About

Huy Dinh is an Assistant Professor at the McArdle Laboratory for Cancer Research at the University of Wisconsin-Madison, with a joint appointment in the Department of Biostatistics and Biomedical Informatics. He completed his Ph.D. at the University of Vienna in Austria, where he developed bioinformatics methods for analyzing epigenomics data through a collaboration between the Center for Integrative Bioinformatics and the Gregor Mendel Institute in Vienna. Following his Ph.D., he pursued postdoctoral research in cancer epigenomics and intratumoral heterogeneity at the University of Southern California Cancer Center and Cedars Sinai Medical Center. Prior to joining UW-Madison as a Human Precision Medicine Cluster Hire assistant professor, he was an instructor at the La Jolla Institute for Immunology, where he studied myeloid heterogeneity and cancer immunology. His research focuses on cancer immunology using high-dimensional data profiling and computational approaches.

Research topics

• Biology
• Genetics
• Computational biology
• Evolutionary biology
• Cancer research
• Immunology
• Statistics
• Cell biology
• Mathematics
• Virology

Selected publications

• Integrated Multiomic Profiling Identifies BRD8/EP400 as a Pivotal Chromatin Module Mediating Anti-HER2 Response in HR+/HER2+ Breast Cancer

  Cancer Research · 2026-03-26

  articleOpen access

  Abstract Patients with hormone receptor–positive/HER2-positive (HR+/HER2+) breast cancer represent a historically underrecognized subgroup demonstrating poor response to combined endocrine and HER2-targeted therapies. In this study, using single-cell transcriptomic and epigenomic sequencing of estrogen receptor–positive (ER+)/HER2+ models, we identified BRD8, an acetyl-lysine reader in the EP400 histone acetyltransferase complex, as a critical mediator of ER/HER2 signaling cross-talk. BRD8 expression rapidly increased following anti-HER2 treatment, whereas its depletion disrupted ER–HER2 interaction and enhanced drug sensitivity. Single-nucleus assay for transposase-accessible chromatin using sequencing revealed that chromatin regions opening after anti-HER2 treatment were enriched for ER, FOX, and ETS transcription factor motifs, coinciding with BRD8-dependent gene activation through EP400-mediated acetylated H2AZ deposition. BRD8 regulated ER-dependent and ER-independent growth pathways, and depletion of BRD8 abolished neratinib-induced ER activation and restored drug sensitivity in resistant cells. A three-gene BRD8 signature successfully predicted anti-HER2 therapy response in two human clinical trials. Together, these findings establish BRD8 as both a predictive biomarker for anti-HER2 response and a therapeutic target to overcome resistance in HR+/HER2+ breast cancer. Significance: Multiomic single-cell profiling of breast cancer identified BRD8 as a critical mediator of resistance to HER2-targeted therapies that enables ER-HER2 cross-talk and serves as a predictive biomarker for treatment response.

  PublisherOA PDFDOI

• SR2P: an efficient stacking method to predict protein abundance from gene expression in spatial transcriptomics data

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-03-07

  articleOpen accessSenior author

  Spatial transcriptomics data are largely available with RNA expression alone, limiting the detection of cell states defined by surface protein abundance. The lack of multi-omics spatial data limits the ability to identify immune cells and their signaling in the tumor microenvironment, as most solid tumors are immunologically poor and exhibit protein-RNA abundance discordance in critical immune cell surface markers. Although emerging technologies enable spatial multi-omics profiling, technical and cost constraints remain a hurdle. We introduce SR2P, a stacking-based machine-learning framework for predicting spatial protein abundance from RNA expression. SR2P integrates 11 complementary predictive models and consistently outperforms existing methods across multiple spatial multi-omics benchmark. We showcased an application of SR2P recovered macrophage-enriched regions and identified potential immune markers associated with therapeutic response from head-and-neck squamous cell carcinoma patients. SR2P enables protein-abundance inference from RNA-only spatial data, extending the analytical capabilities of current spatial platforms for studies of tumor immunology.

