BEGIN:VCALENDAR VERSION:2.0 PRODID:-//wp-events-plugin.com//7.2.3.1//EN TZID:America/Los_Angeles X-WR-TIMEZONE:America/Los_Angeles BEGIN:VEVENT UID:265@escience.washington.edu DTSTART;TZID=America/Los_Angeles:20250423T163000 DTEND;TZID=America/Los_Angeles:20250423T172000 DTSTAMP:20250418T181231Z URL:https://escience.washington.edu/events/uw-data-science-seminar-lauren- buckley-and-cole-martin/ SUMMARY:UW Data Science Seminar: Lauren Buckley and Cole Martin DESCRIPTION:Please join us for a UW Data Science Seminar featuring UW Biolo gy Professor Lauren Buckley and UW Chemical Engineering PhD Candidate Cole Martin on Wednesday\, April 23rd from 4:30 to 5:20 p.m. PT. This seminar will feature two different research projects supported by the eScience Da ta Science and AI Accelerator. \n\n\nThe seminar will be held in Electrica l and Computer Engineering Building 125 – Campus Map.\n"Automating asse ssment of butterfly thermoregulatory traits to uncover responses to climat e variability and change"\n\nAbstract: Heterogenous responses to climate change highlight the need to identify the underlying organismal mechanisms . Rapid progress in disseminating images of museum specimens affords exten sive opportunities to assess shifts in functional traits across space and time to uncover mechanisms. A single butterfly phenotype—wing coloration —shapes organismal responses to climate change at both acute and chronic timescales. Dark wings allow Pierid butterflies to absorb more sunlight t o heat up sufficiently for flight and associated fitness determining activ ities\, but dark wings can result in overheating and damage during thermal extremes. We present a workflow to download specimen images from reposito ries\, detect the bounding box of butterflies\, subset images to butterfli es with a standard resolution\, classify whether the image is of a dorsal or ventral surface\, and segment the butterfly wings. A geometric analysis then evaluates grayscale within the thermally relevant wing regions. We f ind that seasonal plasticity and spatial differences in coloration enable butterflies to balance flight capacity against risk of overheating. Shifts in wing coloration across decades depend on whether climate change poses an opportunity or a stress. Our project demonstrates the potential to leve rage emerging computational tools and museum resources to identify the org anismal mechanisms mediating climate change responses.\n Biography: Laur en Buckley is a professor in Biology at the University of Washington. Her research integrates modelling\, field and lab collection of ecological and physiological data\, and ecoinformatics to examine how biology (morpholog y\, physiology\, and life history) determines an organism’s ecological a nd evolutionary responses to climate change. A focus is characterizing ho w organisms experience and respond to fine scale spatial and temporal envi ronmental variation. Much of her recent work has entailed repeating functi onal experiments and observations on montane insects after several decades of climate change to assess ecological and evolutionary responses. The Tr EnCh project builds computational and visualization tools to Translate Env ironmental Change into organismal responses and improve capacity for ecolo gical and evolutionary forecasting.\n\n\n"Quantification of HIV-1 DNA Isot hermal Amplification Images Using Resnet-18 Networks"\n\nAbstract: Achiev ing the United Nations'\; goal of ending the HIV/AIDS epidemic by 2030 will require 30 million viral load tests annually—yet current diagnostic s fall short in speed\, scalability\, and accessibility. To address this n eed\, the Posner Research Group has developed a novel rapid HIV test based on isothermal nucleic acid quantification via time-resolved fluorescence microscopy of nucleation puncta. Previous DNA quantification using counts of discrete puncta has poor accuracy at high viral loads\, limiting the me thod’s dynamic range. This work leverages a ResNet-18 convolutional neur al network to overcome this barrier. By training a convolutional neural ne twork on both spatial and temporal imaging data\, we dramatically extend t he system’s dynamic range\, achieving 95% accuracy across clinically rel evant viral load classifications. This approach offers a promising path to ward fast\, low-cost\, and widely deployable quantitative diagnostics.\n\n Biography: Cole Martin is a Ph.D. candidate in the Posner Research Grou p in the Department of Chemical Engineering at the University of Washingto n. His current research focuses on developing a quantitative HIV viral loa d test for use in low-resource settings. Prior to this\, he helped create a rapid COVID-19 diagnostic that aimed to retain the sensitivity and speci ficity of PCR in the same readout format as lateral flow style antigen-bas ed tests. Before starting graduate school\, Cole worked at GlaxoSmithKline in vaccine adjuvant development. He holds bachelor’s degrees in chemica l engineering and biological engineering from Montana State University and a master’s in chemical engineering from the University of Washington.\n \n\nThe 2024-2025 seminars will be held in person\, and are free and open to the public. ATTACH;FMTTYPE=image/jpeg:https://escience.washington.edu/wp-content/uploa ds/2025/04/SeminarCombo.png END:VEVENT BEGIN:VTIMEZONE TZID:America/Los_Angeles X-LIC-LOCATION:America/Los_Angeles BEGIN:DAYLIGHT DTSTART:20250309T030000 TZOFFSETFROM:-0800 TZOFFSETTO:-0700 TZNAME:PDT END:DAYLIGHT END:VTIMEZONE END:VCALENDAR