Expanding the coverage of spatial proteomics

Abstract

Motivation: Multiplexed protein imaging methods provide valuable information on complex tissue structure and cellular heterogeneity. However, costs increase and image quality decreases with the number of biomarkers imaged, and the number of markers that can be measured in the same tissue sample is limited. Results: In this work, we propose an efficient algorithm to choose a minimal predictive subset of markers that for the first time allows the prediction of full images for a much larger set of markers. We demonstrate that our approach also outperforms previous methods for predicting cell-level marker composition. We also demonstrate that an effective minimal subset can be selected even if the desired full set is too large to be imaged on the same samples. Availability: All code and intermediate results are available in a Reproducible Research Archive at https://github.com/murphygroup/CODEXPanelOptimization.

MIM-CyCIF: Masked Imaging Modeling for Enhancing Cyclic Immunofluorescence (CyCIF) with Panel Reduction and Imputation

“…Sun et al . iteratively trained a U-Net to reconstruct patch-level images, aiding marker selection for a reduced panel 12 . In contrast to prior research, where panel selection is separate from full panel reconstruction, our method integrates iterative marker selection within a pre-trained model, streamlining panel reduction and reconstruction.…”