Objectives/Goals: We assessed the feasibility of using a large language model (LLM) to create lay language descriptions of study protocols for recruitment, which has the potential to improve accessibility and transparency of clinical studies and enable participants to make informed decisions. Methods/Study Population: All studies from a clinical research recruitment platform were included, which features human-written lay descriptions and titles for study recruitment. Corresponding protocol summaries in the IRB system were extracted and translated into lay language using a LLM (gpt-35-turbo-0613). A subset was used to develop prompt variations through an iterative process. Prompt strategies evaluated include chain-of-thought and few-shot prompting techniques. LLM-generated and human-written descriptions were compared for readability using Flesch–Kincaid and Simple Measure of Gobbledygook (SMOG) reading grade levels and information completeness using Word Movers’ Distance (WMD). Results/Anticipated Results: A total of 55 study descriptions were included – 10 were used to develop prompts and 45 were used for evaluation. The final LLM instructions included multistep prompts. The LLM was first instructed to produce a two- to three-sentence long description without using scientific jargon and included two pairs of examples. The LLM was then asked to shorten the description and finally to provide an engaging title. LLM-generated and human-written summaries were similar in length (median (IQR) 328 (278.5–360.5) vs. 342 (203–532.5) characters, respectively). LLM-generated summaries had lower Flesch-Kincaid grade level (5.15 vs. 8.28, p Discussion/Significance of Impact: An LLM can be used to generate lay language summaries that are readable at a lower grade level while maintaining semantic similarity. This approach can be used to improve the drafting of summaries for recruitment, thereby improving accessibility to potential participants. Future work includes human evaluation and implementation into practice.

Objectives/Goals: Both diabetes and obesity are major health issues in the USA. Though much focus is given on the impact of lifestyle modifications to control diabetes and obesity, more information needs to be established about the association of social determinants of health and them. This study explores these associations, focusing on the DMV area. Methods/Study Population: The data for this study were collected for 158 counties which cover two states Maryland and Virginia, and Washington DC from PolicyMap. PolicyMap is a geography information system (GIS) that aggregates different types of data from different sources for research purposes. County-level data on public health related to different SDoH like median age, and percentages proportions of different ethnicity (Hispanic or Latino), percentages of different race, and nativity (foreign born), gender (ratio of male to female), access to primary healthcare, social vulnerability index (SVI), and median household income was used in this study. Statistical methods like multiple regression, one way ANOVA, and Pearson’s correlation coefficients were used to determine which factors are associated with these two conditions. Results/Anticipated Results: For both diabetes prevalence and obesity, multiple linear regression model with backward elimination was used to select variables which associated with them. The backward elimination process selected the set of factors for which the adjusted R square was the highest. In both cases, median household income, median age of population, social vulnerability level, percentage of white population, and percentage of foreign-born population were found to be significant at 5% level of significant. Pearson’s correlation coefficients showed significant positive relationship between factors like obesity and diabetes, median age and access to health care, and negative relationship between obesity and foreign born. Income, healthcare access, and white population were found to be significantly different SVIs from ANOVA. Discussion/Significance of Impact: This research study found that some SDoH affect diabetes and obesity in the same direction. The association is positive for median age and negative for income, SVI, percentage of white population, and foreign born. The associations were found between actionable and non-actionable factors like percentage of white population with access to health care.

