Embedded Files
Research Areas
I have primarily worked in the following application areas:
Biomedical imaging (including histopathology, neuroimaging, and vascular imaging)
Electronic health records (including transplant registry data)
I consider my machine learning specialties to be:
Unsupervised clustering
High-dimensional regression (especially lasso-based approaches)
Random forest
XGBoost
Ensemble learning
Feature/variable selection
Data harmonization
Active and Completed Projects
Check out the projects in progress or completed by my trainees (or Champs as I sometimes call them to keep with the theme of the lab acronym)!
Active Projects
Kidney Function and Kidney Trasplant Allocation Prediction with Multimodal Clinical and Omics Data
Traditional diagnoses of kidney disease are made by manual visual assessment of renal biopsy tissue. Such diagnoses are subject to intra- and inter-rater variability, and we believe that the use of omics data such as renal biopsy pathomics and genomics data can help improve the prediction of kidney function outcomes and lead to the identification of novel biomarkers of kidney disease. We leverage high-dimensional regression, unsupervised clustering, random forest, XGBoost, and data harmonization approaches to predict continuous and binary outcomes of kidney function as well as whether a kidney was allocated for transplantation using patients' clinical and omics data. In our MACHAMP Lab research for this aim, we consider the heterogeneity of renal biopsy tissue and histologic objects by disease severity/scarring, the hierarchical data structure of image features, batch effects presented by multisite studies, and making the output as well as interpretations of these machine learning models visually accessible to a practicioner with clinical decision support systems.
A Cluster-Aware Pathomic Framework for Predicting Delayed Graft Function and Allograft Outcomes from Deceased Donor Procurement Biopsies (Champ(s): Ketki)
Optimized Renal Allocation via Computational Pathomics (Champ(s): Rachel, Ketki)
Integrating Tubule-Level Procurement Biopsy Pathomics and Clinical Factors for Machine Learning Prediction of Delayed Graft Function (Champ(s): Huiqian)
Adapting Machine Learning Models Using Multisite Histopathology Data for Predicting Kidney Function (Champ(s): Ritesh, Advay, Raymond, Janelle, Siyuan)
A Web-Based Application for Machine Learning Predictions of Kidney Transplant Outcomes (Champ(s): Abby, Emilie, Abhinav)
Mapping Kidney Trait Heritability to Individual Cells Revelals Disease-Specific Remodeling of Genetic Risk Architecture (Champ(s): Huiqian): https://doi.org/10.64898/2026.04.12.717976
Genome-wide association studies (GWAS) have identified hundreds of genetic loci associated with kidney function and disease, yet the cell-type-specific mechanisms through which these variants act remain largely unknown. Here, we construct the Kidney Genetic Disease Cell Atlas by applying single cell disease relevance scoring (scDRS) to map GWAS signals for six kidney-related traits-estimated glomerular filtration rate (eGFR), cystatin C-based eGFR (eGFRcys), blood urea nitrogen (BUN), urinary albumin-to-creatinine ratio (UACR), type 2 diabetes (T2D), and IgA nephropathy (IgAN) onto a comprehensive single-nucleus RNA-seq atlas of 304,652 kidney cells spanning five clinical conditions (healthy reference, acute kidney injury [AKI], COVID-19-associated AKI [COV-AKI], diabetic kidney disease [DKD], and hypertensive chronic kidney disease [H-CKD]).
Statistical Methods for Structured, High-Dimensional Biomedical Data
A central theme of the MACHAMP Lab research is developing statistical and machine learning methods for structured, high-dimensional biomedical data, where standard regression approaches break down due to unbalanced, matrix- or tensor-valued predictors or complex correlation structure. An example of such data arises from renal histopathology image features which can be measured per histologic object per renal biopsy scan subject. To model such complex data structures, we have developed a novel scalar-on matrix regression framework for unbalanced feature matrices called the CLUstering Structured lasSO (https://doi.org/10.1007/s12561-025-09476-7), which accommodates subject-level predictor matrices that are unbalanced in dimension and uses L1-penalization to identify features informative of an outcome. We extended this framework to handle high correlation among these structured features in a follow-up study (https://doi.org/10.1002/sam.70078). Many Champs are building upon this methodological foundation to develop their own L1-penalized regression approaches for structured, high-dimensional biomedical data!
