Completing Spatial Transcriptomics Data for Gene Expression Prediction Benchmarking

2025-05-05

AI Summary

Spatial transcriptomics (e.g., 10x Visium) suffers from high cost and low gene capture efficiency, resulting in severely sparse expression data; meanwhile, inconsistent datasets and training protocols hinder fair benchmarking of deep learning models. Method: We introduce SpaRED—the first standardized benchmark for spatial gene expression prediction—comprising uniformly preprocessed, registered, and evaluated data from 26 institutions. We further propose SpaCKLE, a novel Transformer-based model enabling end-to-end histology-to-full-transcriptome imputation. Results: SpaCKLE reduces MSE by 82.5% across all 26 datasets. SpaRED substantially improves performance across eight state-of-the-art models and establishes the most comprehensive, reproducible evaluation standard to date. Crucially, we demonstrate—for the first time—that imputed expression profiles consistently enhance downstream tasks, validating their broad utility in spatial omics analysis.

Technology Category

• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data
• Machine Learning: Representation Learning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Spatial Transcriptomics is a groundbreaking technology that integrates histology images with spatially resolved gene expression profiles. Among the various Spatial Transcriptomics techniques available, Visium has emerged as the most widely adopted. However, its accessibility is limited by high costs, the need for specialized expertise, and slow clinical integration. Additionally, gene capture inefficiencies lead to significant dropout, corrupting acquired data. To address these challenges, the deep learning community has explored the gene expression prediction task directly from histology images. Yet, inconsistencies in datasets, preprocessing, and training protocols hinder fair comparisons between models. To bridge this gap, we introduce SpaRED, a systematically curated database comprising 26 public datasets, providing a standardized resource for model evaluation. We further propose SpaCKLE, a state-of-the-art transformer-based gene expression completion model that reduces mean squared error by over 82.5% compared to existing approaches. Finally, we establish the SpaRED benchmark, evaluating eight state-of-the-art prediction models on both raw and SpaCKLE-completed data, demonstrating SpaCKLE substantially improves the results across all the gene expression prediction models. Altogether, our contributions constitute the most comprehensive benchmark of gene expression prediction from histology images to date and a stepping stone for future research on Spatial Transcriptomics.

Problem

• Standardizing gene expression prediction benchmarking for Spatial Transcriptomics
• Addressing gene capture inefficiencies and data corruption in Visium
• Improving model comparisons with curated datasets and preprocessing

Innovation

• SpaRED: standardized database for model evaluation
• SpaCKLE: transformer-based gene expression completion
• Benchmarking eight models with SpaCKLE-improved data