Poster Session: 11:15-12:30 CDT

The WID Symposium Poster Sessions will be hosted on Congregate (click here for tips on using Congregate). This page will include a gallery of all posters so that you can browse them at your leisure before, during, and after the symposium.

1Stefan PietrzakDefining Reprogramming Checkpoints From Single-Cell Analysis of Induced Pluripotency
Elucidating the mechanism of reprogramming is confounded by heterogeneity due to the low efficiency and differential kinetics of obtaining induced pluripotent stem cells (iPSCs) from somatic cells. Therefore, we increased the efficiency with a combination of epigenetic and signaling molecules and profiled the transcriptomes of individual reprogramming cells.
Contrary to the established temporal order, somatic gene inactivation and upregulation of cell cycle, epithelial, and early pluripotency genes can be triggered independently such that any combination of these events can occur in single cells. Sustained co-expression of Epcam, Nanog, and Sox2 with other genes is required to progress towards iPSCs. Ehf, Phlda2, and translation initiation factor Eif4a1 play important functional roles in robust iPSC generation. Using regulatory network analysis, we identify a critical role for signaling inhibition by 2i in repressing somatic expression and synergy between the epigenetic modifiers ascorbic acid and a Dot1L inhibitor for pluripotency gene activation. ATAC-seq will also be implemented to examine the chromatin accessibility dynamics during reprogramming.
2Michelle CraftDiscoverIT, UW-Madison, and WID Provide Your IT Services
DiscoverIT, UW-Madison, and WID provide a variety of IT services including communication, desktop support, data, computing, servers, security, networking, and monitoring.
3Yue XieEfficient Algorithms for Bound Constrained Optimization
Bounds are the easiest type of inequality constraints. In this work we discuss efficient algorithms to solve problems with bound constraints, which includes the popular Nonnegative Matrix Factorization (NMF) in machine learning. We propose two practical variants of projected Newton methods and show that they are equipped with global worst-case complexity guarantees.
4Jiani ChenEpigenetic regulation through the control of a de novo DNA methyltransferase in plants
DNA methylation plays crucial roles in cellular development and stress responses through gene regulation and genome stability control. Precise regulation of DOMAINS REARRANGED METHYLTRANSFERASE 2 (DRM2), the de novo Arabidopsis DNA methyltransferase, is crucial to maintain DNA methylation homeostasis to ensure genome integrity. Compared to the extensive studies on DRM2 targeting mechanisms, little is known regarding the quality control of DRM2 itself.
Here, we identified an E3 ligase, COP9 INTERACTING F-BOX KELCH 1 (CFK1), as a novel DRM2-interacting partner that targets DRM2 for degradation via the ubiquitin-26S proteasome pathway. Loss-of-function CFK1 leads to increased DRM2 protein abundance and aberrant induction of CFK1 showed reduced DRM2 protein levels. Consistently, CFK1 overexpression induces genome-wide CHH hypomethylation and transcriptional de-repression at specific DRM2 target loci. Collectively, this study uncovered a distinct mechanism safeguarding de novo DNA methyltransferase by CFK1 to control DNA methylation homeostasis.
5Yuyuan WangpH-Responsive Polymer-Drug Conjugate: An Effective Strategy to Combat the Antimicrobial Resistance
Advances in nanotechnology promise new developments in multifunctional drug/nucleic acid/CRISPR gene-editing tool delivery systems. In particular, nanoparticles encapsulating therapeutic and diagnostic agents while providing specific molecular targeting capabilities are emerging as the next generation of multi-functional nanomedicines for targeted therapy and diagnosis, which will also pave the road for personalized medicine.
Nanomedicine is under intense investigation for the treatment and diagnosis of various major health threats including cancers, as well as cardiovascular, infectious, metabolic, ocular, and autoimmune diseases.

