Transcriptomics-based Discovery of Small Molecules Down Regulating NFE2L2 as Potential Therapeutics for Hepatocellular Carcinoma

Presentation Time: 03:30 PM - 03:42 PM

Abstract Keywords: Cancer Prevention, Drug Discoveries, Repurposing, and Side-effect, Causal Inference
Primary Track: Applications

Hepatocellular carcinoma (HCC) is the predominant liver cancer, characterized by high mortality and significant resistance to chemotherapy. The upregulation of transcription factor NRF2 (encoded by the NFE2L2 gene), enhances HCC tumor progression and drug resistance. Since direct inhibition of NRF2 has proven challenging, we leveraged the Library of Integrated Network-based Cellular Signatures (LINCS) and Gene Expression Omnibus (GEO) RNA-seq transcriptomics data to systematically identify novel small-molecule therapeutics downregulating NFE2L2 expression through intermediate genes. Using LINCS data, we selected the top-ranked compounds effectively downregulating NFE2L2 and validated these results with Bayesian network modeling based on RNA-seq from 138 HCC tissues to infer causal relationships. Among identified intermediates, TARDBP and ABCD3 emerged as critical regulators positively driving NFE2L2 expression. Furthermore, our integrated computational pipeline identified two drug candidates predicted to suppress NFE2L2 expression via these intermediates. Mechanistically, TARDBP influences mitochondrial homeostasis and oxidative stress, whereas ABCD3 regulates peroxisomal fatty acid metabolism. We hypothesize that the pathways might be related to NFE2L2-driven carcinogenesis. Our computational approaches highlight the efficacy of combining transcriptomic profiling with causal inference networks to identify novel therapeutic agents. In further study, these candidate drugs would be validated through biological experiments, which may combine with chemotherapy, offering a new treatment in hepatocellular carcinoma.

Speaker:
Charles Zheng, Bachelors
University of Pennsylvania

Authors:
Charles Zheng, Bachelors - University of Pennsylvania; Tianhua Zhai, Ph.D. - University of Pennsylvania; Sumita Garai, Postdoctoral Researcher/Ph.D - University of Pennsylvania; Alvin Pham, Neuroscience - The University of Texas at Austin; Bin Chen, Ph.D. - Saint Joseph's University; Li Shen, Ph.D. - University of Pennsylvania;

---

Discrete-Event Simulation Model for Cancer Interventions and Population Health in R (DESCIPHR): An Open-Source Pipeline

Presentation Time: 03:42 PM - 03:54 PM

Abstract Keywords: Quantitative Methods, Cancer Prevention, Policy, Public Health, Clinical Decision Support, Healthcare Economics/Cost of Care
Primary Track: Foundations
Programmatic Theme: Public Health Informatics

Simulation models can inform cancer screening policy by combining various data sources to forecast the health and economic impacts of screening. However, there is a lack of integrated guidance on structuring, calibrating, and deploying models with advanced simulation and Bayesian techniques. We introduce the Discrete-event Simulation Model for Cancer Interventions and Population Health in R (DESCIPHR), an open-source framework and codebase for constructing models to evaluate the burdens and benefits of interventions for progressive diseases.

Speaker:
Selina Pi, Bachelor of Science in Engineering
Stanford University

Authors:
Carolyn Rutter, PhD - Fred Hutchinson Cancer Center; Carlos Pineda-Antunez, MSc - University of Washington; Jonathan Chen, MD, PhD - Stanford University Hospital; Jeremy Goldhaber-Fiebert, PhD - Stanford University; Fernando Alarid-Escudero, PhD - Stanford University;

---

MedTaxon: Accelerating Biomedical Taxonomy Construction with A Web-based Interactive Tool using Large Language Models

Presentation Time: 03:54 PM - 04:06 PM

Abstract Keywords: Informatics Implementation, Information Extraction, Large Language Models (LLMs)
Primary Track: Applications
Programmatic Theme: Translational Bioinformatics

In the realm of biomedical research, taxonomies can provide hierarchical classifications of concepts within a specific domain. Traditionally, the construction of biomedical taxonomies has mainly relied on manual curation by domain experts, which is inherently a time-consuming, labor-intensive, and costly endeavor. To address the challenges in constructing a taxonomy from textual knowledge, we propose MedTaxon, a serverless web-based tool that integrates interactive user interfaces and LLMs to assist researchers in extracting and structuring domain-specific taxonomies.