Objectives/Goals: This Weill Cornell Clinical and Translational Science Collaborative (CTSC) project evaluates whether large language models (LLMs) can generate accurate summaries of translational science benefits using the Translational Science Benefits Model (TSBM) framework, aiming to identify optimal LLMs and prompting strategies via expert review. Methods/Study Population: We are using prompt engineering to train multiple LLMs to generate one-page impact profiles based on the TSBM framework. LLMs will be selected via benchmarks, focusing on models excelling in information extraction. Leading LLMs (e.g., Llama 3.2, ChatGPT 4.0, Gemini 1.5 Pro, and Claude) and other high-performing models will be considered. Initial work has utilized Gemini 1.5 Pro. Models use data from CTSC-supported projects in WebCAMP, our local instantiation of a translational research activity tracking system used by >20 CTSA hubs, and manuscripts from the Overton database cited in policy documents. Human experts will evaluate the quality and accuracy of LLM-generated profiles. Results/Anticipated Results: Preliminary results using Gemini 1.5 Pro indicate that LLMs can generate coherent and informative impact profiles encompassing diverse areas within the TSBM. Face validity appears satisfactory, suggesting the outputs align with expectations. We anticipate that further exploration with other LLMs and expert validation will reveal strengths and weaknesses of the LLM approach, including the potential for naccuracies (“hallucinations”), informing further refinement of models and prompting strategies. Analysis of manuscripts cited in policy will provide valuable insights into communicating policy-relevant benefits effectively, and benchmark comparisons will identify optimal LLMs for this use case. Discussion/Significance of Impact: This project demonstrates LLMs’ potential for streamlining and enhancing impact reporting in translational science, enabling broader dissemination of research outcomes and promoting better understanding among stakeholders. Future work will integrate LLM-based reporting into research infrastructure.

Objectives/Goals: Our goal was to determine whether a consensus exists around 1) what the main barriers to innovation in Health artificial intelligence (AI) are 2) where there are gaps in education and training in Health AI and 3) where in their workflows organizations should implement AI to see the most immediate impact on productivity. Methods/Study Population: We employed a three-round Delphi method survey to stakeholders with health and/or engineering expertise. The first round was open-ended to generate responses to the three research questions. The second round asked participants to rank the responses and provide feedback as to their reasoning. The third round provided aggregated results and feedback and asked participants to re-rank the responses. Participants were attendees at a conference that brought people with health and/or engineering backgrounds together to discuss innovation in Health AI. 55 people in total participated across the three rounds. Results/Anticipated Results: Consensus emerged on all three questions: lack of trust was seen as the single greatest barrier to innovation, experience with implementation as the greatest gap in training, and automating health documentation as the point of most immediate impact. Consensus also emerged as to which of the 10–15 responses to each question were top priorities, which were somewhat significant, and which were not that important. Some of the rankings (such as implementation) seemed to reflect hot topics of discussion at the conference, but others (such as documentation) only emerged as significant in the surveys. Discussion/Significance of Impact: We successfully employed the Delphi method to discover what stakeholders think about three important questions in Health AI. Interestingly, although we polled experts from both health and engineering backgrounds, their answers converged on all three questions.

Objectives/Goals: To investigate interventional clinical trial participation overall and by race, gender, and age. Methods/Study Population: We used Epic Cosmos, an aggregated, de-identified EHR platform including over 270 million patients, to examine overall clinical trial participation and the race, gender, and age composition of participants versus non-participants. Patients ≥5 years old with known race and gender and at least one healthcare encounter between 2021 and 2024 were included. Interventional trial enrollment was identified by a “research flag” indicating current or past participation in an interventional study within an Epic system contributing data to Cosmos. Race was categorized as American Indian, Asian, Black, Native Hawaiian, or White. Age-adjusted relative representation (RR) ratios were used to compare participation, with RR >1 indicating over-representation and RR Results/Anticipated Results: Of 130,455,189 patients meeting eligibility criteria, 0.52% (673,425) of patients were active or inactive in an interventional clinical trial. Results are shown in the figure below. The poorest representation was from Asian and NH/PI persons. Representation was most similar to the patient population for whites and AI/AN persons. Black males participated less and women, more than predicted by patient composition. Older patients participated more frequently than younger (age, mean (SD), y, 53 (22) vs. 46 (23); p Discussion/Significance of Impact: This is the first study we know of describing interventional trial participation in the USA across millions and millions of patients. Further research is needed to clarify whether these differences are due to the nature of the studies themselves (e.g., OB/GYN trials including only women, etc.) versus disparities in recruitment or otherwise.

