes speculation within the PKD community about whether heterozygous mutations with one mutated copy of the PKD1 or PKD2 gene are sufficient to induce cystogenesis or if two mutated copies are needed. This concept is referred to as the two-hit hypothesis, a somatic mutation (second hit) occurs during the lifetime of the kidney and gives rise to clonal populations of cells that give rise to cysts. The goal of my project is to investigate the two-hit hypothesis and determine if there is cyst formation in the heterozygous organoids. To do so, I use patient-derived hSPC lines modified using CRISPR to express PKD1 or PKD2 mutations. I follow a 21 day differentiation process to produce PKD organoids. I then pick individual organoids to place in suspension for 14 days and track cyst growth by area and amount. If cysts form in heterozygous PKD organoid models at the same rate as the control, it would suggest that heterozygous mutants are unlikely to cause cyst formation, indicating the two-hit hypothesis applies to PKD. Current data collected on mutations in PKD2 have shown heterozygous organoids do not form cysts, however more data is needed to test the PKD1 site. Future applications involve inducing cyst formation in heterozygous mutants using forskolin, an agonist of adenylyl cyclase shown to induce cyst growth. Establishing if heterozygous mutant organoids form cysts has significant implications in ADPKD research by potentially creating a better representation of the disease process, which has been difficult to study in other models.

This research compares the COVID-19 control policies and the use of Contact Tracing Apps (CTAs) in Shenzhen, China, and Seattle (King County, WA). The research examines the differences between local government's policies of pandemic control, contact tracing response, and the discrepancies between CTAs design, mechanism, and the ways of defining close contacts. This first part provides background for a better understanding of the differences in government policies and CTAs for the two metropolises. The second part of the research analyzes data available to show the effectiveness of contact tracing apps and the results of differences in government pandemic control policies and response time. Visualizations are created using existing data and construct timelines of the pandemic and usage of CTAs, demonstrating the effects of CTAs on certain indicators such as numbers of hospitalizations, people in quarantine, cases, and deaths. The timeline will be divided based on different COVID-19 mutations that were dominant in the areas, as well as the version updates on the CTAs. After analyzing the effectiveness of CTAs on pandemic control, the research explores the overall impacts of CTA, COVID-19, and government’s pandemic responses on society as a whole, including the inequalities between races, different income groups, and migrations. This research may generate insights on the possible policies and ways of utilizing CTAs that are beneficial to the pandemic control as well as the economy and residents' daily life. Besides, as there are few studies on CTAs’ effectiveness on COVID-19 control, this study can also suggest possible directions for other researchers. In the end, the research generates visualizations that tell a story from the beginning of how the government deals with the pandemic and how their responses and CTAs had an effect on the pandemic control and the society, especially on migration and inequalities.

Madjedbebe is the earliest known site of human activity in Australia with artifacts dating to 65,000 years ago. However, the presence of termite mounds in the landscape of the site have led some to question this notable date. Termites have the potential to affect the stratigraphic integrity of archaeological sites, their subterranean movements mixing sediments from deposits of different ages. Such activity may have led to the overestimation of the age of artifacts at Madjedbebe. Our study investigates the extent to which termites disturbed Madjedbebe’s archaeological deposits, thereby providing key insight into the reliability of the ages of the site’s sediments and artifacts. We used micro X-ray fluorescence (μXRF) to quantify and compare elemental concentrations in blocks of resin-impregnated archaeological sediments and sediments collected from termite mounds. We analyzed the μXRF data using log ratio analysis. Our research seeks to identify whether chemical traces of termite activity exist within sediments at Madjedbebe. Our results indicate that termite-impacted sediments have distinctive chemical signatures that distinguish them from archaeological deposits; generally, this signature is not present in any of the Madjedbebe samples. There is therefore little evidence for termite activity and mixing in the archaeological deposits. Our findings are indicative of high stratigraphic integrity at the site, suggesting that the proposed 65,000-year-old ages of the Madjedbebe artifacts are sound. This is important for understanding the timing of the arrival of humans in Australia, the earliest movement of humans out of Africa, and the chronology of human interactions with Denisovans and Neanderthals. Our research also demonstrates the applicability of material science methods to archaeological questions.