Speaker:
Huan He, Ph.D.
Yale University

Authors:
Huan He, Ph.D. - Yale University; Xueqing Peng, PhD - Yale University; Vipina K. Keloth, PhD - Yale University; Lingfei Qian, PHD - Yale University; Qianqian Xie, PhD - Yale University; Na Hong, PhD - Yale University; Hua Xu, Ph.D - Yale University;

---

A Novel Multicomponent Cost Function for Estimating Digital Twins: Estimating Oscillatory Blood Glucose Dynamics with Sparse Data

Presentation Time: 04:06 PM - 04:18 PM

Abstract Keywords: Quantitative Methods, Personal Health Informatics, Precision Medicine
Primary Track: Foundations
Programmatic Theme: Clinical Informatics

This abstract introduces a novel data assimilation method to tackle challenges common to most problems in medical informatics: sparse observations, unreliable models, and non-stationary dynamics. The method employs a multicomponent cost function that ensures model-data agreement, maintains parameter consistency, and penalizes unrealistic parameter changes. Tested on blood glucose estimation, it performs well in both simulated and real data, effectively handling data sparsity and maintaining accurate dynamics.

Speaker:
Melike Sirlanci, PhD
University of Colorado Anschutz

Authors:
Melike Sirlanci, PhD - University of Colorado Anschutz; George Hripcsak, MD - Columbia University Irving Medical Center; Lena Mamykina, PhD - Columbia University; Estaban G. Tabak, PhD - Courant Institute of Mathematical Sciences New York University; David Albers, PhD - University of Colorado, Department of Biomedical Informatics;

---

Predictive Factors and State-Level Barriers to Postpartum Birth Control Usage in the United States: Insights from PRAMS Phase 8

Presentation Time: 04:18 PM - 04:30 PM

Abstract Keywords: Public Health, Delivering Health Information and Knowledge to the Public, Health Equity, Population Health, Policy, Diversity, Equity, Inclusion, and Accessibility
Primary Track: Policy
Programmatic Theme: Public Health Informatics

In the last decade, varied state-level policies on contraception access have highlighted the importance of large-scale public health datasets in assessing the impact of these policies on reproductive healthcare access. This study uses PRAMS Phase 8 (2016-2022) data to examine predictive factors of postpartum birth control use, hypothesizing that state policies impact contraception uptake and barriers, particularly regarding the expansion of immediate postpartum long-acting reversible contraception (LARC) reimbursement policies. Two distinct logistic regression models were constructed, and the inclusion of state as a covariate significantly reduced residual deviance (ΔDeviance = 13.696, p=0.0002). This finding indicates that state of residence is a statistically significant predictor of postpartum birth control usage. This study underscores the significant impact of state-level and institutional policies on birth control usage and LARC uptake, emphasizing the need for informed policy changes and patient-centered strategies to address disparities and improve postpartum reproductive health outcomes.

Speaker:
Pitchaya Chairuengjitjaras, MS in Health Informatcs
Yale University

Authors:
Amber Tran, MS Health Informatics - Yale University; Ameesha Masand, MS in Health Informatics - Yale University;

---

Quantifying Cognitive Decline and Balance in Older Adults with Mild Cognitive Impairment using Wearables in Clinical Environments

Presentation Time: 04:30 PM - 04:42 PM

Abstract Keywords: Artificial Intelligence, Fairness and elimination of bias, Machine Learning, Mobile Health, Diversity, Equity, Inclusion, and Accessibility
Primary Track: Applications
Programmatic Theme: Clinical Research Informatics

Monitoring gait in Mild Cognitive Impairment (MCI) is crucial for early intervention. We collected waist-worn Inertial Measurement Unit data from 45 older adults over 6 months during free-movement and social interactions in a real-world therapeutic environment. Machine learning models predicted Montreal Cognitive Assessment and Mini-Balance Evaluation Systems Test scores with mean absolute errors 2.45 and 4.11 respectively, identifying stride variability and direction change as most relevant features. Bias mitigation reduced demographic disparities, offering equitable outcomes.

Speaker:
Hyeokhyen Kwon, Ph.D.
Emory University

Authors:
Dharini Raghavan, Graduate Student - Georgia Institute of Technology; Bolaji Omofojoye, Clinical Research Coordinator - Emory University; Soheil Saghafi, Postdoctoral Fellow - Emory University; Yashar Kiarashi, Postdoctoral Fellow - Emory University; Allan Levey, Professor - Emory University; Amy Rodriguez, Assistant Professor - Emory University; Gari Clifford, Professor - Georgia Institute of Technology and Emory University; Hyeokhyen Kwon, Assistant Professor - Emory University and Georgia Institute of Technology;

---