Objectives/Goals: Ventilator-associated pneumonia (VAP) is an infection caused by bacteria, viruses, or fungi during mechanical ventilation. We analyzed a cohort of COVID-19 patients admitted to the intensive care unit with respiratory failure with different VAP outcomes. We hypothesize that the multiomics data can help predict VAP development within this cohort. Methods/Study Population: We recruited participants from a cohort on a NYU IRB protocol (i22–00616), who had COVID19 respiratory failure, admitted to ICU, and required invasive mechanical ventilation (n = 245). We collected and analyzed research specimens (bronchoalveolar lavage [BAL, n = 213], tracheal aspirates [n = 246], background [n = 18]) and clinical cultures (sputum and BAL) for 245 participants. A panel of experts adjudicated VAP within the cohort, resulting in 92 VAP diagnoses. We annotated metatranscriptome (Illumina NovaSeq) using a Kraken/Bracken database, and KEGG for functional annotation of transcriptome data (Illumina HiSeq). We used edgeR (v.4.0.16) to analyze differential expression of metatranscriptome and transcriptome data. Results/Anticipated Results: We diagnosed VAP in n = 92 (38%) participants. We found significant differences in days of overall hospital stay (p Discussion/Significance of Impact: VAP is a serious complication of mechanical ventilation, and oral commensals alter the lung microbiome and host immunity. We identified a transcriptome-metatranscriptome signature that identifies those at VAP risk. VAP was associate with both pro- and anti-inflammatory gene expression resulting in increased risk for lower airway infection.

Objectives/Goals: We aim to develop an intuitive interface to understand the possible relationships between data from different RNA-Seq technologies. It can help novice users and educators to understand, analyze, explain, and visualize such datasets from diverse platforms, all without the need for additional software installations or strong programming expertise. Methods/Study Population: An online interactive interface is developed, integrating robust algorithms for three distinct types of analyses: DESeq2 for bulk RNA-seq, CIBERSORTx for deconvolution, and Seurat for single-cell analysis, with plans to include more algorithms. It allows a demo mode for training using the sample datasets and option for tailored analysis using user’s partially processed data. The interface provides capability to process bulk RNA-seq data from raw counts or a differential gene list. Further, deconvolution analysis for bulk RNA-seq data can be done using raw counts and single-cell data analysis can be performed using processed sequence reads, organized into three key files: barcodes, matrix, and features. Users also have an option to download the results as a zipped file, for samples from human and mouse studies. Results/Anticipated Results: Users with an active internet connection can access the interface from any major web browser. They can adjust parameters – such as genome type, cutoff thresholds, and batch effect correction – according to their specific needs. Bulk RNA-seq results are presented in the form of volcano plots, heat-maps, clusters, gene expression across samples, DEGs, and enrichment plots from KEGG and GO analyses. Deconvolution analysis can be performed using either the “LM22” signature matrix (for human leukocyte cell types) or Derm22 (for skin-specific cell types). The single-cell workflow provides results including quality control metrics, UMAP clustering, gene expression plots/tables, and cluster annotation using CellTypist. Comprehensive details on methods and tutorials are available in the GitHub repository. Discussion/Significance of Impact: Although multiple workflows are available to process bulk and single cell RNA-Seq data along with deconvolution methods to bridge the gap between the two, this is the first online interface to provide the capability to explore and analyze data from all three approaches in one place, without requiring strong computational expertise or resources.

Objectives/Goals: This research aims to harness electronic health records (EHR) combined with machine learning (ML) to predict necrotizing enterocolitis (NEC) in preterm infants using data up to their first 14 days of life. We aim to provide interpretable results for clinical decisions that can reduce infant mortality rates and complications from NEC. Methods/Study Population: Through a retrospective cohort study using data from the University of Florida Integrated Data Repository and One Florida, we will develop machine learning models suitable for sequential data to predict NEC. Our inclusion criteria include very low birth weight (VLBW; <1500g) infants born <32 weeks gestation and EHR data availability from the first 14 days of life. We will include infants with NEC and infants without NEC to train our ML model. Exclusion criteria include infants diagnosed with spontaneous intestinal perforation and severe congenital anomalies/defects requiring surgery. Results/Anticipated Results: We anticipate that our model will provide an accurate and dynamic prediction for the risk of NEC in neonates using data up to the first 14 days of life. Our model will be interpretable to identify key risk factors and can integrate real-world clinical insights to increase early detection and improve patient outcomes. Discussion/Significance of Impact: The development of a model to predict NEC could be used in neonatal intensive care guidelines and protocols and could ultimately decrease mortality, reduce complications, improve the overall quality of care, and lower healthcare costs associated with NEC.