  PublisherOA PDFDOI

• Abstract 510: Early metabolic and differentiation remodeling in <i>BRCA1/2</i> high-risk fallopian tube epithelium revealed by single-cell multi-omics.

  Cancer Research · 2026-04-03

  article

  Abstract Germline mutations in BRCA1 and BRCA2 greatly increase the risk of developing high-grade serous carcinoma (HGSC), yet the earliest molecular events driving malignant transformation in the fallopian tube epithelium, the likely site of origin, remain poorly understood. Defining these early alterations is essential for improving risk prediction and informing preventive strategies for BRCA-associated ovarian cancer.To investigate the initial cellular and molecular perturbations in high-risk individuals, we performed single-cell transcriptomic and multi-omic profiling on fallopian tube fimbrial samples obtained through exfoliative cytology brushings and conventional tissue specimens. Single-cell RNA sequencing was conducted on 14 brushings from BRCA1/2 mutation carriers, 3 brushings from average-risk individuals, and 12 tissue samples from average-risk controls. In parallel, single-cell multi-omic analysis was applied to short-term epithelial cultures derived from 4 high-risk and 2 average-risk individuals, enabling an integrated assessment of transcriptional states, chromatin accessibility, and gene regulatory programs.Brushings were enriched for epithelial and immune populations, facilitating high-resolution characterization of epithelial differentiation. Cells from BRCA1/2 mutation carriers exhibited disrupted secretory-ciliated differentiation trajectories, accompanied by upregulation of mitochondrial respiration and oxidative phosphorylation genes, suggesting early mitochondrial and metabolic remodeling in high-risk epithelia. Multi-omic integration further identified a distinct epithelial cluster enriched in BRCA1/2 carriers characterized by increased RUNX3 transcription factor activity. Although these cells remain transcriptionally aligned with secretory epithelium, RUNX3-associated programs indicate a partial or intermediate differentiation state—potentially reflecting early lineage instability preceding malignant transformation. Subtle alterations in immune-related pathways were also observed, pointing to microenvironmental changes that may support a permissive niche for tumor initiation.Together, these data reveal early transcriptomic, metabolic, and gene-regulatory remodeling in the fallopian tube epithelium of BRCA mutation carriers. By defining epithelial states and pathways perturbed prior to neoplasia, this work provides new mechanistic insight into hereditary ovarian cancer predisposition and highlights potential biomarkers for early detection and prevention. Citation Format: Quentin Chartreux, Marcela Haro, Josh Brand, Andrew Li, Bobbie J. Rimel, Patrick Sung, Simon Gayther, Huy Dinh, Fabiola Medeiros, Kate Lawrenson. Early metabolic and differentiation remodeling in BRCA1/2 high-risk fallopian tube epithelium revealed by single-cell multi-omics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 510.

  PublisherDOI

• Abstract 375: BRD8 chromatin regulator mediates anti-HER2 response in HR+/HER2+ breast cancer.

