Caleb Hallinan
Biomedical Engineering PhD Candidate · Johns Hopkins University
I am a PhD candidate in Biomedical Engineering at Johns Hopkins University, working in the JEFworks Lab under the mentorship of Dr. Jean Fan, where I develop machine learning and computational methods for spatial transcriptomics, digital pathology, and tissue-scale biological analysis. I care about building practical, reproducible, open-source tools that help biologists make sense of complex tissues.
Research
My research sits at the intersection of machine learning and tissue biology, with an emphasis on making spatial and imaging technologies more accurate, interpretable, and usable.
Spatial transcriptomics
Computational analysis and benchmarking of imaging- and sequencing-based spatial transcriptomics data to study the spatial organization of tissues.
Digital pathology & deep learning
Predicting spatial gene expression from H&E histology images, and understanding how data quality shapes model performance and interpretability.
Data & probe quality
Assessing measurement fidelity in spatial platforms, including off-target probe binding, to improve the biological interpretability of spatial data.
Subtype discovery
Feature-selection and metric-learning methods that preserve heterogeneity to uncover disease and cell-state subtypes from transcriptomic data.
Live-cell image analysis
Deep learning for cell segmentation, tracking, and morphodynamic profiling of live-cell imaging.
Open-source software
Building user-friendly, reproducible computational tools for biologists who may not be programmers.
Selected Works
First-author publications. Names in bold indicate me; an asterisk (*) denotes co-first authorship. See All Publications for the complete list.
eLife, 2026 Journal
bioRxiv, 2025 Preprint
Nature Communications, 2025 Journal
Featured Software
Open-source tools for spatial and imaging-based transcriptomics. More on my GitHub.
Off-target Probe Tracker (OPT)
A tool associated with the Xenium off-target probe binding work for identifying putative off-target probe binding through sequence-alignment-based analysis.
P-Het
Software for Preserving Heterogeneity / subtype discovery from transcriptomic data, associated with the Nature Communications P-Het paper.
SpatialMNN
spatialMNN, an algorithm that integrates multiple spatial transcriptomic samples and identifies spatial domains.
STARIT
Converts transcripts within segmented cells in imaging-based spatial transcriptomics data into rasterized image tensors, enabling deep-learning models to characterize cell states from subcellular molecular heterogeneity rather than gene counts alone.
News
Jun 2026
Attended the HuBMAP Final Hurrah and Spatial Biology meetings at the NIH.
Milestone
May 2026
Version of record published in eLife: off-target probe binding in 10x Genomics Xenium gene panels, introducing the Off-target Probe Tracker (OPT) tool.
Publication
Dec 2025
New bioRxiv preprint, STARIT, characterizing cell states with subcellular molecular heterogeneity in spatial transcriptomics data.
Publication
Apr 2025
PHet published in Nature Communications — a heterogeneity-preserving feature-selection method for disease-specific subtype discovery (co-first author).
Publication
Teaching
I care deeply about teaching and mentorship, with a long-term goal of becoming a teaching professor. I currently mentor two students and have designed and taught two original courses on neural networks and deep learning for spatial transcriptomics, in addition to five teaching assistant roles at Johns Hopkins, Harvard Medical School, and the University of Virginia. See Teaching & Mentorship for the complete history.
Experience & Education
Education
PhD, Biomedical Engineering Johns Hopkins University — JEFworks Lab (Advisor: Jean Fan), Baltimore, MD
Aug 2023 – Present
B.A., Statistics & Biology University of Virginia, Charlottesville, VA
2017 – 2021
Experience
Research Assistant II Boston Children’s Hospital & Harvard Medical School — Lee Lab (Advisor: Kwonmoo Lee), Boston, MA
Sep 2021 – Jun 2023
Undergraduate Researcher University of Virginia (Advisors: Tianxi Li, Frederic Padilla), Charlottesville, VA
2018 – 2021
Contact
The best way to reach me is by email. You can also find me on GitHub, Google Scholar, ORCID, and LinkedIn, or download my CV.