Biomedical Informatics Commons^™

Essays On Ai-Driven Analytics For Enhanced Healthcare Delivery: From Onset To Outcomes In Neurological Diseases Management, Gabriel Owusu

Theses and Dissertations

Chronic neurological diseases, particularly Alzheimer's disease and related dementias (ADRDs), pose mounting global healthcare burdens. Despite advances in AI and health information technologies, early detection, precise diagnosis, and effective caregiver support remain elusive. This dissertation proposes innovative AI-enabled analytical frameworks to address these challenges across the full ADRD continuum.

The first essay designs an explainable AI-enabled clinical decision support system (CDSS) using a graph neural network for early AD identification, evaluated through usability studies demonstrating improved risk awareness and clinical engagement. The second essay introduces MADT-RNN, a multi-level attention-based deep transfer recurrent neural network that fuses convolutional and recurrent architectures …

Computational Discovery Of Sars-Cov-2 Viral Entry Inhibitory Peptides From Androctonus Mauretanicus Scorpion Venom: Molecular Docking And Molecular Dynamics Simulations Targeting The Spike Protein, Reda Chahir, Salaheddine Redouane, Jacob Galan, Hicham Hboub, Lahoussaine Aserrar, Salma Chakir, Ahmed Salim Lahlou, Hinde Aassila, Rachid El Fatimy, Naoual Oukkache

School of Medicine Publications

Background and Objective: 

While vaccination remains central to controlling the COVID-19 pandemic, the emergence of SARS-CoV-2 variants with partial resistance to immune responses has highlighted the need for complementary therapeutic strategies. Among these, antiviral agents that inhibit viral entry mechanisms are of particular interest. Animal venoms, especially scorpion venoms, are a rich source of bioactive peptides with potential antiviral properties. This study aimed to evaluate peptides derived from the Moroccan scorpion Androctonus mauretanicus as inhibitors of SARS-CoV-2 spike glycoprotein, which mediates virus entry into host cells via ACE2 receptor binding.

Material and Methodology: 

Six peptides from the venom of the …

A Picture Tells A Thousand Words—, But Ecg Signals Have More To Say, Ashley N. Gomez

Theses and Dissertations

With the increasing adoption of deep learning classification models in the medical domain, a critical challenge remains: achieving high predictive accuracy while maintaining clinical Inter-pretability. This study examines how model architecture, dataset origin, and the use of full versus subset data affect both classification performance and Interpretability in Electrocardiogram (ECG) signal analysis. ResNet18 is evaluated using an open-source ECG Image Dataset, thus a custom dataset derived from digitized ECG images. Post-hoc explainability methods, such as Integrated Gradients, are applied to determine which time steps have the most significant influence on model decisions. The findings demonstrate that model architecture and dataset …

A Molecular Renaissance In Alzheimer's Disease Research: The Rise Of Systems Biology And Spatial Omics, Lucia M. Perez-Navarro, Juan Lopez-Alvarenga

School of Medicine Publications

Recent advances in multi-omics and spatial proteomics are reshaping our understanding of Alzheimer's disease. Guo et al.1 applied integrative multi-omics to stratify mild cognitive impairment into biologically distinct subtypes with divergent progression trajectories: one metabolically impaired and slow-progressing, the other immune-activated and rapidly declining. Current techniques such as STC-DESI, LCM-MS, and machine learning enhance regional proteomic resolution, supporting biomarker discovery and spatially targeted interventions. This work exemplifies a broader shift toward precision medicine and a systems-level molecular framework in neurodegenerative disease research.

Inferred Global Dense Residue Transition Graphs From Primary Structure Sequences Enable Protein Interaction Prediction Via Directed Graph Convolutional Neural Networks, Islam A. Ebeid, Haoteng Tang, Pengfei Gu

Computer Science Faculty Publications

Introduction: Accurate prediction of protein-protein interactions (PPIs) is crucial for understanding cellular functions and advancing the development of drugs. While existing in-silico methods leverage direct sequence embeddings from Protein Language Models (PLMs) or apply Graph Neural Networks (GNNs) to 3D protein structures, the main focus of this study is to investigate less computationally intensive alternatives. This work introduces a novel framework for the downstream task of PPI prediction via link prediction.

