Divij Mathew

Deep learning representations of human Immune Health for precision immunology

The paper develops MAESTRO, a self-supervised deep learning framework to transform immune profiles into embeddings capturing immune health, enabling clinical prediction and precision immunotherapy.

PD-1 blockade drug holiday improves exhausted progenitor CD8 T cell (Tpex) reinvigoration by avoiding Tpex adaptive resistance

The paper shows that a drug holiday during PD-1 blockade improves reinvigoration of exhausted T cells by restoring progenitor cell function and reducing adaptive resistance.

High-Throughput De Novo Protein Design Yields Novel Immunomodulatory Agonists

The paper designs novel cytokines by fusing computationally generated receptor-binding domains, identifying several that activate signaling pathways and have specific biological effects.

Geometric Tuning of Cytokine Receptor Association Modulates Synthetic Agonist Signaling

This paper designs synthetic cytokine agonists with controlled receptor geometry to modulate signaling outcomes and functional effects.

Cleanet: robust doublet detection in cytometry data based on protein expression patterns

Cleanet detects and classifies doublets in cytometry data using protein expression patterns, improving accuracy over traditional bivariate gating methods.

Deciphering Cell Fate and Clonal Dynamics via Integrative Single-Cell Lineage Modeling

Clonotrace integrates gene expression and clonotype data to infer cell state transitions and fate biases, improving understanding of lineage dynamics in various biological systems.

Divergent ontogeny of Tissue Resident Memory and Tissue Resident Exhausted CD8+ T cells underlies distinct functional potential

The paper identifies TR-TEX cells as distinct from TRM cells, showing they have unique transcriptional regulation and respond differently to PD-1 inhibition.

Automated Cytometric Gating with Human-Level Performance Using Bivariate Segmentation

This paper introduces UNITO, a deep learning framework for automated cytometric gating that matches human performance and provides interpretable results without requiring manual input.

Cytometry Masked Autoencoder: An Accurate and Interpretable Automated Immunophenotyper

The paper introduces cyMAE, a masked autoencoder for automated immunophenotyping, showing it improves accuracy and interpretability over manual methods in cytometry data analysis.

Deep Phenotyping of the Lipidomic Response in COVID and non-COVID Sepsis

The paper identifies lipidomic biomarkers and therapeutic targets in sepsis patients, including those with COVID-19, and develops an interactive network analysis tool for multiomic data exploration.

Signal recovery in single cell batch integration

This paper introduces CellANOVA, a statistical model to recover biological signals lost during single-cell data integration by distinguishing batch effects from true biological variation.