High-throughput and Machine Learning Optimization of Fluorescent Sensors for Multiplexed Imaging in vivo

/ ABSTRACT Monitoring neuronal activity modulation is pivotal for elucidating brain functionality and addressing neurological disorders. Despite the advancements brought by green fluorescent calcium indicators like GCaMP and neuromodulator sensors, a considerable gap persists in the development of red fluorescent sensors that match the properties of their green counterparts. This gap, characterized by limitations in dynamic range, photostability, and kinetics, restricts a more comprehensive exploration of neuronal interactions, especially in multiplexed, dual-imaging imaging scenarios. Additionally, the iterative engineering approach for new sensor development is notoriously slow and labor-intensive. Our central goal is to leverage our sensor screening platform, Opto-MASS, as well as our recent successes in using machine learning to expedite the optimization of fluorescent sensors. This project aims to engineer red fluorescent calcium and neuromodulator sensors that match the kinetics and dynamic range of green sensors and further enhance their properties. Our objectives include the rigorous benchmarking of these sensors against the best-in-class for properties such as dynamic range, kinetics, and photostability, followed by comprehensive in vivo validation across multiple laboratories and application scenarios using fiber photometry and two and 3-photon imaging. Our project is innovative because it utilizes a high-throughput screening assay capable of evaluating over 10,000 sensor variants from library collections in under an hour, a significant advancement over current methods. Coupled with pioneering machine learning models that identify key residues affecting sensor performance, we will significantly accelerate fluorescent sensor development, particularly for red calcium and GPCR-based sensors. Importantly, we aim to achieve these goals while reducing time and resource commitments. Our project directly addresses critical needs outlined in this FOA, including a broader range of reliable sensors in neuroscience research that facilitate nuanced, multidimensional studies of brain activity. By developing sensors with improved dynamic ranges, kinetics, and photostability, we aim to overcome existing barriers to multiplexed imaging of neuronal dynamics in vivo. Ultimately, the successful completion of this project would not only fill a vital gap in neuroscientific research tools but also align with the NIH BRAIN Initiative's objectives to advance neurotechnology and set new standards for molecular tool development and in vivo validation in neuroscience. Project Number: 1UM1MH138543-01A1 | Fiscal Year: 2026 | NIH Institute/Center: National Institute of Mental Health (NIMH) | Principal Investigator: Andre Berndt (+3 co-PIs) | Institution: UNIVERSITY OF WASHINGTON, SEATTLE, WA | Award Amount: $2,305,031 | Activity Code: UM1 | Study Section: Special Emphasis Panel[ZMH1 ERB-S (04)] View on NIH RePORTER: https://reporter.nih.gov/project-details/11216452