Why do identical twins have different lifespans? Beyond genes, what else might influence the aging process? Variation in any phenotype is due to the combined effects of genetic variation and environmental variation. In fact, there are two types of environmental variation—one is extrinsic environmental variation, such as food, temperature, etc., and the other is intrinsic environmental variation, which can lead to subtle differences in behavior, such as how much an individual eats, how long it sleeps, etc. We hypothesize that these differences can be predicted by an individual’s underlying metabolism. There are thousands of molecules that make up the structural and functional building blocks of all organisms, a domain known as the metabolome. Previously, many studies have shown that genotypes vary in lifespan, but even within a single genotype there is enormous variation in lifespan. Here we address how intrinsic environmental variation influences aging by controlling the genetic and extrinsic environmental variation under lab conditions. We designed an experiment using Drosophila melanogaster, and since Drosophila has a natural tendency to climb upwards against gravity, and climbing ability of flies decreases with age, we hypothesized that we might use climbing ability as a biomarker of future mortality risk. Using mid-life climbing ability, we separated genetically-identical flies and then analyzed each group’s lifespan. We found that within a genotype, strong climbers had a longer lifespan than non-climbers. Finding strong support for this hypothesis led us to propose that the metabolome between climbers and non-climbers might be different. Our goal is to understand the role of intrinsic variation in aging. If we can find metabolites that associate with climbing ability, and as we have shown, climbing ability is associated with aging, we might be a step closer to explaining how intrinsic environmental variation influences aging.

Organisms are required to maintain genome stability for the correct propagation of genetic information. Glycation damage is one of the most important types of DNA damage that cause genome instability. Glycation damage in organisms is caused by the covalent attachment of parts of sugar molecules to proteins and DNA. Accumulation of the Advanced Glycation End (AGE) products may cause Parkinson's disease, cancer, and other oxidative stress-induced diseases. The protein deglycase, known as DJ-1 in plants and Parkinson Disease Protein 7 (PARK7) in animals, can prevent glycation damages in many organisms. In maize, plastid genome stability is maintained by Whirly ssDNA-binding proteins. Our lab recently showed that the demise of plastids and their DNA is associated with increased DNA damage due to oxidative and glycation damages during maize seedling development. Here, our objective is to understand the role of Whirly proteins in glycation damage. Our approach involves quantifying the glycation and deglycation levels in maize plastids. We isolated plastids and their proteins from the wild-type (wt) and whirly (why) mutant maize seedlings and performed glycation and deglycation assays. We find a significant difference in deglycation levels between wt and why mutant plants. Our study should provide a better understanding of the role of ssDNA binding proteins in glycation damage. 

Throughout our lives, we generally base our idea of age upon someone’s ‘chronological age’, or how many years they’ve been alive. This, however, is not always the best indicator of aging, as people reach social and biological milestones at different ages. As an alternative, someone’s ‘biological age’ can be more representative of their progression through life. As such, research has focused on identifying biomarkers of biological age to help us better understand aging. Recent work in our lab has sought to determine the impact of several metabolites - biomolecules used for metabolism - on the biological age of the fruit fly, Drosophila melanogaster. Among the metabolites studied, histamine - a neurotransmitter involved in wakefulness and visual processing - had one of the strongest correlations with lifespan, suggesting that it plays a role in aging. In this study, we attempted to manipulate the biological age of female D. melanogaster by altering either their metabolic levels of histamine, or their perception of histamine. To do this, flies were given food supplemented with histamine or with the antihistamine hydroxyzine, a competitive inhibitor of histamine receptors. These experimental conditions were compared to control food that lacked additives. Treatment was administered continuously starting at 4 weeks and the lifespans of flies in each condition were measured. Based on our previous results, we expected to see a negative effect of added histamine on lifespan and an increase in lifespan in response to antihistamine. Our study could highlight histamine’s role in aging and lay the foundation for demonstrating that biological age can be influenced by a single metabolite.

Patterns in three dimensional artefact orientation can provide insights into the environmental and cultural processes that form archaeological deposits. A range of post-depositional processes can be inferred from non-random patterns of artefact orientations. We apply statistical and graphical techniques to artefact orientation data from Madjedbebe, a site in northern Australia with evidence of human activity at 65 thousand years ago. The three-dimensional positions of artefacts were recorded in situ using a total station (survey equipment), and analysed following McPherron’s (2018) methods. Through this I investigate how accurate the site's dates are through analyzing deposition processes and site formation processes. I compare the orientations of the archaeological artefacts to orientations of experimentally trampled artefacts, and to orientations from simulated slopes and terrains. I found that the archaeological orientations are significantly different from the trampled artefacts, which have a strongly linear orientation. The orientations of archaeological artefacts are most similar to simulated orientations from sloped and irregular surfaces. The upper phases at Madjedbebe present more isotropic orientations than the lowest phases, which are strongly planar. These patterns suggest minimal post-depositional disturbance in the deposits containing the earliest artefacts at Madjedbebe, which provides support for a stable stratigraphy, which provides added confidence for the association of stone artefacts and the dates of 65 kya. This methodology can be used to reinforce conclusions about site dates in other sites with questionable stratigraphy.

