The paper details the development and validation of PASTA (PAthway-oriented Spatial gene impuTAtion), a novel computational framework designed to enhance spatial transcriptomics data. By integrating scRNA-seq reference data with cell type and spatial proximity information, this method accurately predicts unmeasured gene expressions at the biological pathway level. Unlike traditional techniques that focus on individual genes, PASTA aggregates signals from functionally related gene sets to reduce noise and provide more robust biological interpretations. The researchers demonstrate its effectiveness through extensive simulations and applications on real-world datasets, including human lung cancer and mouse brain tissues. Results indicate that PASTA consistently outperforms existing tools in maintaining prediction stability and capturing complex developmental trajectories. This advancement offers a more precise way to translate spatial findings into meaningful insights for disease research and clinical diagnostics.

References:

• Li R, Yang P, Di Pilato M, et al. Accurate imputation of pathway-specific gene expression in spatial transcriptomics with PASTA[J]. Nature Communications, 2025.