Mitigation of Multi-scale Biases in Cell-type Deconvolution for Spatially Resolved Transcriptomics Using HarmoDecon

Abstract

Motivation: The advent of spatially resolved transcriptomics (SRT) has revolutionized our understanding of tissue molecular microenvironments by enabling the study of gene expression in its spatial context. However, many SRT platforms lack single-cell resolution, necessitating cell-type deconvolution methods to estimate cell-type proportions in SRT spots. Despite advancements in existing tools, these methods have not addressed biases occurring at three scales: individual spots, entire tissue samples, and discrepancies between SRT and reference scRNA-seq datasets. These biases result in overbalanced cell-type proportions for each spot, mismatched cell-type fractions at the sample level, and data distribution shifts across platforms. Results: To mitigate these biases, we introduce HarmoDecon, a novel semi-supervised deep learning model for spatial cell-type deconvolution. HarmoDecon employs a multi-scale bias correction strategy, including spot-level bias correction, sample-level bias correction, and cross-platform bias correction, to enhance deconvolution accuracy. We systematically evaluated HarmoDecon using both simulated and real SRT datasets across multiple platforms. The experimental results demonstrated that HarmoDecon significantly outperformed existing methods in accurately estimating cell-type proportions while maintaining computational efficiency.

Yu Xu 许煜

Ph.D. Candidate

My research interests include computational metabolomics, computational genomics, and large language models in metabolomics.