Scalar-on-Tensor Regression with Unbalanced Tensor Predictors (Champ(s): Alec)
False Discovery Rate-Controlled Scalar-on-matrix Regression for Unbalanced Feature Matrices (Champs(s): Danny)
Scalar-on-matrix Logistic Regression for Unbalanced Feature Matrices (Champ(s): Hedwig)
Ensemble Learning and Conformal Prediction for Patient Outcomes
Rather than using individual machine learning models for prediction of patient outcomes, what if we jointly leveraged the predictive abilities of multiple machine learning models for predictions of these outcomes? In the MACHAMP Lab, we are developing new, flexible ensemble learning models to better classify patient health outcomes in a way that prioritizes minimizing the computational complexity as well as maximizes the interpretability of our ensemble models. We believe that such ensemble learning approaches will also offer the potential for biomarker discovery by capturing more complex relationships/dependencies between the predictors and patient health outcomes that may be missed by less comprehensive ensemble models or individual machine learning models.
While many machine learning and deep learning architectures have shown promise of great predictive accuracy on biomedical data, these models are often overconfident in their predictions and measures are not taken to appropriately quantify the uncertainty around such predictions. Conformal prediction offers a framework to rigorously quantify the uncertainty in our machine-learning predictions of patient outcomes and can assist practitioners with clinical-decision making such as disease staging per subject, promoting personalized medicine. In the MACHAMP Lab, we have explored conformal prediction in the context of renal histopathology data for prediction of kidney function (https://doi.org/10.1007/s10742-026-00392-x), and we are looking to expand our application of conformal prediction more data types, such as gait video data.
Uncertainty Quantification of Predicting Gait Outcomes Using Multimodal Conformal Prediction (Champ(s): Jai)
Flexible Ensemble Learning-based Classification of Health Outcomes (Champ(s): Chris)
Completed Projects
A Pathomic-Ensemble Strategy for Exploring Histological Signatures of eGFR Decline in IgAN (Champ(s): Connie, Lylybell, Huiqian): https://doi.org/10.1117/12.3087852
IgA nephropathy (IgAN) is the most common glomerulonephritis worldwide. Clinicians rely on kidney histology and clinical data, such as estimated GFR (eGFR), to obtain a patient’s prognosis. Current clinical tools, including a histologically obtained MEST-C score (histological object-level features) and the International IgAN Prediction Tool (IIgAN PT), estimate the patient’s odds of end-stage kidney disease (ESKD). The histopathologic evaluation of the kidney biopsies relies on inter-observer reproducibility and qualitative interpretation. Pathomics, which uses computational image analysis to extract quantitative features, offers an objective alternative to visual scoring and can uncover histologic signatures with prognostic significance. Cluster-Aware enseMblE Learning with pathOMIC featureS (CAMELOMICS) uses non-sclerotic glomeruli pathomic data and clinical data from IgAN patients at a single institution to estimate 5-year eGFR. Our pipeline clusters the glomeruli and trains high-dimensional regression models on the cluster-level information that may better capture objects’ heterogeneity in disease presentation.
A Pathomics-Integrated Approach Toward Improved Prediction of Kidney Survivability Up to 5 Years Post-Biopsy in IgA Nephropathy Patients (Champ(s): Lylybell, Connie): https://doi.org/10.1117/12.3087510
Immunoglobulin A nephropathy (IgAN) is the most common glomerular disease worldwide and is a leading cause of chronic kidney disease. The current paradigm of IgAN risk stratification, the International IgAN Prediction Tool (IIgAN-PT), relies on clinical and histologic data collected at the time of biopsy. Clinical data includes variables such as estimated glomerular filtration rate (eGFR), proteinuria, and blood pressure. Histologic data from the diagnostic kidney biopsy are assessed by the Oxford classification system; however, its reliability can be significantly affected by interobserver variability. The emergence of “pathomics,” in which morphometric features are extracted from histologic objects, aims to address this issue and provides a more objective and reproducible assessment. In this study, we created a dataset of clinical variables and averaged pathomic features from non-globally sclerotic glomeruli, arteries/arterioles, and tubules. We used this data to train and test logistic regressions (LGs) to predict whether patients would experience kidney function decline (i.e., eGFR decline >50% from baseline or dialysis initiation within 5 years post-biopsy).
Please reach out if you are interested in working with us! - you could be featured on this page!
Lab membership comes with your own custom Pokémon avatar (see below for Pokémon avatars submitted by trainees or ones me or my trainees made of collaborators)!