The mechanistic target of rapamycin (mTOR) is a protein kinase that is closely linked to growth and nutrient control in a multitude of organisms. Inhibiting TOR with the drug rapamycin has been shown to increase lifespan in many species. In the fruit fly, Drosophila melanogaster, rapamycin slows development, an outcome that is perhaps closely related to its effect on lifespan. Recent work in the Promislow lab on larval development has shown that the effect of rapamycin varies greatly across different genotypes, from no impact in the time of development to a nearly doubling of development time. However, it has not yet been determined which of the three larval stages is most sensitive to rapamycin. My project attempts to answer this question. I tested the delay in larval development of larvae treated with rapamycin across six different fly genotypes, four that are known to be sensitive to rapamycin treatment, and two that are resistant. After transferring eggs to food containing rapamycin or control food, I collected larvae over three days and staged them based on specific characteristics of each stage. The data were compared between treatments and genotypes to see if there was a delay in specific larval stage development that resulted in the overall delay seen in previous experiments. These data were analyzed using R, and the results indicate that there is a significant delay in development of the first instar larvae of the sensitive strains, and no delay in the resistant strains. Based on these results, I will next use single cell sequencing of first instar larvae raised on rapamycin-treated or normal food, with the goal of better understanding the specific mechanisms by which rapamycin leads to a decrease in larval development time, and the genetic basis of variation in the response to this treatment.

Massive binary star systems influence the formation and evolution of galaxies through production of energy and heavy elements. These systems can evolve into X-ray producing high mass binaries: a neutron star or black hole accreting material from a high-mass companion. While high-energy mass accretion makes these sources especially important, it makes them exceedingly difficult to model. One extremely useful characteristic for testing models of their formation and evolution is their temporal variability. By observing the properties of these sources over time, we can directly compare observed variability against predictions from models. M33, a large spiral galaxy just under three million light-years away from the Earth, is home to a collection of these X-ray binaries, last catalogued in 2015. With the goal of constraining the variability of the binaries, I have analyzed five separate epochs of Chandra imaging data taken after the production of this catalogue. I compiled a preliminary list of observed X-ray sources in the images and used positions of the previously cataloged sources to correct the astrometry, ensuring consistent alignment among observations. I then selected the most accurate position estimate for each source by-eye. Combining data from all aligned observations, I have now extracted a catalogue of reliable source detections, as well as source properties such as position and flux. With these values, I have assessed source validity and created lightcurves to study their variability characteristics. Through this work, I have found a total of 56 bright sources that meet my criteria for validity, of which 49 sources were previously identified in literature and 7 sources are new, appearing only in our observations. Looking at the characteristics of both new sources and previously observed sources, I will constrain source variability to determine how much the X-ray binaries of M33 changed in brightness over the course of the observations.

In Southeast Asia, it is common for there to be a lack of clear typological categories in stone artefact archeology, but this is typically informally observed rather than empirically demonstrated. In Northern Vietnam, Mau A, an open air site located near the Red River in Yen Bai Province, Northern Vietnam, stone artefact flakes were excavated and analyzed in order to answer these questions: What variation exists within the assemblage of flakes, and are there distinct typological groupings corresponding to different tool categories? 1058 stone artefacts were excavated from four square meters using hand tools. The artefacts were measured using calipers, and these measurements were converted into outlines that we analysed using geometric morphometry and Principal Components Analysis (PCA). The PCA plot we made revealed that there was a great amount of overlap between reduction categories with a lack of typological groupings distinct from one another, although there was high variation in the amount of dorsal scarring. Primary flakes (the flakes that have been worked upon the least) have the least variability in terms of shape, with tertiary flakes having the most variability. We also created additional boxplots which revealed a lack of shape and material variation as well as confirming the PCA's results that dorsal scarring impacts flake variation. Our results suggest that flake shape is sensitive to assemblage reduction intensity, and may give useful comparative insights when other attributes show little variation. These results are important for understanding stone artefact assemblages from Southeast Asia which often yield little variation when analysed with traditional approaches.