Nanomedicines can improve many of the pharmacological properties of free drugs:
• Enhance the solubility of hydrophobic drugs in aqueous solutions
• Increase the elimination half-time
• Offer both passive and active tissue/cell-targeting abilities
• Prevent the drug from premature in vivo degradation
• Provide controlled and sustained drug release profiles
• Offer multi-functionality
6Spencer HawsEpigenetic Adaptation to Methyl-Metabolite Depletion
S-adenosylmethionine (SAM) is the methyl-donor substrate for DNA and histone methyltransferases that regulate cellular epigenetic states. This metabolism-epigenome link enables the sensitization of chromatin methylation to altered SAM abundance. However, a chromatin-wide understanding of the adaptive/responsive mechanisms that allow cells to actively protect epigenetic information during life-experienced fluctuations in SAM availability are unknown.
We identified a robust response to SAM depletion that is highlighted by preferential cytoplasmic and nuclear de novo mono-methylation of H3 Lys 9 (H3K9) at the expense of global losses in histone di- and tri-methylation. Under SAM-depleted conditions, de novo H3K9 mono-methylation preserves heterochromatin stability and supports global epigenetic persistence upon metabolic recovery. This unique chromatin response was robust across the mouse lifespan and correlated with improved metabolic health, supporting a significant role for epigenetic adaptation to SAM depletion in vivo. Together, these studies provide the first evidence for active epigenetic adaptation and persistence to metabolic stress.
7Kushin MukherjeeFinding meaning in simple sketches: How do humans and deep networks compare?
Picasso famously showed that a single unbroken line, curved and angled just so, can depict a dog, penguin, or camel for the human viewer. What accounts for the ability to discern meaning in such abstract stimuli? Deep convolutional image classifiers suggest one possibility: perhaps the visual system, in learning to recognize real objects, acquires features sufficiently flexible to capture meaningful structure from much simpler figures.
Despite training only on color photographs of real objects, such models can recognize simple sketches at human levels of performance (Fan, Yamins, & Turk-Browne, 2018). We consider whether the internal representations arising in such a model can explain the perceptual similarities people discern in sketches of common items. Using a triadic comparison task, we crowdsourced similarity judgments for 128 sketches drawn from 4 categories—birds, cars, chairs, and dogs (Mukherjee, Hawkins, & Fan, 2019). On each trial, participants decided which of two sketches was most perceptually similar to a third. From thousands of judgments we computed low-dimensional nonmetric embeddings, then compared these human-derived embeddings to representational structures extracted for the same sketches from the deepest fully-connected layer of the VGG-19 image classifier. VGG-19 representations predicted human triadic comparison judgments with 59% accuracy–reliably better than chance, but still quite poor given chance performance of 50%. Embeddings derived from human judgments predicted held-out judgments with 75% accuracy. 2D embeddings derived from VGG-19 vs triadic-comparison differed starkly, with semantic category structure dominating the human-derived embedding and only weakly discernable in network representations. And yet network representations reliably captured some semantic elements: latent components predicted whether a given sketch depicted a living or non-living thing with 90% accuracy. Thus while the visual features extracted by VGG-19 discern some semantic structure in sketches, they provide only a limited account of the human ability to find meaning in abstract visual stimuli.
8Wallace LiuSmall Molecule Targeting of UHRF1 to Reverse Oncogenic Chromatin Binding
Can an oncogenic protein-protein interaction be selectively targeted by small molecules?
When cancer cells replicate, daughter cells inherit the existing program of gene expression states from parental cells, enabling oncogenes to remain highly expressed and tumor suppressor genes (TSGs) to be suppressed.
Faithful replication of these patterns is partly mediated by UHRF1, a protein that recognizes post-translational modifications (PTMs) on chromatin and recruits enzymes to propagate the same PTM patterns on newly replicated DNA. These PTMs subsequently promote cellular mechanisms that enforce gene expression. Thus, we reason that small molecule inhibition of UHRF1-chromatin binding constitutes a unique approach to reverse aberrant gene expression in cancer. To identify novel inhibitors, we conducted a high-throughput screening campaign and filtered the resulting hits through various in vitro validation assays. We find that eight compounds exhibit moderate potency for UHRF1-chromatin binding, the tightest series of inhibitors identified to date. These molecules will be followed up with structural efforts to discover the mode of binding, and cellular validation experiments to confirm UHRF1 displacement from chromatin and re-expression of TSGs.