Lab Members/Champs
The man who depicts himself as riding the Machamp!
Jeremy Rubin, Lab Director and Self-Appointed Pun Expert
Clinical Assistant Professor of Biostatistics, University of Maryland, College Park
Email: jrub@umd.edu
X/Bluesky: @super_jrub
Jeremy Rubin, PhD is a Clinical Assistant Professor of Biostatistics at the University of Maryland, College Park. He received his undergraduate degrees in Statistics and Mathematics from the University of Maryland, Baltimore County before matriculating to the University of Pennsylvania, where he completed his PhD in Biostatistics. Jeremy's main research focuses are the development and application of statistical methods for renal histopathology and kidney transplantation data, but he also previously done research in application areas including computed tomography angiography (CTA) imaging, structural magnetic resonance imaging (MRI), wearable device data, kidney disease (including patient-reported outcomes, autosomal recessive polycystic kidney disease), and inflammatory bowel disease. While Jeremy believes that most problems can be solved with a high-dimensional regression model (especially lasso-based approaches), he also worked with statistical techniques including random forests, survival analysis, mixed models, linear/logistic regression analyses, and conformal prediction. Besides research, Jeremy is also passionate about finding coffee, playing badminton, making puns, as well as watching and performing stand-up comedy!
The Working Folks!
Abby Sun, She'll be there in a dash(board)!
Research Associate, Institute for Asthma & Allergy
Email: abbyjsun@gmail.com
Abby Sun is currently a Research Associate at the Institute for Asthma & Allergy, where she specializes in flow cytometry and supports clinical and validation studies in immunotherapies and immunological diagnostics. A recent MSPH graduate in Global Disease Epidemiology and Control from Johns Hopkins Bloomberg School of Public Health and an UMD alumni, she is passionate about applying evidence-based and translational approaches to advance clinical health outcomes for underserved and vulnerable populations. Her past research has focused on infectious disease epidemiology, vaccine science, and clinical immunology, including coordinating a respiratory disease surveillance study in rural Zambia and contributing to vaccine coverage and serosurveillance research. She is excited to join the MACHAMP Lab to further explore machine learning and statistical approaches for modeling and improving clinical insights. Outside of academia, Abby enjoys traveling, hiking, and playing Pokémon Go.
Emilie Chang
Recent ScM Graduate in Epidemiology, Johns Hopkins University
Email: ychen750@alumni.jh.edu
Emilie recently completed her ScM in Epidemiology at Johns Hopkins University. Some highlights of her experience include mixed-methods analysis nested in an oncology clinical trial, pharmacovigilance data mining, and linear mixed effects modeling on longitudinal cohort data from the Chronic Kidney Disease in Children (CKiD) study. She is passionate about advancing skills and knowledge in machine learning and its application to healthcare data and outcome prediction. Emilie enjoys practicing the piano, running, and playing Pikmin Bloom during her personal time.
Siyuan Chen
Research Assistant, Arizona State University
Email: flyingricole@gmail.com
Danny Cao
Email: caodanny2000@gmail.com
The Postdocs!
Huiqian Hu, Badminton Player with Great Taste in Racket Specs
Postdoc in Transplant Immunology, Stanford University
Email: huiqian.hu@utah.edu
Huiqian Hu recently received his PhD in Molecular Pharmaceutics at University of Utah and just started his postdoc in transplant immunology at Stanford. He is interested in statistical and machine learning methods for high-dimensional biological data analysis. His research focuses on developing random forest and XGBoost frameworks for predictive modeling in precision medicine, with expertise in multi-omics integration and computational biology. He is excited to collaborate with the MACHAMP Lab as a trainee under the supervision of Dr. Rubin to explore novel statistical approaches—particularly conformal prediction and high-dimensional regression methods—for improving model reliability in clinical applications.
Ketki Joshi, everyone clusters around to see the great work she's doing!
Postdoc in Statistical Genetics, Cornell University
Email: joshiketki4@gmail.com
Ketki Joshi recently completed her PhD in Cell and Molecular Biology at the University of Texas at Dallas and started a postdoc in Statistical Genetics at Cornell University. Her doctoral work centered on developing multi-omics computational pipelines, including Hi-GREx, a 3D genome-guided framework that integrates Hi-C spatial data to improve gene expression prediction-achieving a 77% accuracy improvement and identifying new Alzheimer's disease–associated genes. She also completed an internship in Oncology R&D Data Science at AstraZeneca, where she developed causal inference pipelines for driver gene identification and validated actionable biomarkers for precision medicine programs. With expertise spanning bulk RNA-seq, statistical genetics, and machine learning, Ketki is excited to bring her computational toolkit to the MACHAMP Lab and explore novel statistical approaches for clinical applications. Outside the lab, she unwinds with Agatha Christie mysteries- because apparently solving the secrets of the genome isn't enough of a puzzle.
