AlphaFold-Assisted Affinity Probe Resource for Scalable Brain Protein Mapping with Reference Datasets

About This Grant

PROJECT SUMMARY Despite the broad use of antibody tools, their reliability and scalability remain major challenges for reproducible cellular and anatomical profiling in large tissue volumes, such as whole-mouse and human brains. In this project, we propose to develop an AI-assisted platform for recombinant antibody resources that meets the high standards required to advance brain research. By integrating our teams' complementary expertise in AI algorithms, structural analysis, antibody validation, and brain imaging, we aim to create a synergistic technical and resource platform for curating recombinant antibody tools, establishing a new paradigm for antibody applications in brain research. As part of the BICCN/BICAN effort, we have validated an extensive collection of commercial and open- source monoclonal antibodies using a high-content screening pipeline, providing a strong experimental foundation for recombinant antibody curation. However, the current experimental approach has limitations in efficiently selecting and prioritizing thousands of monoclonal antibody sequences for cost-effective conversion and validation. It also relies solely on existing sequences without effective optimization or design using structural information for diverse applications. Our integrated approach, leveraging our expertise in AI-based protein structure prediction, modeling, and simulation, will analyze antibody sequences to provide a comprehensive understanding of recombinant antibody structural properties, including antigen binding sites and affinities across target species. This will significantly accelerate the conversion and validation of recombinant antibodies. Additionally, we will continue refining our approach to enhance flexibility in optimizing and designing recombinant antibodies, tailoring them to specific epitopes and cross-species targets for diverse applications. We will also develop a curated database of recombinant antibody resource for easy reference and adaptation within the field. This database will include a comprehensive collection of validated recombinant antibodies, featuring their defined sequences, predicted antibody-antigen binding structures, 2D IHC data in mouse and human, and 3D whole mouse brain labeling datasets for key BICCN targets. To ensure broad accessibility, the database will be available through a user-friendly online portal. This integrated recombinant antibody platform, rAb-GenAI, will be open to incorporating advanced AI algorithms and continuously refined through iterative experimental feedback. Both our AI pipeline and recombinant antibody resource will be highly scalable and adaptable, maximizing cost efficiency to support consistent large-scale profiling, including human brain cohort mapping. Through this project, we aim to establish a new paradigm for antibody-based research, laying the foundation for brain-wide proteomic investigations across species and enabling studies on whole brain cellular and structural dynamics in health and disease.