  Cancer Research · 2026-04-03

  article

  Abstract Chromatin remodeling complexes are critical regulators of transcription and therapeutic resistance in cancer, yet the mechanisms governing their recruitment and activity remain poorly understood. Here, we identify BRD8, a largely uncharacterized bromodomain protein within the EP400 chromatin remodeling complex, as a master regulator of treatment-induced transcriptional reprogramming in hormone receptor-positive/HER2-positive (HR+/HER2+) breast cancer. Using integrative multi-omics approaches—including paired single-cell RNA-seq and ATAC-seq, comprehensive ChIP-seq profiling, and functional genomics—we uncover a novel epigenetic mechanism whereby BRD8 orchestrates chromatin accessibility and transcription factor cooperation to drive therapeutic resistance. BRD8 depletion reduces H2A.Z acetylation and disrupts interactions between the EP400 complex and key transcription factors. Through both ER-dependent and ER-independent pathways, BRD8 cooperates with H2A.Z and acetylated H2A.Z (H2A.Zac) to facilitate recruitment of the EP400 complex to transcription factors including ER, FOXA1, and ETS family members, thereby promoting H2A.Z acetylation, chromatin accessibility, and expression of growth-promoting genes. BRD8 expression is markedly elevated in anti-HER2-resistant models, and its depletion sensitizes resistant cells to anti-HER2 therapy. To translate these findings therapeutically, we developed a potent and selective BRD8 PROTAC degrader with a DC50 of ∼100 nM. Combined treatment with the BRD8 degrader and anti-HER2 agents synergistically inhibits HR+/HER2+ breast cancer cell growth with superior efficacy and no detectable toxicity. Notably, the BRD8 degrader demonstrates fivefold greater potency than the BRD8 inhibitor DN02 across multiple HR+/HER2+ breast cancer cell lines. Finally, we developed a 3-gene BRD8 signature that predicts anti-HER2 therapy response in two independent clinical trials, establishing BRD8 as both a predictive biomarker and a central epigenetic regulator of therapeutic resistance. Our findings identify dual BRD8 and HER2 targeting as a promising precision medicine strategy for overcoming resistance in HR+/HER2+ breast cancer. Citation Format: Wei Xu, Ang Gao, Parth Khatri, Gui Ma, Huy Dinh, Charles Perou. BRD8 chromatin regulator mediates anti-HER2 response in HR+/HER2+ breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 375.

  PublisherDOI

• Integrated Single-Cell and Spatial Analysis Reveals Context-Dependent Myeloid–T Cell Interactions in Response to Immune Checkpoint Blockade in Head and Neck Cancer

  Clinical Cancer Research · 2026-03-16

  articleOpen accessSenior author

  PURPOSE: We conduct a systematic evaluation of cell-cell interactions between tumor-infiltrating immune cells in patients with head and neck cancer squamous cell carcinoma (HNSCC) who have been treated with immune checkpoint blockade (ICB) using spatial and single-cell omics data. EXPERIMENTAL DESIGN: We employ complementary techniques from both Visium spot-based spatial transcriptomics and CosMx SMI single-cell spatial omics, utilizing a 64-plex protein panel and a 1000-gene RNA panel, which includes 435 ligands and receptors. We conducted integrated bioinformatics analyses to identify cellular neighborhoods of co-localizing cell types and Ligand-Receptor interactions across different single-cell and spatial data modalities. RESULTS: With 522,399 single cells profiled for both RNA and protein from 23 patients, along with spot-resolved spatial RNA-seq data from 8 patients treated with ICB, and through bioinformatics analysis of publicly available single-cell and bulk RNA-seq, we identified a spatial and cell-type specific context dependency in the differences of myeloid and T cell interactions between Responders and Non-Responders samples. We further defined the cellular neighborhood and sources of chemokine CXCL9/10-CXCR3 interactions, emphasizing the specificity of this marker in Responders samples, an emerging target in ICB, as well as other underappreciated markers and targets for ICB response in HNSCC, such as CXCL16-CXCR6, CCL4/5-CCR5. CONCLUSIONS: We have provided a valuable resource for analyzing spatial and cell-cell ligand-receptor interactions, including the cellular and spatial contexts of ICB response markers. Our data suggest that future mechanistic studies should consider this context specificity when evaluating ICB response biomarkers and targets.

  PublisherOA PDFDOI

• Data from: Integrated single-cell and spatial analysis reveals context-dependent myeloid-T cell interactions in the response to immune checkpoint blockade in head and neck squamous cell carcinoma