Objectives/Goals: Healthcare organizations must track electronic health record (EHR) activity at the user level, including logons, accessed records, and viewed or entered documentation. There is little standardization in EHR audit logs and nurse workload has not been explored using these data. In this project, we characterized nurse actions from EHR audit logs. Methods/Study Population: We performed an analysis of EHR audit log data collected from 8,149 nurses over 5 years at University of Utah Health. We preprocessed nursing-centric EHR audit logs from the Epic EHR by cleaning and preparing the data for analysis. We calculated basic statistics for the variables labeled user_id (nurse) and metric_id (action). We reviewed the actions used by nurses and categories the actions as navigation, view, and entry. To capture the clinical context of the actions, two nurses categorized each action. A third nurse resolved any discrepancies. Results/Anticipated Results: We found that of the 4,419 available metrics, nurses used 1,461 unique metrics during the timeframe. The actions most used by nurses were 1) report with patient data viewed, 2) inpatient system list, and 3) storyboard viewed. Most of the metrics were categorized as navigation. The number of nurses interacting with the EHR increased each year and on average, we found that 1000 unique metrics were used by each nurse user in a 24-hour period. The expected outcome is a set of actions that can be mapped to higher level nursing interventions and in the future contribute to models for nursing workload measurement. Discussion/Significance of Impact: We found great value in using EHR audit logs to provide insights into nursing actions. Information gleaned can benefit organizations that are crafting interventions to decrease workload. Ultimately, the goal is to ensure that nurses have an appropriate workload allowing for safe and high-quality patient care while maintaining their well-being.

Objectives/Goals: Despite the documented link between ultraviolet (UV) exposure and skin cancer, the potential protective role of pigmentation in preventing skin carcinogenesis is not fully understood. This project will test the hypothesis that individuals with light skin tones are more susceptible to skin cancer due to a higher accumulation of somatic mutations from UV exposure. Methods/Study Population: Skin biopsies were collected from various anatomical sites of seven donors aged 70 to 80 years with dark skin tones and used to generate clones of normal skin cells. The clones were then subjected to whole-exome and RNA sequencing. We developed a workflow that allows us to accurately detect somatic mutations in clonal expansions of individual cells with high specificity and sensitivity. For comparison, we also analyzed mutations in skin cells obtained from individuals with light skin tones, who served as our control group. In the future, we plan to employ admixture analysis to clarify the impact of race on somatic mutations by examining ancestral genetic contributions. Results/Anticipated Results: Our preliminary analysis of the existing data supports our hypothesis. The average mutation burden in individuals with light skin tone was found to be 3.778 Mut/Mb, whereas in those with dark skin tones, was only 0.818 Mut/Mb. These findings align with the prevailing hypothesis that melanin acts as a protective factor against skin cancer, as it appears to reduce the mutagenic effects of UV radiation. This suggests that individuals with darker skin may have a lower cumulative mutational load, potentially contributing to their reduced risk of developing skin cancers such as melanoma. Similarly, our results showed variations in mutational burdens across different anatomical sites, which seemed to be influenced by differing levels of UV exposure, with the highest burdens observed in areas with the greatest exposure. Discussion/Significance of Impact: Understanding how skin cancer operates in darker-skinned individuals is imperative for tailoring effective screening and targeted therapies to meet their specific needs. In the long run, elucidating the mechanisms underlying skin cancer development in this demographic will help refine screening protocols and prevention recommendations.

