Doubling multiplexed imaging capability via spatial expression pattern-guided protein pairing and computational unmixing

Area of Science:

• Neuroscience
• Biotechnology
• Microscopy

Background:

• Three-dimensional (3D) multiplexed fluorescence imaging is crucial for neuroscience research.
• Cyclic immunofluorescence is common for 2D imaging but time-consuming for 3D specimens due to lengthy staining and imaging cycles.

Purpose of the Study:

• To develop a faster and more efficient method for 3D multiplexed fluorescence imaging.
• To reduce the number of imaging cycles required for volumetric protein analysis.

Main Methods:

• Proposed SEPARATE (Spatial Expression PAttern-guided paiRing And unmixing of proTEins) method.
• Utilized a neural network for signal unmixing based on 3D spatial expression patterns.
• Employed a feature extraction network to quantify spatial protein distinctions and identify suitable protein pairs for co-imaging.

Main Results:

• Validated the feature extraction network with ten proteins, demonstrating a strong correlation between spatial pattern distinction and signal unmixing accuracy.
• Successfully performed volumetric multiplexed imaging of six proteins using only three fluorophores.
• Achieved signal unmixing by pairing proteins based on feature-based distances and using protein separation networks.

Conclusions:

• SEPARATE significantly reduces imaging time for 3D multiplexed fluorescence imaging.
• The method enables efficient volumetric imaging of multiple proteins by leveraging spatial expression patterns and neural network-based signal unmixing.
• This technique offers a more practical approach for complex 3D neuroimaging studies.