Introduction
The AI in Genomics Lab is a cross-disciplinary research group which aims to transcend the disciplinary boundary of genomics. We integrate data science, machine learning, and biology approaches to decipher disease mechanisms by finding cell morphologies and molecular profiles associated with disease progression. Such knowledge is demanded for precision medicine and the therapeutic development of small molecules and their delivery to target regions.
This research lab is directed by Dr. Jian Hu, Assistant Professor of Human Genetics at Emory School of Medicine.
Research
Developing machine learning methods for spatial multi-omics integration
Multi-modal spatial omics represents a paradigm shift in understanding complex biological systems by integrating diverse omics modalities within their native tissue contexts. Different spatial omics techniques capture distinct molecular information, including transcriptomics, proteomics, metabolomics, chromatin accessibility, and histone modifications. Together, these modalities provide a comprehensive view of cellular and tissue functions, which is crucial for unraveling the molecular mechanisms of diseases. One focus of my lab is to develop state-of-the-art machine learning analytical tools to help researchers better mine the rich information in spatial multi-omics data.
Developing statistical tools for single-cell RNA sequencing analysis
The advent and rapid development of single-cell technologies have made it possible to study cellular heterogeneity at an unprecedented resolution and scale. Cellular heterogeneity underlies phenotypic differences among individuals, and studying cellular heterogeneity is an important step toward our understanding of disease molecular mechanisms. Single-cell technologies offer opportunities to characterize cellular heterogeneity from different angles, but linking cellular heterogeneity with disease phenotypes requires careful computational analysis.
Developing AI-driven analytical tools for digital pathology imaging data
The adoption of digital pathology has enabled the curation of large repositories of gigapixel whole-slide images (WSIs), which are invaluable for examining cellular morphology and its changes during embryonic development or disease progression. One focus of my lab is to develop label-free, AI-driven methods for medical imaging data analysis. These tools can be easily applied to studies with limited training samples, eliminating the need for cumbersome labeling steps.
News
1/15/2025: New Nature Methods Paper on multi-omics integration!
Excited to share our latest #genomics work published in Nature Methods. MISO (MultI-modal Spatial Omics) is a versatile algorithm for feature extraction and clustering, capable of integrating multiple modalities from …
09/03/2024: Hu Invited to Speak at UNC Biostatistics Seminar
Dr. Hu presented on multi-layered applications of AI for detailed study of tumor microenvironments. Link …
08/25/2024: New Nature Communications Paper on Deep Profiling of Tumor Ecosystems
Excited to share our new article on building and evaluating METI, an end-to-end framework for deep profiling of tumor ecosystems using spatial transcriptomics. We’ve built on our previous success in …