Objectives/Goals: NIH requires researchers submit Data Management and Sharing (DMS) Plans with their grant applications. Librarians developed an assessment tool for the plans and completed a pilot assessment in order to leverage the plans and understand current institutional research data management and sharing. Methods/Study Population: The assessment tool includes questions related to evaluations of DMS Plans as well as questions related to the content of the plans. Evaluation questions were adapted from the Federation of American Societies for Experimental Biology evaluation rubric developed for the DataWorks! Data Management Plan (DMP) Challenge. Fields were added to collect information on the content of DMS Plans, including data type, institutional resources, data repositories, data standards, and data dissemination timelines. The assessment tool was tested in a pilot implementation. Seven library workers were trained and completed paired review samples of 27 DMS Plans (54 evaluations total) in order to test for tool reliability. Results/Anticipated Results: Results include findings on the reliability of the tool as well as preliminary results from an assessment of DMS Plans. Findings on the reliability of the tool include assessments of the paired reviewers for each question included in the tool. Paired reviewers generally agreed, but tended to differ on specific questions, including questions pertaining to the data types generated or used in a research project. Questions with high levels of agreement included subjects of study and code sharing practices. Results on the content of the DMS Plans include information such as data repositories used, data oversight responsibilities, and data and metadata standards employed. Discussion/Significance of Impact: DMS Plans present an opportunity to better understand data management and sharing practices, and good data management supports high-quality, reproducible research. Developing and testing assessment tools for these plans is a key step toward understanding and improving current research data management practices.

Objectives/Goals: Although several studies have identified significant associations between specific social determinants of health (SDoH) and adverse outcomes, little is known about how SDoH co-occur to form subtypes and their outcome-based risks. Here we analyze how SDoH co-occur across all participants with a cancer diagnosis in the All of Us program. Methods/Study Population: Data: All participants (n = 3361) with cancer and their responses to 110 survey questions related to SDoH. Independent variables: 18 SDoH factors aggregated from the questions to address uneven granularity. Dependent variables: depression, delayed medical care, and ER visits in the last year. Analytical Method. (1) Bipartite network analysis with modularity maximization to identify participant-SDoH biclusters, measure the degree of their biclusteredness (Q), and estimate the significance of Q. (2) Visualization of the results using the ExplodeLayout force-directed algorithm. (3) Multivariable logistic regression (adjusted for demographics and corrected through FDR) to measure the odds ratio (OR) of each bicluster compared pairwise with the other biclusters to estimate their risk for the 3 outcomes. Results/Anticipated Results: As shown in Fig. 1A (http://www.skbhavnani.com/DIVA/Images/Cancer-SDoH.jpg), the analysis (n = 3361, d = 18) identified 4 biclusters with significant biclusteredness (Q = 0.13, random-Q = 0.11, z = 9.94, P Discussion/Significance of Impact: Currently, many health equity policies allocate resources based on sociodemographic factors like race and income to address disparities. The 4 distinct subtypes and their outcome-based risks suggest that such policies could be more precise if they were based directly on combinations of need using SDoH subtypes and their risk stratification.

Objectives/Goals: Dravet syndrome is a developmental and epileptic encephalopathy associated with refractory seizures and a high risk of sudden unexpected death in epilepsy. A pathogenic biallelic variant in SCN1B, SCN1B-p.R98C, was identified in three patients with Dravet syndrome. Here we investigate SCN1B-p.R98C on neuronal function in vivo. Methods/Study Population: Scn1b-p.R98C mice were previously generated using CRISPR-Cas9 gene editing. Homozygous animals exhibit increased susceptibility to hyperthermia induced seizures at postnatal day (P) 15, 100% expression of spontaneous generalized seizures by P30, and ~20% undergo SUDEP by approximately P60. Here we examined the neuronal phenotype of P17–28 male and female Scn1b-p.R89C mice. We used whole-cell patch clamp electrophysiology approaches to measure effects of the variant on passive membrane properties, intrinsic excitability, and single action potential properties of parvalbumin positive (PV+) interneurons and pyramidal neurons in layers 5/6 of the somatosensory cortex and CA1 region of the hippocampus. Wild-type littermates were used as controls. Results/Anticipated Results: Our results show no differences between genotypes in any measure for somatosensory cortical PV+ interneurons or pyramidal neurons. In the CA1 region of the hippocampus, we found no differences for any measure in PV+ interneurons. In contrast, CA1 pyramidal neurons were hyperexcitable, however, with no changes in passive membrane properties or single action potential properties. Discussion/Significance of Impact:

