Adithya Madduri

I am a student at Harvard College studying Molecular & Cellular Biology and Statistics, with a secondary in Government. I work on machine learning for genomics and neuroscience. I am advised by Professor Chirag Patel at Harvard Medical School, and work with Professor Bernardo Sabatini at Harvard Medical School and Professor Nídia Trovao at the NIH Fogarty International Center.

Email / Scholar / Github

Research

I work at the intersection of machine learning and biomedicine, with a focus on single-cell genomics and neurodegenerative disease. I am interested in how we can use high-dimensional genomic data to identify disease mechanisms and therapeutic targets.

	Single-cell machine learning uncovers genetically anchored, cell-type specific programs of Alzheimer's disease Adithya V. Madduri, Randall Ellis, Chirag M. Lakhani, David A. Bennett, Chirag J. Patel medRxiv, 2026 medRxiv Cell-type-resolved machine learning on single-cell RNA-seq data from the prefrontal cortex uncovers genetically anchored programs driving Alzheimer's disease.
	Protein Compositional Ratio Representation (PCRR) Systematically Improves Human Disease Prediction Adithya V. Madduri, Randall J. Ellis, Chirag J. Patel International Conference on Research in Computational Molecular Biology (RECOMB), 2026 bioRxiv A protein representation method based on compositional ratios that systematically improves prediction of human diseases.
	Selection of Genetic Conditions for Multi-State Genomic Newborn Screening in BEACONS-NBS Nina B. Gold, Britt A. Johnson, Harini Somanchi, ..., Adithya V. Madduri, ..., Robert C. Green medRxiv, 2026 medRxiv A multi-state study integrating whole genome sequencing into newborn screening; validated a condition list through expert consensus and community input.
	Sparse Control of Disease-Aligned Gene Programs in Single-Cell Transcriptomics Adithya Madduri, Chirag J. Patel ICLR 2026 Workshop on Generative AI in Genomics & ICLR 2026 Workshop on Machine Learning for Genomics Explorations OpenReview Sparse manifold control methods applied to single-cell transcriptomics identify donor-robust perturbation targets and disease-aligned gene programs.
	Cell-type–Resolved Machine Learning Reveals Excitatory Neuron Perturbations in Parkinson's Disease Prefrontal Cortex Adithya V. Madduri, Chirag J. Patel 20th International Conference on Alzheimer's & Parkinson's Diseases, 2026 abstract Single-cell machine learning identifies excitatory neuron-specific perturbations in Parkinson's disease prefrontal cortex, with proteostasis and excitatory neuron vulnerability as key findings.
	A paradigm for skilled forelimb reaching by head-restrained mice Annabel Chang, Adithya Madduri, Janet Berrios Wallace, Junzheng Wang, Kimberly Reinhold bioRxiv, 2026 bioRxiv A complete pipeline for head-restrained forelimb reaching in mice, with a customizable pellet-presentation rig, training protocol, and analysis code to detect and classify reaches. Applied to reach interruption, it resolves fine phase-dependent features of skilled motor behavior.
	Integrating Gene Expression, Pseudotime, and Imputed SNPs to Predict Alzheimer's Disease Across Major Brain Cell Types Using Machine Learning Adithya V. Madduri, Randall J. Ellis, Chirag J. Patel The American Society of Human Genetics Annual Meeting, 2025 abstract Integrates gene expression, pseudotime trajectories, and imputed SNPs to build cell-type-specific predictive models for Alzheimer's disease across major brain cell types.
	The role of tomato wild relatives in breeding disease-free varieties Hamid Khazaei, Adithya Madduri Genetic Resources, 3(6), 64–73, 2022 paper Reviews the role of wild tomato relatives as sources of disease resistance genes for cultivating improved varieties.

Modified from Jon Barron.

Recent News

Research