PICASSO

Unmix spectral spillover

PICASSOnn

PICASSOnn is an algorithm to remove spillover fluorescence by minimizing the mutual information (MI) between sink and source images¹. MI between pairs of sink and source images is estimated and minimized with a neural network (MINE²)using stochastic gradient descent and GPU acceleration.

Mixing model

\[unmixed sink = sink - \sum_{i} \alpha_i(source_i - \beta_i)\]

References

1. Seo, J. et al. PICASSO allows ultra-multiplexed fluorescence imaging of spatially overlapping proteins without reference spectra measurements. Nat Commun 13, 2475 (2022).
2. Belghazi, M. I. et al. MINE: Mutual Information Neural Estimation. arXiv:1801.04062 [cs, stat] (2018).

Input

Zarr store of processed images YAML section summary

Output

YAML unmixing parameters

Log

Script log