Objectives/Goals: Methicillin-resistant Staphylococcus aureus (MRSA) is a human bacterial pathogen and is classified as a serious threat. MRSA has become resistant to most B-lactam antibiotics (penicillins and cephalosporins). The goal of this study is to identify an antibiotic adjuvant capable of resurrecting B-lactams for the treatment of MRSA infections. Methods/Study Population: A fluorescence-reporter assay was used to screen a compound library. Minimum-inhibitory concentrations were assessed for the compounds against various MSSA and MRSA strains. A common resistance mechanism to B-lactams by MRSA is by the function of the bla operon. One gene in this operon encodes for a B-lactam sensor/signal transducer protein BlaR, the primary target of this study. Inhibition of BlaR by compound 1 (best potentiator of oxacillin) was studied by nano-differential scanning fluorimetry (nanoDSF), surface plasmon resonance (SPR), scanning electron microscopy (SEM), and time-kill assays. Results/Anticipated Results: We identified 80 compound hits from a 1,974-compound NCI library. Twenty-four compounds showed potentiating ability (2- to 4,096-fold decrease in MIC for oxacillin). Seven compounds exhibited melting temperature shifts by nanoDSF of BlaR, indicating binding. SPR determined compound 1 has a binding affinity of 31 micromolar to BlaR-SD. SEM images showed disruption in the S. aureus cell wall on exposure to compound 1 and oxacillin. S. aureus N315 showed 3-log reduction in bacterial count treated with a mixture of compound 1 and oxacillin. Discussion/Significance of Impact: Compound 1 targets BlaR-SD, which restores S. aureus susceptibility to treatment by oxacillin. There are currently few antibiotics available in the clinic capable of treating MRSA infections. The combination hold promise of a treatment option for MRSA.

Objectives/Goals: We hypothesized that adding cytomegalovirus (CMV) viral Fc gamma receptors (vFcγRs) to a glycoprotein B (gB) protein subunit vaccine would improve vaccine-elicited Fc mediated effector functions such as antibody dependent cellular phagocytosis (ADCP) and cytotoxicity (ADCC), over gB subunit alone. Methods/Study Population: We immunized rabbits (n = 4 per group) at Weeks 0, 4, and 8 with 20μg gB alone or with one vFcgR (gp34, gp68, or gp95) at 20μg or 40μg, adjuvanted with squalene emusion, Addavax. Plasma from immunized rabbits was analyzed for antigen-specific IgG binding via enzyme-linked immunosorbent assays (ELISAs). ADCP was measured by conjugating whole virions to a fluorescent marker (AF647), incubating the fluorescent virus with rabbit plasma, and measuring uptake of virus by THP-1 monocytes via flow cytometry. ADCC was measured by natural killer cell degranulation via flow cytometric detection of CD107a expression following co-incubation with CMV-infected fibroblasts and rabbit plasma. Results/Anticipated Results: Each vFcγR demonstrated immunogenicity, although average vFcγR-binding IgG titers were between 4- to 10-fold higher in animals receiving the 40μg dose of each vFcγR compared to the 20μg dose. We observed similar IgG binding responses against gB among all vaccine groups. Comparing groups at peak immunogenicity (Week 10), ADCP responses were improved over gB alone by approximately twofold in animals receiving 40ug of each vFcγR. This effect was maintained across several human CMV strains with variable vFcγR genes. ADCC responses were undetectable in all animals immunized with gB alone, yet those receiving 40μg gp34 or gp95 demonstrated detectable ADCC. Discussion/Significance of Impact: HCMV-specific ADCP and ADCC are associated with protection against vertical CMV transmission, so a vaccine including vFcγRs which can improve vaccine-elicited Fc-effector responses is promising toward reducing the immense global impact of congenital CMV and associated neurologic birth defects.

