Completing spatial transcriptomics data for gene expression prediction benchmarking

Open Webpage

Daniela Ruiz, Paula Cárdenas, Leonardo Manrique, Daniela Vega, Gabriel Mateo Mejía Sepúlveda, Pablo Arbeláez

Published: 01 Jan 2025, Last Modified: 04 Nov 2025Medical Image Anal. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Release of SpaRED, a preprocessed set of 26 Spatial Transcriptomics datasets.•Propose SpaCKLE, a transformer-based method to complete missing gene expressions.•Benchmarking of eight common models for gene expression from histology images.•Open-source database and gene completion model with an easy-to-use Python library.