  Open MIND · 2026-02-17

  dataset1st authorCorresponding

  Purpose: We conduct a systematic evaluation of cell-cell interactions between tumor-infiltrating immune cells in patients with head and neck cancer squamous cell carcinoma (HNSCC) who have been treated with immune checkpoint blockade (ICB) using spatial and single-cell omics data. Experimental Design: We employ complementary techniques from both Visium spot-based spatial transcriptomics and CosMx SMI single-cell spatial omics, utilizing a 64-plex protein panel and a 1000-gene RNA panel, which includes 435 ligands and receptors. We conducted integrated bioinformatics analyses to identify cellular neighborhoods of co-localizing cell types and Ligand-Receptor interactions across different single-cell and spatial data modalities. Results: With 522,399 single cells profiled for both RNA and protein from 23 patients, along with spot-resolved spatial RNA-seq data from 8 patients treated with ICB, and through bioinformatics analysis of publicly available single-cell and bulk RNA-seq, we identified a spatial and cell-type specific context dependency in the differences of myeloid and T cell interactions between Responders and Non-Responders samples. We further defined the cellular neighborhood and sources of chemokine CXCL9/10-CXCR3 interactions, emphasizing the specificity of this marker in Responders samples, an emerging target in ICB, as well as other underappreciated markers and targets for ICB response in HNSCC, such as CXCL16-CXCR6, CCL4/5-CCR5. Conclusions: We have provided a valuable resource for analyzing spatial and cell-cell ligand-receptor interactions, including the cellular and spatial contexts of ICB response markers. Our data suggest that future mechanistic studies should consider this context specificity when evaluating ICB response biomarkers and targets.

  DOI

• Fluoro-forest: a random forest workflow for cell type annotation in high-dimensional immunofluorescence imaging with limited training data

  Bioinformatics Advances · 2025-12-22 · 1 citations

  articleOpen accessSenior author

  Motivation: Cyclic immunofluorescence (IF) techniques enable deep phenotyping of cells and help quantify tissue organization at high resolution. Due to its high dimensionality, workflows typically rely on unsupervised clustering, followed by cell type annotation at a cluster level for cell type assignment. Most of these methods use marker expression averages that lack a statistical evaluation of cell type annotations, which can result in misclassification. Here, we propose a strategy through an end-to-end pipeline using a semi-supervised, random forest approach to predict cell type annotations. Results: Our method includes cluster-based sampling for training data, cell type prediction, and downstream visualization for interpretability of cell annotation that ultimately improves classification results. We show that our workflow can annotate cells more accurately compared to representative deep learning and probabilistic methods, with a training set <5% of the total number of cells tested. In addition, our pipeline outputs cell type probabilities and model performance metrics for users to decide if it could boost their existing clustering-based workflow results for complex IF data. Availability and implementation: Fluoro-forest is freely available on GitHub under an MIT license (https://github.com/Josh-Brand/Fluoro-forest).

  PublisherDOI

• 630 Anti-CD40 and epigenetic modifier inhibitors to augment treatment of high-risk neuroblastoma

  Regular and Young Investigator Award Abstracts · 2025-11-01

  articleOpen access
  PublisherOA PDFDOI

• Identification of anti-tumoral neutrophil states from pan-cancer single-cell transcriptome analysis of immunotherapies in preclinical mouse models and humans 3571

  The Journal of Immunology · 2025-11-01

  articleOpen access1st authorCorresponding

  Abstract Description Neutrophils, the most abundant immune cells, have multifaceted functions in cancers. While most previous work reported pro-tumoral neutrophils, emerging data have shown anti-tumoral phenotypes directly killing tumor cells and activating cytotoxic T-cell immunity in different immunotherapy contexts. Therefore, there is a need to define anti-tumoral neutrophil markers and potential therapeutic targets. Here, we interrogate anti-tumoral phenotypes using bioinformatics analysis of ∼95,000 neutrophils from scRNA-Seq of 8 mouse models with different treatments, including Immune Checkpoint Blockade (ICB) and CD40 agonist. In addition, we generated multi-omics single-cells of neutrophils from immunological cold models using a combined neutrophil activation therapy. We found consistent markers of interferon-stimulated neutrophil phenotypes in multiple cancer types and treatments. The compendium of pan-cancer pan-treatment data helps identify potential markers of neutrophils with tumor-killing phenotypes found in microbial stimulation and tumor gene knock-out settings. Furthermore, our new scRNA-Seq data determined a subset associated with TNF and migration phenotypes. Lastly, cross-species analysis with data across 17 cancer types, including those treated with ICB, determined similar signatures in humans. Our results report the emerging phenotypes in anti-tumoral neutrophils revealed by single-cell omics, providing a foundation for future mechanistic and translational studies. Topic Categories Computational and Systems Immunology (COMP)

