BIOMED PhD Research Proposal

Title: 
Improving Molecular Diagnostics for Rare Mendelian Disorders Using Long Read Sequencing

Speaker:
Tanaya Jadhav, PhD Candidate
School of Biomedical Engineering, Science and Health Systems
Drexel University

Advisors:
Ahmet Sacan, PhD
Associate Professor
School of Biomedical Engineering, Science and Health Systems
Drexel University

Ramakrishnan Rajagopalan, PhD
Senior Principal Scientist
Division of Genomic Diagnostics 
Children's Hospital of Philadelphia (CHOP)

Details:
Rare Mendelian disorders, while individually rare, collectively affect 25–30 million Americans. Despite the implementation of short-read sequencing (SRS) as the standard of care, approximately 50–60% of patients remain without a definitive molecular diagnosis. This diagnostic gap is in part driven by the inherent technical limitations of SRS, including the inability to resolve variants in "dark" genomic regions, determine phase, or capture DNA methylation without fragmented, multi-tiered testing. While Long-Read Sequencing (LRS) offers a transformative solution by consolidating structural variant detection, phasing, and methylation into a single assay, its adoption in research and clinical settings is hindered by a lack of standardized, validated bioinformatics infrastructure.

This work aims to bridge this gap through the development of two integrated workflows: WAFL (Workflow for Annotation, Filtering, and Prioritization of LRS variants) and MitoPac. WAFL provides a comprehensive, platform-agnostic framework for prioritizing the expanded variant spectrum of the nuclear genome, enabling the identification of compound heterozygous variants, tandem repeat expansions, and imprinting disorders through native methylation and phasing data. MitoPac addresses the critical shortcomings of standard-of-care mitochondrial diagnostics by enabling precise characterization of multiple mitochondrial deletions, SNV calling, and accurate heteroplasmy quantification from LRS without the amplification biases inherent in traditional SRS-based assays.

Without specialized annotation workflows, the expanded variant spectrum provided by LRS, including large SVs, tandem repeats, phased haplotypes, and differential methylation, remains largely overlooked and underutilized. The objective of this work is to bridge this gap by developing integrated variant analysis workflows designed specifically for LRS data. These pipelines will provide a comprehensive solution for filtering, annotating, and prioritizing variants across both nuclear and mitochondrial genomes. We hypothesize that integrating multi-omic data from long-read sequencing in a streamlined variant analysis workflow will provide additional diagnostic yield for rare Mendelian disorders. 

BIOMED Seminar

Title:
Biomedical Data Science for Brain Health: From Patient Cohorts to Cellular Circuits

Speaker:
Raha Dastgheyb, PhD
Assistant Professor of Neurology
Co-Director
JH-CAHN Biomarker Core
School of Medicine 
John's Hopkins University

Details:
Biomedical brain research increasingly generates rich, heterogeneous data across scales: from longitudinal neuropsychological testing and electronic health records to blood biomarkers, imaging, and patient-derived cellular models. The central challenge is no longer data scarcity, but biological and clinical heterogeneity: patients with similar symptoms may have distinct underlying mechanisms, while shared mechanisms may manifest as different clinical phenotypes across individuals and populations. Unraveling this heterogeneity and linking phenotypes to mechanism is essential for building more precise and targetable strategies to preserve neurological health.

This talk will illustrate how biomedical data science can bridge patient cohorts and cellular circuits to enable more mechanistic, scalable, and ultimately targetable approaches to brain health across diverse populations and diseases, it will show how engineering processes can translate to data-driven methods and reveal mechanistic insights that are not apparent within any single modality. At the cellular and circuit level, it will highlight how high-dimensional electrophysiology, including multi-electrode array recordings, can be used to quantify neural dysfunction and perturbation responses in ways that complement patient-scale inference and emphasize practical machine learning approaches that prioritize reproducibility and interpretability. 

Biomedical data science is most powerful when it is mechanistically grounded and methodologically rigorous. The goal is not simply more data, but better questions and better tools for understanding and improving human health. 

Biosketch:
Raha Dastgheyb, PhD, is an Assistant Professor of Neurology at the Johns Hopkins School of Medicine whose work sits at the intersection of engineering, neuroscience, and data science. She received her PhD in Biomedical Engineering from Drexel University, where her dissertation under the mentorship of Dr.Ken Barbee investigated the mechanisms of secondary axonal pathology in traumatic brain injury, asking why damage can continue to unfold long after the initial insult.

After Drexel, she completed postdoctoral training at Johns Hopkins that paired quantitative methods with high-throughput experimental neuroscience, including multi-electrode array studies in neurons and stem cell derived organoids. That combination of mechanistic biology and scalable computation became a theme of her work: using rigorous analytic approaches to extract meaning from complex signals, without losing sight of the underlying biology.

This foundation led her to translate advanced analytic tools to human studies, including work that has produced the largest cognitive phenotyping analysis to date in women with HIV and helped catalyze similar efforts internationally. At Johns Hopkins, she now co-directs two translational cores: the Data Science and Mathematical Modeling(DSMM) Core of the Brain Health Program and the Biomarker Core of the Johns Hopkins Center for HIV Associated Neurocognitive Disorders (JH-CAHN). In these roles, she integrates multi-omics, neuroimaging, and both supervised and unsupervised machine learning methods to identify drivers of cognitive decline and mental health changes across multiple neurological and neurodegenerative conditions. 

Rooted in her training in Biomedical Engineering, Dr.Dastgheyb is guided by the principle that scientific questions, not existing tools, should dictate the methods. She is also deeply committed to reproducibility, interpretability, and making analytic work accessible.

The School of Biomedical Engineering, Science and Health Systems is pleased to announce its 8th Annual Immune Modulation & Engineering Symposium (IMES).