9Rui PanAgDH: A Distributed System for Gathering and Disseminating Dairy Data
Dairy farms have been incorporating modern data-tracking services, which generate an enormous amount of data of myriad types (e.g. genetic, nutritional, reproductive). The organic nature by which the different types of automation systems have arisen and developed has resulted in a highly heterogeneous arrangement of different systems from different companies that often have difficulty integrating with one another.
Additionally, dependency on high-cost hardware systems (e.g. milking machines) makes it difficult for milk producers to switch service providers, which can disincentivize adaptation of the existing technologies. Modern agricultural analytics relies on the ability to integrate data from all of these data streams. To that end, we present the Agricultural Data Hub (AgDH), which revolutionizes dairy data collection and interpretation by providing uniformed data for future analyses through an extraction, transformation, and loading (ETL) process. This system establishes the relationships between these data, integrates those data, and makes them available from a single source, thus making it easier for dairy farmers to make management decisions. Later, these uniformed data will be sent to the Dairy Brain analytics services for additional analyses, and facilitate the visualization of the raw data and the analysis outcomes for easier consumption.
10Mingzhou YepH-Responsive Polymer-Drug Conjugate: An Effective Strategy to Combat the Antimicrobial Resistance
Antimicrobial resistant (AMR) infection is a growing threat to public health and force people to pursuit antibacterial drugs with enormous systemic toxicity. Here we report a dextran coated stimuli-responsive nanoparticle (NP) that encapsulates hydrophobic antibiotic, and specifically binds bacteria to overcome severe AMR infections. The NP shows strong affinity with a variety of pathogens in vitro and can effectively accumulate in the bacterial infected tissues.
It can be activated by either low pH or high reactive oxygen species (ROS) in the infectious microenvironment, and release both cationic polymer and rifampicin that show a strong synergy to combat AMR of the pathogens. The NP carrier also enables the antibiotic to effectively penetrate into the bacterial biofilm or enter the mammalian cells, thus allowing the elimination of biofilm induced AMR or intracellular infections. The NP shows great biosafety in vivo and exhibits remarkable therapeutic efficacy in two infection models induced by either Gram-negative or Gram-positive AMR pathogens, demonstrating its broad-spectrum antimicrobial capacity as well as its promising prospect in overcoming AMR infections.
11Bryce SprecherDiving in Data
The Virtual Environments Lab, in partnership with the National Park Service - Submerged Resources Center, capture and digitize shipwrecks in hard to reach places. Though a number of 3D scanning methods are available, many are not suitable for underwater operations. Utilizing the photography skills and systems already established at the NPS-SRC the Virtual Environments Lab reconstructs shipwrecks from the dive data using photogrammetry. These challenging shipwreck locations range from tropical waters of the Florida Keys to deep cold waters of Lake Superior.
Data is collected by NPS-SRC professional SCUBA divers using handheld DSLR cameras and a custom high resolution multi camera imaging platform, SeaArray. The SeaArray was built by the National Park Service and Marine Imaging Technologies, with consultation from the Virtual Environments Lab. Many challenges still remain: color and visibility are significantly reduced underwater, efficient capture is vital with limited dive times, and sometimes too much resolution and too many images can lead to suboptimal results. We continue to work together to address these problems to ensure that high quality results are available for site monitoring and management, archeological record, and public outreach.
12Zhen PengFrom chemical ecosystem to the origin of life
How life arose from abiotic world has been an unresolved fundamental question in natural sciences for a long time. Multiple theories were proposed to explain how life might emerge, but the lack of realistic chemical kinetics hinders the application of these theories to directing experimental studies as well as to analyzing specific life-like dynamics. By integrating chemical kinetics, ecology, evolutionary biology, and computer simulation, we developed a theoretical framework called chemical ecosystem ecology which not only depicts a promising route by which simple chemical rules can be organized to generate life-like systems but also can be used to search for candidate signals of life-like systems in experimental studies.
13Rohit KannanIntegrated Learning and Optimization
We consider optimization for data-driven decision-making in which parameters within the optimization model are uncertain, but predictions of these parameters can be made using available covariate information. We investigate a framework for integrating a machine learning prediction model within a sample-based approximation for approximating the solution to such problems. We derive conditions under which solutions to these approximations approach optimality when we use increasing amounts of data in the construction of the prediction model. Computational studies demonstrate the advantages of our data-driven formulations over existing approaches even when the prediction model is misspecified.