Methods: We introduce a two-stage graph representation learning framework, ProtGram-DirectGCN. First, we developed ProtGram, a novel approach that models a protein's primary structure as a hierarchy of …

Bioinformatics-Guided Identification And Quantification Of Biomarkers Of Crotalus Atrox Envenoming And Its Neutralization By Antivenom, Auwal A. Bala, Anas Bedraoui, Salim El Mejjad, Nicholas K. Willard, Joseph D. Hatcher, Anton Iliuk, Joanne E. Curran, Elda E. Sanchez, Montamas Suntravat, Emelyn Salazar, Rachid El Fatimy, Tariq Daouda, Jacob Galan

School of Medicine Publications

Quantitative mass spectrometry-based proteomics of extracellular vesicles (EVs) provides systems-level exploration for the analysis of snakebite envenoming (SBE) as the venom progresses, causing injuries such as hemorrhage, trauma, and death. Predicting EV biomarkers has become an essential aspect of this process, offering an avenue to explore the specific pathophysiological changes that occur after envenoming. As new omics approaches emerge to advance our understanding of SBE, further bioinformatics analyses are warranted to incorporate the use of antivenom or other therapeutics to observe their global impact on various biological processes. Herein, we used an in vivo BALB/c mouse model and proteomics approach …

Machine Learning Quantification Of High-Resolution Tissue Microarray (Tma) Image On Muc13 Ihc Analysis, Beibei Huang, Aiko Yamaguchi, Jianbo Wang, Shilpa Sharma, Zhiwen Liu, Henry Charles Manning

Research Symposium

Background High-resolution tissue microarray (TMA) technology allows prompt molecular profiling of multiple tissue specimens, making it ideal for analyzing candidate biomarkers quickly and effectively [1, 2]. Integrating TMA with digital pathology and machine learning enhances high-throughput, cost-effective studies, offering advanced image analysis and improved diagnostic accuracy. MUC13 (Mucin 13) is a transmembrane glycoprotein frequently overexpressed in colorectal cancer (CRC) [3]. MUC13 contributes to colonic tumorigenesis, progression and metastasis [4, 5], making it an attractive target for antibody-guided radiotheranostics in CRC. This study investigates the expression pattern of MUC13 and its association with patients' clinical characteristics in primary and metastatic CRC …

Gene Co-Expression Networks And Descriptive Statistical Patterns In Cancer Subtypes, Arely Solis, Marzieh Ayati

Research Symposium

There are many cancers that are affecting the human population, with some being more common and studied than others. These cancers have mostly been studied individually until 2012 when scientists began studying through comparison analysis of different cancers to see possible connections on the genomic and molecular level and have resulted in the categorization of tumors into types. The result from molecular analysis has re-classified types of tumors into new clusters, which aid doctors in deciding the optimal way of treating tumors. In this study, we analyze a dataset comprising over 2,000 cancer samples, focusing on six cancer types: breast, …

Symbol-Temporal Consistency Self-Supervised Learning For Robust Time Series Classification, Kevin Garcia, Cassandra Garza, Brooklyn Berry, Yifeng Gao

Computer Science Faculty Publications

The surge in the significance of time series in digital health domains necessitates advanced methodologies for extracting meaningful patterns and representations. Self-supervised contrastive learning has emerged as a promising approach for learning directly from raw data. However, time series data in digital health is known to be highly noisy, inherently involves concept drifting, and poses a challenge for training a generalizable deep learning model. In this paper, we specifically focus on data distribution shift caused by different human behaviors and propose a self-supervised learning framework that is aware of the bag-of-symbol representation. The bag-of-symbol representation is known for its insensitivity …

Empowering The Biomedical Research Community: Innovative Sas Deployment On The All Of Us Researcher Workbench, Izabelle Humes, Cathy Shyr, Moira Dillon, Zhongjie Liu, Jennifer Peterson, Chris St Jeor, Jacqueline Malkes, Hiral Master, Brandy Mapes, Bassent Abdelbary

Physician Assistant Studies Faculty Publications

Objectives: The All of Us Research Program is a precision medicine initiative aimed at establishing a vast, diverse biomedical database accessible through a cloud-based data analysis platform, the Researcher Workbench (RW). Our goal was to empower the research community by co-designing the implementation of SAS in the RW alongside researchers to enable broader use of All of Us data.