Several age-related neurodegenerative diseases include abnormal protein deposition of tau, amyloid-β (Aβ), and TDP-43. The most common dementia, Alzheimer’s Disease (AD), is tauopathic, meaning it is mainly characterized by tau deposition. In neurons, tau normally functions to bind and stabilize microtubules (MTs), proteins made of tubulin dimers that provide neuronal structure and assist in molecular transport along axons. However, in AD, tau becomes hyperphosphorylated, resulting in disassociation from MTs and aggregation in the cell. Previous genetic screening identified several new mutations in genes encoding tubulin proteins that ameliorate the effects of tau toxicity in tauopathy models of Caenorhabditis elegans; however, the various genes confer suppression to varying levels. Given that tubulin genes are expressed differentially in neurons, I hypothesized that the level of gene expression may correlate with the level of tau toxicity suppression for a given gene. To test this, we constructed transgenic C. elegans strains overexpressing mutant tubulin at differing levels. I characterized age-matched worms for tau-induced motility defects and assessed tubulin transgene expression by qPCR, observing that higher levels of mutant tubulin gene expression correlated with higher levels of toxicity suppression. Additionally, since previous work showed that tubulin mutations could confer suppression in a tau-Aβ copathology model, I sought to elucidate whether this result is generalizable to a copathology model with tau and TDP-43. To test this, I generated strains of worms expressing human tau and TDP-43 as well as mutant tubulin, and assessed them for motility defects. I expect that suppression will be conferred in these animals, although not to the same extent as in a model without copathology. These experiments will help increase our understanding of the molecular mechanisms underlying mutant tubulin mediated tau suppression. Additionally, we will gain knowledge about the mechanisms behind tauopathic disease progression in models of pure tauopathy and copathology.

Recently, the Promislow lab found that levels of metabolites in the carnitine pathway can be used to estimate age in the fruit fly, Drosophila melanogaster. This ‘metabolite clock’ not only predicts an individual’s age, but also shows that when an individual’s predicted age is older than its chronological age, it has a higher mortality rate than other flies its age, and vice-versa. The carnitine pathway is required for energy production via fatty acid oxidation, for which carnitine also removes cellular waste products, and which may influence aging. I hypothesized that higher levels of carnitine would be associated with a longer lifespan, sustained by ongoing energy production and reduced cellular toxin accumulation. To test the effect of the carnitine pathway on fly aging, I measured the lifespan of flies while either supplying additional carnitine, or treating with the carnitine biosynthesis inhibitor etomoxir. I expect flies treated with supplemental carnitine to live longer than control flies, and that etomoxir-treated flies will live shorter than control flies. Approximately 125 female Drosophila melanogaster were assigned to food vials in each condition, plus a control condition that lacked added carnitine or etomoxir. I recorded deaths every two days, transferring remaining flies to fresh vials. Once all flies are dead, I use survival analysis to determine if either treatment affects lifespan, thus testing for a role of the carnitine pathway in fly mortality. Should the results support my hypothesis, I may explore the role that fatty acid oxidation has in aging, or to what degree the metabolome clock is affected by manipulation of the carnitine pathway.

Monitoring is a key component of any successful ecological restoration project. Being able to see how an ecosystem responds to restoration treatments is not only vital for planning out future restoration work, but also for more fully understanding how an ecosystem functions. The aim of this study was to examine the relationship between two culvert removals done in Carpenter Creek, a tidal creek feeding an estuarine wetland in North Kitsap, WA and the texture of the alluvial sediments bedded by the creek. The R programming language was used to analyze and visualize sediment texture data collected by Stillwaters Environmental Center, an environmental monitoring non-profit operating out of the local area. Generated data visualizations suggest that the first culvert removal, done near the mouth of the creek in 2012, resulted in a change in overall texture and an increase in fine sediments, while in 2014 and 2016 sediment texture drifted back towards the base state seen in 2011. In 2018, when the second culvert, this time located upstream near the marsh, was removed, a greater change in texture and increase in fine sediment was observed. These results suggest that an increased level of stream connectivity has been achieved, thereby allowing a freer and more natural sediment transportation regime. This increased understanding of Carpenter Creek’s evolution was made possible through a long-term monitoring effort. Understanding how estuarine wetlands respond to restoration treatments is key for successfully planning out future restoration work, as both adaptive management in existing projects and new restoration projects rely upon past experiences to inform management decisions.