14Sarah MillerTiny Earth: Expanding a Diverse Network to Discover Antibiotics while Pivoting to Online
Tiny Earth is a network of students and instructors focused on crowdsourcing antibiotic discovery from soil bacteria. In 2020-21, our priorities are to expand the global network, increase diversity and antiracism in the curriculum, and pivot the research-based Tiny Earth course to online. In this poster, we provide an overview of the network's diversity and reach as well as tools we are using to pivot to online with equity and accessibility in mind.
15Kaivalya MoluguTracking and Predicting Reprogramming using Nuclear Characteristics
Reprogramming of human somatic cells to induced pluripotent stem cells (iPSCs) generates valuable resources for disease modeling, cell therapy, and regenerative medicine. However, the reprogramming process can be stochastic and inefficient, creating many partially-reprogrammed intermediates and non-reprogrammed cells in addition to fully-reprogrammed iPSCs. Much of the work to identify, evaluate, and enrich for iPSCs during reprogramming relies on methods that fix, destroy, or singularize cell cultures, thereby disrupting each cell’s microenvironment.
To overcome these shortcomings, we developed a micropatterned substrate that allows for nondestructive dynamic live-cell microscopy of hundreds of microscale cell subpopulations undergoing reprogramming while also preserving many of the biophysical and biochemical cues within the cells’ microenvironment. On this substrate, we were able to both watch and physically confine cells into discrete micron-sized islands during the reprogramming of human somatic cells from skin biopsies and blood draws obtained from healthy donors. Using high-content analysis, we identified a combination of eight nuclear morphometric characteristics that can be used to track the progression of reprogramming and distinguish partially-reprogrammed cells from those that are fully-reprogrammed. Non-cell autonomous characteristics, such as clustering of nuclei, were highly informative in classifying the progression of reprogramming, and were used to generate a predictive computational model of the process. This quantitative approach to track reprogramming in situ using micropatterned substrates could aid in biomanufacturing of therapeutically-relevant iPSCs, and be used to elucidate multiscale cellular that accompany human cell fate transitions.
16Ardhendu TripathyGeneralized Chernoff Sampling: A New Perspective on Structured Bandits
Structured stochastic bandit problems are intimately related to the classical problem of sequential experimental design. This paper studies new algorithms for best-arm identification in structured stochastic bandits settings inspired by an experimental design method proposed by Chernoff in 1959.
The contributions of the paper are: 1) a new algorithm for structured bandits; 2) novel sample complexity bounds for the new algorithm and Chernoff’s original design method; 3) experimental results demonstrating that the new algorithm is computationally light-weight and its performance is competitive with state-of-the-art methods for linear bandit problems.
17Alex PlumSpatial Structure in Autocatalytic Chemical Ecosystems at the Origins of Life
We consider life as a general process, distinguished by an ability to self-propagate and a capacity for open-ended evolution. How exactly self-propagation and adaptive evolvability can emerge in out-of-equilibrium chemical processes remains an open question. Autocatalytic cycles, whose constituent chemicals collectively catalyze their own continuous recreation, seem to have the potential to manifest both such properties and may play a substantial role in abiogenesis.
Mathematical models of the dynamics of chemical reaction networks situated in well-mixed reactors, continuously diluted, and driven out of equilibrium by a constant flux of food chemicals demonstrate that distinct autocatalytic processes within those chemical reaction networks can act analogous to distinct species in biological ecosystems. Further, larger networks can exhibit long term dynamics that resemble succession and evolution. At the level of autocatalytic pre-life, “individuals” contain no spatial structure, or perhaps because there is no spatial structure there are no individuals. Nevertheless, spatial structure imposed by the environment might scaffold the emergence of individuality as well as permit a richer variety of selective pressures and more complex ecological dynamics. Past modeling has relied on mass-action kinetics so that concentrations are continuous, events are deterministic, and all spatial structure is abstracted away. Here, we introduce new models to stochastically simulate artificial, out-of-equilibrium chemical ecosystems in well-mixed flow reactors, reaction diffusion systems, and nested compartments. We find that spatial structure can reshape chemical ecological dynamics and that certain regimes of interaction with that spatial structure lead to higher levels of ecological complexity and autocatalytic process stability. This computational modeling approach may provide critical insights into the emergence of evolvable chemical systems prior to the emergence of genetics.
