High-dimensional imaging using combinatorial channel multiplexing and deep learning.

Area of Science:

• Biotechnology
• Molecular Imaging
• Cell Biology

Background:

• Accurate tissue structure and function analysis requires quantifying multiple proteins at single-cell resolution with spatial information.
• Current imaging methods are limited in throughput and scalability due to single-channel protein detection.

Purpose of the Study:

• To introduce Combinatorial Multiplexing (CombPlex), a novel platform to significantly enhance the number of proteins measurable by imaging modalities.
• To demonstrate CombPlex's ability to decompress agglomerated protein images into individual protein data using deep learning.

Main Methods:

• CombPlex employs a combinatorial staining strategy where multiple proteins are imaged within fewer channels.
• A deep learning algorithmic framework is utilized to decompress these combined signals into distinct protein expression data.
• The method was validated using both fluorescence and mass-based imaging techniques across diverse tissue types.

Main Results:

• Accurate reconstruction of 22 proteins compressed into five imaging channels was achieved.
• Successful application of CombPlex was demonstrated in various tissues and cancer types.
• The platform significantly increases protein quantification capacity without requiring specialized instrumentation.

Conclusions:

• CombPlex offers a scalable solution for high-content protein analysis in biological and clinical research.
• This technology can be integrated with existing imaging modalities to expand multiplexing capabilities.
• CombPlex advances single-cell spatial proteomics, enabling deeper understanding of tissue structure and function.