  PublisherOA PDFDOI

• Abstract 4308: BRD8 is a therapeutic vulnerability for overcoming resistance to dual ER/HER2 blockade therapy in HR+/HER2+ breast cancer

  Cancer Research · 2025-04-21

  article

  Abstract Hormone receptor (HR)-positive, HER2-positive breast cancers often develop resistance to endocrine and anti-HER2 therapies due to the heterogeneous expression of estrogen receptor (ER) and HER2 and the crosstalk of these growth-promoting pathways. However, how anti-HER2 agents activate ER and other growth-promoting pathways remains unknown. Single-cell RNA sequencing of BT474 breast cancer cells identified Bromodomain Containing Protein 8 (BRD8), an acetyl-lysine reader protein in the histone acetylase EP400 complex, as a pivotal mediator to activate ER in response to neratinib treatment. BRD8 expression was rapidly induced by various anti-HER2 agents (neratinib, lapatinib, and trastuzumab), and its depletion disrupted the crosstalk between ER and HER2 signaling pathways and rendered HR+/HER2+ cells and PDxOs more sensitive to anti-HER2 agents. BRD8, ER, and ER target genes are co-induced by neratinib in single-nucleus RNA and ATAC sequencing of a patient-derived xenograft (PDX). SnATAC-seq also reveals that the activated genes share open chromatin regions enriched in ER, forkhead box (FOX), and ETS family transcription factors (TF) binding motifs. FOX family TFs are well-known for regulating estrogen signaling, and ETS proteins have been shown to promote tumorigenesis. Since EP400 enhances H2AZ deposition and acetylation on chromatin, we performed H2AZ and H2AZac ChIP-sequencing in the presence or absence of BRD8 and neratinib treatment. We found that, in response to neratinib treatment, BRD8 activates ER, FOX, and ETS target genes through modulating H2AZac deposition and chromatin decompaction. This finding coincides with RNA-sequencing where BRD8 promotes cell growth in an ER-dependent and -independent manner. In line with these findings, patients who responded poorly to the anti-HER2 therapies exhibited higher levels of BRD8 target gene signature as compared to the responders. Furthermore, BRD8 knockout ablates the ER and HER2 signaling crosstalk and re-sensitizes neratinib-resistant HR+/HER2+ cells to neratinib. In summary, this work not only explains why ER signaling is activated upon anti-HER2 therapies but also identifies BRD8 as a druggable vulnerability for treating HR+/HER2+ breast cancer. Citation Format: Ang Gao, Parth Khatri, Gui Ma, Peng Liu, Mark Burkard, Kari Wisinski, Charles Perou, Huy Dinh, Wei Xu. BRD8 is a therapeutic vulnerability for overcoming resistance to dual ER/HER2 blockade therapy in HR+/HER2+ breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4308.

  PublisherDOI

Frequent coauthors

• Rory Johnson

  University Hospital of Bern

  94 shared

• Roland Eils

  84 shared

• Thomas J. Mitchell

  Wellcome Sanger Institute

  82 shared

• Lars Feuerbach

  German Cancer Research Center

  76 shared

• L. Sylvia

  Mirai Hospital

  73 shared

• Geoff Macintyre

  Spanish National Cancer Research Centre

  72 shared

• Keiran Raine

  Wellcome Sanger Institute

  70 shared

• Taobo Hu

  Peking University People's Hospital

  70 shared

Labs

• DINH LABORATORYPI

Education

• Ph.D., Biostatistics

  University of Wisconsin–Madison

  2004

• M.S., Biostatistics

  University of Wisconsin–Madison

  2001

• B.S., Mathematics

  University of Wisconsin–Madison

  1998