Objectives/Goals: Research supports the use of music to improve the care and well-being of adults living with dementia; however, the practice and implementation of music in elder care communities is not regulated. The goal of this qualitative study was to survey elder care communities in Northeast Kansas to determine the use of music with people living with dementia. Methods/Study Population: We interviewed staff (n = 10) at five elder care communities in the Kansas City Metro area and observed musical activities and artifacts in shared living spaces within each community. Interview questions included details of the frequency and purpose of using music, who determined which music to use, and any effects, positive or negative, the interviewee believed to be associated with the use of music. Musical events, visiting musicians or music therapists leading group sing-alongs were observed at two communities, and music-related activities led by staff were observed at two others. Results/Anticipated Results: Music was used in some way at each of the five communities. Each location had recorded music available to residents in the shared living spaces, and most had a piano in the main lounge area. During the sing-along and music-related activities, residents were observed singing along to songs from memory, engaging with one another and the group leader and smiling. Staff employed by each community varied in their level of musical training and experience, from none to a full-time music therapist in residence. Staff interviewed said they believed music was helpful to aid memory recall, reduce anxiety, and to engage interest. Interestingly, a music therapist at one site also described how music during mealtimes created too much of a distraction for residents and interfered with dietary care. Discussion/Significance of Impact: It is clear from both the staff interviews and direct observations of musical activities that music is important to consider for people living with dementia in care communities. Guidelines for implementation and minimum standards would be helpful to ensure all care community residents can experience benefits highlighted by staff in this study.

Objectives/Goals: Aged keratinocytes are less proliferative than adult, and aged skin heals more slowly. We examined the proliferation kinetics of aged and adult human keratinocytes. We then tested whether an extrinsic agent, sh-oligopeptide-72, can ameliorate these defects. Methods/Study Population: We used live cell imaging (LCI) to examine the proliferation kinetics of aged (73–92y) and adult (34–49y) passage zero human keratinocytes. We then incubated aged keratinocytes with a peptide, sh-oligopeptide-72 (purported to improve keratinocyte proliferation), or vehicle (PBS). Lineage trees of cell divisions were constructed to determine cell cycle duration and the proliferation/differentiation outcomes of each division. To assess wound healing, cells were isolated from 3 patients, 82–92y, and plated in 2-well culture dishes with inserts. Wells were treated with sh-oligopeptide-72 (100 ng/ml) or vehicle (PBS). At confluence, the insert was removed leaving a well-defined 500 μm gap. The time until 100% closure of the defect was obtained using LCI and the wound healing size tool. Results/Anticipated Results: There was no significant difference in the number of stem cell (SC) colonies between aged and adult keratinocytes. However, aged keratinocytes produced more aged committed progenitor (CP) colonies (P<0.0001). Adult CP, but not stem, colonies were significantly larger than aged (P = 0.0001), and this was associated with earlier terminal differentiation (P = 0.0005). Aged SC and CP colonies exhibited a higher proportion of differentiation divisions, and their cell cycle duration (CCD) was increased. Sh-oligopeptide-72 rescued the increased terminal differentiation as well as decreased the CCD in SC colonies. Sh-oligopeptide decreased the mean closure time of the wound assays (143h vs. 204h, P = 0.04). Discussion/Significance of Impact: Sh-oligopeptide-72 reversed many of the proliferation defects that develop in aged SC colonies. Wound assays show that this results in improved keratinocyte function. These results suggest that the age-related changes in growth dynamics can be modified in response to extrinsic signals in vitro.