The Grad Students!
Jialu Zhou
PhD student in Statistics at The University of Maryland, College Park
Email: jzhou122@umd.edu
Yasin Saleem
MS student in Data Science at The University of Maryland, College Park
Email: ysaleem@umd.edu
Hedwig Nordlinder, If there's a will then there's a theorem for that!
MSc Mathematical Statistics student at Stockholm University
Email: hedwignordlinder@gmail.com
Biostatistical Research Assistant, Karolinska Institute. Voting member Swedish Society of Actuaries. Research interests include high dimensional statistics, Bayesian inference, MCMC, Phylogenetics and generative machine learning models.
The undergrads!
Ritesh Reddy Thipparthi, Don't sleep on his models!
Undergraduate Student in Computer Science, University of Maryland, College Park
Email: rthippar@umd.edu
Ritesh is a junior CS major at UMD with interests in machine learning for biomedical and clinical applications. Before joining the MACHAMP Lab, he built real-time systems for EEG focus tracking and voice-driven systems, which convinced him that nothing is more exciting than a pipeline that actually works. In the lab, he studies pathomic features from donor biopsies to predict kidney transplant outcomes, and hopes his models won’t be as overfit as his sleep schedule. Outside of his work, you can find him at hackathons, experimenting with side projects, or playing soccer.
Advay Monga, Always working in harmony with others!
Undergraduate Student, University of Maryland, College Park
Email: amonga@terpmail.umd.edu
Advay is an undergraduate student at the University of Maryland, College Park. Before joining the lab, he has had some experience in computational biology and has done some personal projects in machine learning. He is interested in and enjoys learning about machine learning and math in applications to immune biology. In his free time, he likes playing badminton or volleyball with friends and trying new foods.
Alec Zhang, he thrives under pressure, even when things get a little tens(or)!
Undergraduate Student in Computer Science, University of Maryland, College Park
Email: azzhang@terpmail.umd.edu
Alec is currently a freshman studying computer science at UMD, with interests in machine learning and its applications. He is interested in applications of machine learning especially in biology and sports injuries. Outside of school, he enjoys playing soccer, trying new food places (and rating them on beli), and exploring new places.
Raymond Chen, A master of his domain!
Undergraduate Student in Computer Science, University of Maryland, College Park
Email: rchen989@terpmail.umd.edu
Raymond is a third-year student studying computer science at the University of Maryland. He is interested in all things ML, but particularly in machine vision and perception. Outside of the lab, you can find him either at the gym, outside skateboarding, or helping out at his family's restaurant! His other fun facts include that he's in a professional technology fraternity (KTP), has a terrible sleep schedule, plays a lot of TFT and League of Legends as well volleyball and badminton recreationally.
Janelle Vo, If there's a difference, she'll find it!
Undergraduate Student in Public Health Science, University of Maryland, College Park
Email: jvo9@terpmail.umd.edu
Janelle is an undergraduate student studying public health at the University of Maryland. She is particularly interested in using her background in biology to analyze and interpret biological data. She aims to create and interpret data visualizations while applying statistical methods, such as hypothesis testing, to support clinical outcomes and build her data analysis skills. In her free time, Janelle enjoys playing Roblox, dabbling in arts and crafts, traveling, and collecting trinkets!
Chris Wu, The Match Point Converter
Undergraduate Student in Computer Science, University of Maryland, College Park
Email: cwu12314@terpmail.umd.edu
Chris Wu is currently an undergraduate studying Computer Science at the University of Maryland - College Park, pursuing a concentration in Machine Learning. He enjoys learning and talking about all things related to software and ML, and you can often find him on a badminton court or on a run in his free time. He is eager to start his work at and contribute to the MACHAMP lab!
Rachel Carreras, she will survive!
Undergraduate Student in Social Data Science (Public Health Track), University of Maryland, College Park
Email: rcarrer1@terpmail.umd.edu
Rachel is an undergraduate student studying data science at the University of Maryland. She is interested in applying her data analysis experience in a public health context. She wants to refine her statistical and machine learning skills in that field. In her free time, she is likely sketching, immersed in a new book, or watching a horror film.