Materials and methods: Researchers from various fields and with different SAS experience levels participated in co-designing the SAS implementation through user experience interviews.

Results: Feedback and lessons learned from user testing informed the final design of the SAS application.

Discussion: The …

Optimization Of Gait Analysis System For Clinical Applications, Gerardo Medellin, Katherine S. Bolado, Daniel Salinas, Kelsey Potter-Baker

Research Symposium

Background: Adequate gait function is pivotal for many activities of daily living and high quality of life. Following many neurodegenerative diseases, such as Parkinson’s Disease, gait abnormalities can manifest and range from reduced stride length, inability to turn, foot drop or shuffling. To track and monitor such changes in gait, gait analysis techniques are gaining clinical popularity and have the ability to gather a range of data in a short duration. Gait analysis techniques go beyond simple visual observation and include instrumental gait analysis and weight distribution of the gait cycle. Here, we sought to optimize the gait analysis …

Assessing Gait Metrics For Early Parkinson's Disease Prediction: A Preliminary Analysis Of Underfit Models, Daniel Salinas, Gerardo Medellin, Katherine Bolado, Tomas Gomez, Kelsey Potter-Baker, Nawaz Khan Abdul Hack, Ramu Vadukapuram

Research Symposium

Background: Parkinson's Disease (PD) is characterized by both motor and non-motor symptoms, and its diagnosis primarily relies on clinical presentation. There is a growing need for diagnostic tools to identify the early signs of PD, particularly the initial motor impairments often manifested as gait abnormalities. Here we seek to present preliminary findings to address this need. Our study focuses on using Machine Learning techniques (ML) to predict the PD clinical stage most efficiently and accurately. Specifically, we have sought to evaluate how spatiotemporal characteristics and other locomotor performance variables obtained on a walkway system can be utilized to identify the …

Uncovering The Complex Genetic Architecture Of Human Plasma Lipidome Using Machine Learning Methods, Miikael Lehtimäki, Binisha H. Mishra, Coral Del-Val, Leo-Pekka Lyytikäinen, Mika Kähönen, C. Robert Cloninger, Olli T. Raitakari, Reijo Laaksonen, Igor Zwir, Terho Lehtimäki, Pashupati P. Mishra

School of Medicine Publications

Genetic architecture of plasma lipidome provides insights into regulation of lipid metabolism and related diseases. We applied an unsupervised machine learning method, PGMRA, to discover phenotype-genotype many-to-many relations between genotype and plasma lipidome (phenotype) in order to identify the genetic architecture of plasma lipidome profiled from 1,426 Finnish individuals aged 30–45 years. PGMRA involves biclustering genotype and lipidome data independently followed by their inter-domain integration based on hypergeometric tests of the number of shared individuals. Pathway enrichment analysis was performed on the SNP sets to identify their associated biological processes. We identified 93 statistically significant (hypergeometric p-value < 0.01) lipidome-genotype relations. Genotype biclusters in these 93 relations contained 5977 SNPs across 3164 genes. Twenty nine of the 93 relations contained genotype biclusters with more than 50% unique SNPs and participants, thus representing most distinct subgroups. We identified 30 significantly enriched biological processes among the SNPs involved in 21 of these 29 most distinct genotype-lipidome subgroups through which the identified genetic variants can influence and regulate plasma lipid related metabolism and profiles. This study identified 29 distinct genotype-lipidome subgroups in the studied Finnish population that may have distinct disease trajectories and therefore could be useful in precision medicine research.