Objectives/Goals: The goal of this work is to understand the physiological profile of phage susceptibility and identify candidate phage defense mechanisms. Additionally, it aims to determine the host receptors targeted by bacteriophages to infect E. coli O157:H7 through random bar code transposon-site sequencing (RB-TnSeq). Methods/Study Population: A collection of 109 E. coli O157:H7 strains from environmental, food, and animal sources were analyzed, representing phylogenetic lineages corresponding to clades 2, 3, 5, 6, 7, and 8. Phage susceptibility profiles were determined using 23 bacteriophages, assessing plaque morphology. Using the O157:H7 genomes, a genomic analysis was conducted with the Prokaryotic Antiviral Defense Locator (PADLOC), which identified putative phage defense systems through sequence homology. Additionally, 5 RB-TnSeq libraries were generated in representative strains to study loss-of-function mutations. These libraries will be screened against a subset of diverse phage to identify the receptors involved in phage adsorption. Results/Anticipated Results: The phage resistance patterns showed susceptibility varied across clades, suggesting distinct mechanisms. Several defense systems were identified using PADLOC, including restriction-modification, Cas, Lamassu, and Druantia. Phage defense candidate (PDC) systems were identified, showing homology to known systems, though their specific function remains unknown. Clade 7.2 exhibited higher phage resistance and a greater presence of PDCs compared to the other clades. Five saturated RB-TnSeq libraries were constructed in O157:H7, achieving 84.5–89% gene coverage. These libraries will facilitate the identification of receptors involved in phage adsorption and resistance. Discussion/Significance of Impact: This study deepens our understanding of phage resistance in E. coli O157:H7 by identifying key defense systems and receptors. The discovery of novel antiviral mechanisms offers promising targets for phage-based interventions, potentially enhancing strategies for controlling this dangerous pathogen.

Objectives/Goals: This study will assess population heterogeneity in Candida bloodstream infections by quantifying antifungal resistance, fitness, and genomic diversity to understand clonality and develop a high-throughput screening tool to detect population-level resistance to update clinical practice. Methods/Study Population: This study assesses antifungal resistance and population heterogeneity in Candida bloodstream isolates collected through multiple Midwest hospitals. Blood samples are plated to isolate single colonies and population samples, which are then archived. We test resistance to key antifungals using EUCAST guidelines, conduct growth curve assays, and perform whole-genome sequencing to determine genetic diversity. A high-throughput screening method tracks colony growth under different drug conditions using time-lapse imaging and custom analysis software. The findings will reveal the extent of antifungal resistance and genetic variation within infecting populations, informing better clinical management. Results/Anticipated Results: Preliminary analysis of Candida glabrata bloodstream isolates show significant heterogeneity in colony morphology, antifungal resistance, and fitness. Some single colonies exhibit higher minimum inhibitory concentration values for micafungin and fluconazole than the overall population, while others show reduced susceptibility to amphotericin B, highlighting diverse resistance profiles. Growth assays reveal distinct fitness phenotypes within these populations. This variation underscores the limitations of single-colony testing and suggests a need for population-level resistance screening. We anticipate genomic analyses will identify genetic diversity underlying these differences, supporting a more comprehensive clinical approach to treatment. Discussion/Significance of Impact: This study reveals notable intrapopulation heterogeneity in Candida glabrata, including variations in colony morphology, antifungal resistance, and fitness. Findings highlight genomic diversity, introduce a novel screening method for resistance, and emphasize the need for population-level testing in clinical practice.

Objectives/Goals: Communication between clinicians and parents of seriously ill infants is understudied. This study aims to 1) define high-quality communication with parents of critically ill infants and 2) evaluate the psychometric properties and validity of a measure of high-quality communication in parents of critically ill infants. Methods/Study Population: 1) Using participant observation and semi-structured interviews of 35 parents of hospitalized infants, I will conduct content analysis to describe high-quality prognostic communication with parents of infants in the pediatric intensive care unit. Using descriptions captured during participant observation and in semi-structured interviews, I will produce a novel definition of high-quality communication with parents of seriously ill infants. I will also explore parent experiences of communication by race. 2) I will validate a measure of communication quality in parents of 200 neonatal and pediatric intensive care unit patients. I will use factor analysis to evaluate the extent to which responses map onto an established construct and assess dimensionality and reliability. Results/Anticipated Results: 1) I anticipate finding that identification of high-quality communication will be consistent between participant observation and interviews and will track with Wreesmann’s framework. I hypothesize that minoritized parents are more likely to receive low-quality communication. 2) I hypothesize that the measure of communication quality will be valid and reliable in the neonatal and pediatric intensive care units. Discussion/Significance of Impact: I will explore communication quality in a novel setting for which limited data are currently available, establishing a measure for future pediatric communication research and identifying targets for interventions to improve communication quality. Better understanding of communication with parents of sick infants will lead to improved outcomes.