18Michael O’NeillPerformance of Nonconvex Optimization Methods with Worst-Case Guarantees
We investigate the practical efficiency of nonconvex optimization methods with worst-case guarantees when applied to machine learning problems. We compare algorithms with theoretical guarantees against those known to be effective in practice. These algorithms are applied to a nonconvex regression problem as well as low-rank matrix completion.
19Clare MichaudData Science Hub: Here to Help With Your Data Science Needs
The Data Science Hub creates connections among UW-Madison researchers, students, faculty, and staff, and with professionals in the community, around topics related to data science, to help researchers across domains and industries enhance their research with data science techniques. We do this through consultations, regular trainings, events, and the communications channels of our newsletter and Twitter account.
20Alexis LawtonRevealing dynamic protein acetylation across subcellular compartments
Acetylation of the ε-amino of lysine residues is a widespread, reversible post-translational modification that regulates many cellular functions, including protein-protein interactions, protein-DNA interactions, cellular localization, protein stability, and enzymatic activity. Historically, the first well-characterized example of acetylation was found on histone tails, but more recent proteomics analyses have identified thousands of acetylation sites on non-histone proteins.
To address which acetylation sites are functionally relevant, our lab has developed and validated an improved, complementary proteomics method to calculate global and site-specific acetylation stoichiometry. Using the acetylation stoichiometry analyses, the goal of this study is to understand acetylation dynamics in serum stimulated cells and to define the cellular mechanisms involved in regulating the unique responses and rapid changes in protein acetylation.
21Sailendharan SudakaranMicrobiome Hub
Microbiome hub is a joint venture between the UW Biotechnology Center and the Wisconsin Institute for Discovery, making use of current prowess at UWBC in sequencing and bioinformatics and WID’s strength in data storage and management. Microbiome research is one of the rapidly growing areas of science, encompassing a broad range of systems from human body, animals, soil, plants, lakes, food, wastewater treatment systems, and test tubes.
Continuous advancements in sequencing technologies have enabled researchers in utilizing various omics methodology to gain a better insight in to the complex relationships between organisms, genes and their environment and in turn this has led to the generation of vast amounts of data. Therefore, there is a need to ensure that adequate expertise and infrastructure is in place to meet the challenge of storing, analyzing and distributing the data generated as well as a robust and reliable framework for interpreting the data. Microbiome hub will serve as a campus wide resource to address these challenges at UW Madison. It will support researchers interested in tackling a broad spectrum of microbiome related studies from investigating fundamental questions, developing new techniques and methodologies to understanding the impact of microbiome across different fields. The hub will provide a broad spectrum of services such as technical expertise (data analysis), consultation, coordinating collaborative research, data management and conducting outreach activities. Overall, by harnessing the significant research already taking place at UW, world class expertise and facilities we aim to develop the Microbiome Hub as prominent center for microbiome research at UW.
22Sunnie Grace McCallaExamining cell-cell interactions to define population structure from single cell RNA-seq datasets
The advent of single-cell RNA sequencing (scRNA-seq) has permitted the profiling of mRNA levels from nearly all genes in individual cells in a heterogeneous sample. The availability of these mRNA profiles from individual cells presents new opportunities to characterize cellular heterogeneity on a scale that has been previously impossible, including the ability to define new cell subpopulations and their lineage structure. A key step to cell subpopulation identification entails defining a graph of interactions among cells, which is then used for grouping cells based on the similarity of their interactions.
Existing approaches have used pairwise metrics, such as a k-nearest neighbor graph, however, it is unclear how much of the subsequent subpopulation discovery is dependent upon the input graph structure. This may impact the accuracy of defining cell subpopulations and affect the biological inferences drawn from the results. Here we study the impact of the input graph on the identification of cell types using a mouse reprogramming scRNA-seq dataset and examine new directions to define cell subpopulations using methods that incorporate causal network learning approaches to define relationships among individual cells. Our early analysis suggests that more principled approaches to defining cell-cell relationships can improve the recovery population level relationships and can benefit the inference of cell population identity.