Mosaic-PICASSO: accurate crosstalk removal for multiplex fluorescence imaging

Area of Science:

• Biomedical Imaging
• Optical Microscopy
• Computational Biology

Background:

• Ultra-multiplexed fluorescence imaging allows simultaneous visualization of multiple targets.
• Mutual information (MI)-based techniques like PICASSO simplify demixing up to 15 fluorophores.
• MI techniques struggle to distinguish spatial colocalization from spectral mixing, causing overcorrection.

Purpose of the Study:

• To address the overcorrection limitation in MI-based demixing techniques.
• To improve the accuracy of spectral demixing in ultra-multiplexed fluorescence imaging.
• To enhance the reliability of spatial colocalization analysis in complex biological samples.

Main Methods:

• Developed a novel approach to spectral mixing measurement within multiplex images.
• Utilized regions with low-spatial similarity to isolate spectral information.
• Applied this method to refine mutual information-based demixing.

Main Results:

• Demonstrated that measuring spectral mixing in low-spatial similarity regions significantly improves demixing accuracy.
• Showed a reduction in overcorrection errors compared to standard MI techniques.
• Validated the effectiveness of the proposed method for ultra-multiplexed imaging.

Conclusions:

• The proposed method enhances the accuracy of ultra-multiplexed fluorescence imaging by mitigating overcorrection.
• This technique facilitates more reliable spatial colocalization analysis.
• The findings promise to accelerate the adoption and application of advanced multiplex imaging.