Hepatitis C (HCV) is a liver disease caused by the bloodborne HCV virus. When left untreated, HCV can lead to cirrhosis and liver failure. Recent developments in therapeutics present a cure for HCV; however, treatment must be received soon after infection to be effective. Thus, limited availability of HCV testing creates a barrier to treatment distribution as chronic HCV is identified through a detectable viral load. Current HCV testing involves polymerase chain reaction (PCR) testing of blood samples, requiring a central laboratory and technicians to run them. The delay between appointments, sample transportation, running PCR, and receiving results can lead to lost contact with patients, making it difficult to connect them with timely treatment. The goal of the project is to develop a rapid point-of-care assay for HCV nucleic acid testing that allows healthcare providers to diagnose chronic HCV in 30 minutes and immediately prescribe treatments. We designed and validated an isothermal nucleic acid amplification assay for detecting HCV RNA: a two-step process involving reverse transcription of HCV RNA into complementary DNA (cDNA) which is detected by recombinase polymerase amplification (RPA). RPA is an isothermal process held at 40℃ with a runtime of 15 minutes, where a fluorometer collects data from the reaction. We compared the results of our RPA detection assay to the PCR-HCV assay used by the UW Clinical Virology Lab. We tested RNA from all six major genotypes using serum samples from Harborview Liver Clinic, where we had a limit-of-detection of 25 copies per reaction. We were able to match the results of the RPA and PCR assays with 100% agreement. By developing a streamlined detection assay for HCV, we will contribute to HCV testing without the need for expensive machinery or trained technicians, increasing the testing availability to increase HCV treatment rate and decrease HCV prevalence.

Towards Universal Newborn and Early Childhood Hearing Screening in Kenya (TUNE) is a project designed to introduce low-cost phone applications, akin to a standard tool, to screen for middle ear fluid in children. The device, which can be used by non-specialists, has the potential to provide low- and middle-income countries the opportunity to address problems in ear and hearing health that could otherwise impair childhood development. Kenya does not have a universal newborn and early childhood audiology screening program due to a lack of trained screeners and the high cost of assessment tools. The TUNE team met twice weekly via Zoom since colleagues lived in both Seattle and Nairobi, Kenya. My early responsibilities included documentation of stakeholder input. In September 2021, we sponsored a healthcare worker training program in Nairobi. In support of this work, I generated the training videos, created a website, and developed slide decks. These tools were used to provided healthcare workers with a background on hearing health and the importance of early identification of hearing loss, general information about the phone application, and trained them to use the device with a detailed instructional video. Healthcare workers were then given time to practice using the application. Finally, they were asked to complete a survey about the application. Most respondents reacted positively to the tool, with 100% of participants reporting they “would be comfortable using the device ‘as is’ in their work.” Most indicated that they would recommend using the application but noted that there were updates that could make the application more user friendly. We used the resulting qualitative data to inform future iterations of the application, a stakeholder workshop and policy recommendations. Ultimately this technology may help ensure that children’s growth, development, and educational needs are met in low- and middle-income countries.

Biological aging is the greatest risk factor for most major causes of mortality. Our lab operates under the hypothesis that slowing the rate of aging will also lower the risk of associated diseases. Unfortunately, biological aging research is often limited by the need for labor-intensive manual scoring of lifespan experiments. To solve this problem, we have created a pipeline using machine learning and robotics for the automated processing of lifespan experiments in Caenorhabditis elegans, a type of roundworm often used in biomedical research. Plates of C. elegans are first placed into a “WormBot” image capture robot, which takes images of the plates throughout the day. These photos are processed using YOLO, an object detection system that creates bounding boxes used to track general worm motion. I have implemented a semantic segmentation network that uses these bounding boxes to determine the general shape of the worm. This lets us gather morphological and behavioral data such as length, width, and position. This information is used to infer when the worm stops moving and can be called dead. The time of death this provides shows whether a treatment was successful in extending life. The morphological data can also be used to estimate whether or not the worm was healthier as it aged. This analysis lets us understand how effective various drugs, such as metformin, are at modulating the biological aging process and lessens the time required to run large numbers of trials. Our framework increases the rate at which experiments can be performed and also creates predictive models that provide suggestive data on the effectiveness of an intervention before the end of life. . By automating lifespan scoring, we accelerate the discovery rate of potential interventions that may eventually work in humans.

When trying to develop the combinatorics of the p-parts of a multiple Dirichlet series given a group of functional equations isomorphic to the Weyl group of a type A root system, one runs into two natural combinatorial defintions of these p-parts in terms of Gelfand-Tsetlin patterns. The fact that these two natural definitions are equivalent is proved in Brubaker, Bump, and Friedberg's Weyl Group Multiple Dirichlet Series: Type A Combinatorial Theory. Their proof is not bijective. Further developing the combinatorics of these generalizations of the Riemann zeta function and other Dirichlet series, and finding a bijective proof of the result of Brubaker-Bump-Friedberg, remains an active area of research. I worked with a large group of undergraduate mathematicians, mentored by Ben Brubaker himself as part of the online collaborative Polymath Jr. Program, and we introduced new combinatorial objects, namely a new kind of colored lattice model, with which we can explicitly conjecture the existence of such a weight-preserving bijection for result of Brubaker-Bump-Friedberg in the most general setting, with p-parts corresponding to certain metaplectic Whittaker functions.