Learn more →
Back to Expert Scholars
research / researchNMF matrix factorization, multi-omics AI, cancer systems biology, scRNA-seq, spatial omics, CoGAPSBispecific T-cell Engager Pioneer

Elana J. Fertig

伊拉娜·费尔蒂格

PhD

🏢Johns Hopkins University(约翰斯·霍普金斯大学)🌐USA

Professor of Oncology; Director of Convergence Institute; Co-Director of the Convergence Institute for Cancer Informatics, Johns Hopkins University School of Medicine肿瘤学教授;融合研究所所长;约翰斯·霍普金斯大学医学院癌症信息学融合研究所联合主任

38
h-index
4
Key Papers
5
Awards
4
Key Contributions

👥Biography 个人简介

Elana J. Fertig, PhD is Professor of Oncology and Director of the Convergence Institute at Johns Hopkins University School of Medicine, where she leads a research program integrating computational mathematics, machine learning, and multi-omics biology to understand cancer progression and treatment resistance. Dr. Fertig is best known for developing CoGAPS (Coordinated Gene Activity in Pattern Sets), a Bayesian non-negative matrix factorization algorithm that decomposes multi-omics datasets (bulk RNA-seq, scRNA-seq, ATAC-seq, spatial transcriptomics) into biologically interpretable transcriptional programs, widely used to identify cancer-associated gene activity patterns across patient cohorts and single cells. Her ProjectR framework enables transfer learning of NMF-derived patterns across datasets and modalities, allowing computational transfer of biological programs from training cohorts to new datasets, enabling cross-study generalization of cancer molecular signatures. Dr. Fertig's laboratory has made major contributions to understanding treatment resistance in head and neck, colorectal, and pancreatic cancers using multi-omics single-cell approaches, identifying transcriptional programs that emerge under therapeutic selection. She co-directs the JHU Convergence Institute for Cancer Informatics and leads multi-institutional computational oncology consortia integrating multi-omics data from clinical trials. Dr. Fertig is a computational biologist whose mathematical foundations and cross-disciplinary training make her work uniquely rigorous at the interface of AI, genomics, and cancer biology.

Share:

🧪Research Fields 研究领域

Non-negative Matrix Factorization (NMF) for Multi-Omics Cancer Analysis — CoGAPS Algorithm多组学癌症分析的非负矩阵分解(NMF)——CoGAPS算法
Single-Cell and Spatial Omics AI — Computational Methods for scRNA-seq and Spatial Transcriptomics单细胞和空间组学AI——scRNA-seq和空间转录组学的计算方法
Cancer Systems Biology — Network-Based AI for Tumor Microenvironment Characterization癌症系统生物学——用于肿瘤微环境表征的基于网络的AI
Treatment Resistance AI — Identifying Transcriptional Programs Driving Therapy Resistance治疗耐药AI——识别驱动治疗耐药的转录程序
Transfer Learning Across Omics Modalities — ProjectR and Cross-Study Pattern Generalization跨组学模态迁移学习——ProjectR和跨研究模式泛化

🎓Key Contributions 主要贡献

CoGAPS — Bayesian NMF for Multi-Omics Cancer Pattern Discovery

Developed and maintains CoGAPS (Coordinated Gene Activity in Pattern Sets), a widely-used Bayesian non-negative matrix factorization algorithm for decomposing multi-omics datasets into interpretable transcriptional activity patterns. CoGAPS identifies latent biological programs in bulk RNA-seq, scRNA-seq, and ATAC-seq data that correspond to cell state transitions, pathway activities, and microenvironmental interactions in cancer, and is distributed as an open-source R/Bioconductor package with tens of thousands of downloads. It has been applied across cancer types and therapeutic contexts to identify treatment-predictive transcriptional signatures and microenvironmental programs.

ProjectR — Transfer Learning for Cross-Omics and Cross-Cohort Pattern Generalization

Developed ProjectR, a transfer learning framework that projects NMF-learned biological patterns from reference datasets onto new omics data across different modalities (bulk to single-cell, RNA-seq to ATAC-seq, human to mouse), enabling computational knowledge transfer of cancer biology insights across studies. Published in Bioinformatics with widespread adoption, ProjectR enables cross-study generalization of tumor transcriptional programs learned from one clinical cohort to new patient datasets, and has been applied to transfer bulk RNA-seq resistance signatures to single-cell and spatial transcriptomics analyses.

Single-Cell Multi-Omics of Treatment Resistance in Head and Neck and GI Cancers

Applied CoGAPS, scRNA-seq, and integrated multi-omics to characterize transcriptional plasticity and resistance mechanisms in head and neck squamous cell carcinoma, colorectal cancer, and pancreatic cancer under immunotherapy, targeted therapy, and radiation treatment. Identified epithelial plasticity programs, EMT-like transcriptional states, and immunosuppressive microenvironmental programs that emerge under therapeutic selection pressure, providing molecular targets and biomarkers for overcoming treatment resistance in multi-institutional clinical datasets.

Spatial Transcriptomics AI — Tumor Microenvironment Spatial Organization

Extended NMF and transfer learning methods to spatial transcriptomics platforms (Visium, Slide-seq, MERFISH), developing computational frameworks for identifying spatially organized gene expression programs in tumor microenvironments. Demonstrated that spatial co-localization patterns of cancer cell states and immune infiltrates revealed by AI methods predict response to immunotherapy and targeted therapy, and published methodological advances in spatial omics integration that enable joint analysis of spatial gene expression and protein marker data.

Representative Works 代表性著作

[1]

CoGAPS 3: Bayesian Non-Negative Matrix Factorization for Single-Cell Analysis with Asynchronous GPU Processing

Bioinformatics (2020)

Updated CoGAPS implementation with GPU acceleration and single-cell optimizations for large-scale scRNA-seq pattern discovery, enabling application to datasets with millions of cells.

[2]

Determining Gene Expression Patterns from Single-Cell Transcriptomics Data with Non-Negative Matrix Factorization

Genome Biology (2017)

Demonstration of NMF-based transcriptional program discovery in single-cell RNA-seq data for cancer cell state characterization across multiple tumor types.

[3]

Integrated Analysis Reveals Multi-scale Transcriptional Heterogeneity with Novel Prognostic and Predictive Value for Head and Neck Cancer

Cancer Research (2021)

Multi-scale NMF analysis of HNSCC bulk and single-cell RNA-seq identifying transcriptional programs predictive of immunotherapy response and survival.

[4]

Uncovering Axes of Variation Among Single-Cell Cancer Specimens with UINMF

Nature Methods (2022)

UINMF framework for multi-modal single-cell data integration using NMF, enabling joint analysis of scRNA-seq, scATAC-seq, and spatial transcriptomics cancer datasets.

🏆Awards & Recognition 奖项与荣誉

🏆NIH NCI Outstanding Investigator Award (R35)
🏆ASCO Conquer Cancer Foundation Career Development Award
🏆Johns Hopkins Provost Award for Excellence in Mentoring
🏆AACR Women in Cancer Research Award
🏆American Mathematical Society Mathematics in Medicine Award

📄Data Sources 数据来源

Last updated: 2026-04-06 | All information from publicly available academic sources

关注 伊拉娜·费尔蒂格 的研究动态

Follow Elana J. Fertig's research updates

留下邮箱,当我们发布与 Elana J. Fertig(Johns Hopkins University)相关的新研究或访谈时,我们会通知你。

我们不会泄露你的信息,也不会发送无关内容。随时可以退订。

Explore More Experts

Discover the researchers shaping the future of cancer treatment