Jai Mohale, he's got a nice stride to his step!
Undergraduate Student in Biological Sciences and Information Science, University of Maryland, College Park
Email: ajxym12@terpmail.umd.edu
Jai is a sophomore Biology and Information Science major at UMD. With a strong interest in biomechanics and machine learning, he is eager to deepen his understanding of data analysis and machine learning as a part of the MACHAMP lab. Outside of academics, Jai loves playing volleyball, basketball, and NBA 2K.
Abhinav Karumudi
Undergraduate Student in Computer Science, University of Maryland, College Park
Email: karumudi@terpmail.umd.edu
Ajay Maloo
Undergraduate Student in Bioengineering, University of Maryland, College Park
Email: ajxym12@terpmail.umd.edu
Ajay Maloo is a rising senior studying Bioengineering. He is very interested in the exploration of how technological innovation can help revolutionize the prediction and treatment of medical conditions. At UMD Ajay is an undergraduate teaching fellow, resident assistant, and student consultant at the QUEST honors program. For fun Ajay loves to listen to new (and old) music, play the piano, and workout!
Collaborators
Ahmed Sultan, Assistant Professor, University of Maryland School of Dentistry, Director of the Division of Artificial Intelligence Research
Rose (Rong) Wang, Associate Professor, Department of Oral and Craniofacial Sciences, University of Missouri-Kansas City
Adarsh Subbaswamy, Assistant Professor of Practice, Sciences, and Health Outcomes Research, University of Maryland School of Pharmacy
Chaegeun Song, Assistant Professor of Mathematics, Bryn Mawr College
Yulin Hswen, Associate Professor, The Artificial Intelligence Interdiscliplinary Institute at Maryland (AIM), College of Computer, Mathematical, and Natural Sciences (CNMS), Department of Epidemiology and Biostatistics, School of Public Health at the University of Maryland
Jarcy Zee, Assistant Professor of Biostatistics at the University of Pennsylvania (and was my PhD adviser!)
Pinaki Sarder, Associate Professor of AI in the Section of Quantitative Health of the Department of Medicine and Associate Director for Imaging in the Intelligent Critical Care Center at the University of Florida
Anindya S. Paul, Assistant Scientist in the Computational Microscopy Imaging Laboratory and Intelligent Clinical Care Center (IC3) in the Department of Medicine at the University of Florida
Luís Rodrigues, Nephrologist at Centro Hospitalar and Universitário of Coimbra PhD Student
Robert Moy, Fellow with the Division of Nephrology at the Children's Hospital of Philadelphia
James E. Wiseman, Assistant Professor of Surgery in the Division of Trauma, Surgical Critical Care, and Acute Care Surgery, University of Maryland School of Medicine
Cher Dallal, Associate Professor of Epidemiology at the University of Maryland School of Public Health, College Park
Michael Tran, University of Maryland School of Public Health, College Park MPH in Epidemiology Graduate and Expert Champ Recruiter
John Wesley Wiggins: University of Maryland School of Public Health, College Park PhD Student in Environmental Health Sciences and Pokémon Enthusiast
Champs who have hung up the gloves/alumni
Connie Gao, The Long-Range Predictor
Medical Student, University of South Florida Morsani College of Medicine
Email: conniegao@usf.edu
Connie Gao, currently a medical student at the University of South Florida Morsani College of Medicine. She got her undergraduate degree in Biomedical Engineering at the University of Michigan and is passionate about advancing the intersection of technology and medicine!
Lylybell Zhou, Acronym and Pun Second-in-Command
Medical Student, University of South Florida Morsani College of Medicine
Email: lylybell@usf.edu
Lylybell is a second-year medical student at the University of South Florida. She graduated from the University of Florida with a degree in Medical Geography. Her research interests are wide-ranging, but she is overall interested in projects bridging her experiences in the basic, translational, and clinical sciences. She is excited to work with the MACHAMP and CMIL labs this summer (and perhaps beyond)!
Aneesh Chepuri, he can't control his excitement, but he can control his false discoveries!
MS graduate in Data Science, University of Maryland, College Park
Email: achepuri@umd.edu
Aneesh will be starting an internship as a Deep Learning Engineer at a biostatistics research organization!
Page updated
Report abuse