Abstract:Abstract Summary Spatial transcriptomics has changed our way to study tissue structure and cellular organization. However, there are still limitations in its resolution, and most available platforms do not reach a single cell resolution. To address this issue, we introduce SpatialDDLS, a fast neural network-based algorithm for cell type deconvolution of spatial transcriptomics data. SpatialDDLS leverages single-cell RNA sequencing data to simulate mixed transcriptional profiles with predefined cellular composition, which are subsequently used to train a fully connected neural network to uncover cell type diversity within each spot. By comparing it with two state-of-the-art spatial deconvolution methods, we demonstrate that SpatialDDLS is an accurate and fast alternative to the available state-of-the art tools. Availability and implementation The R package SpatialDDLS is available via CRAN-The Comprehensive R Archive Network: https://CRAN.R-project.org/package=SpatialDDLS. A detailed manual of the main functionalities implemented in the package can be found at https://diegommcc.github.io/SpatialDDLS.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the challenge of cell - type deconvolution in spatial transcriptomics data. Specifically, although spatial transcriptomics techniques have significantly enhanced our ability to study tissue structure and cell organization, most existing techniques still cannot achieve single - cell resolution. Therefore, computational methods are required to identify the exact cell combinations within each sequencing site. The paper introduces a fast neural - network - based algorithm, SpatialDDLS, for cell - type deconvolution of spatial transcriptomics data. By comparing with two existing state - of - the - art spatial deconvolution methods, the advantages of SpatialDDLS in terms of accuracy and speed are demonstrated. ### Main problems 1. **Resolution limitations of spatial transcriptomics data**: Most existing spatial transcriptomics platforms cannot achieve single - cell resolution, which restricts the precise analysis of cell - type diversity within tissues. 2. **Requirement for computational methods**: To overcome the resolution limitations, computational methods need to be developed to identify the cell - type composition within each sequencing site. 3. **Limitations of existing methods**: Existing deconvolution methods usually rely on predefined cell markers and are computationally costly and have long running times. ### Solutions - **SpatialDDLS**: An R package based on neural networks for cell - type deconvolution of spatial transcriptomics data. - **Main steps**: 1. **Simulating mixed transcriptional profiles**: Use single - cell RNA - sequencing (scRNA - seq) data to generate mixed transcriptional profiles with known cell compositions. 2. **Neural - network training and evaluation**: Use the simulated mixed transcriptional profiles to train and evaluate the neural - network model. 3. **Deconvolution and spatial regularization**: Use the trained model to predict the cell composition of each sequencing site and regularize it by considering the spatial information of neighboring sites. 4. **Optional features**: Provide a model - explanation module and a clustering module to enhance the interpretability and practicality of downstream analysis. ### Performance evaluation - **Benchmark tests**: Compared with two state - of - the - art spatial deconvolution methods (cell2location and RCTD), the results show that SpatialDDLS performs well on multiple datasets, especially in predicting major cell types. - **Computational efficiency**: SpatialDDLS has an advantage in terms of computational resource consumption and can be easily implemented in a regular workflow. In conclusion, this paper aims to improve the accuracy and efficiency of cell - type deconvolution in spatial transcriptomics data by developing the SpatialDDLS tool, thereby better understanding the cell composition and function within tissues.

SpatialDDLS: an R package to deconvolute spatial transcriptomics data using neural networks

NLSDeconv: an efficient cell-type deconvolution method for spatial transcriptomics data

SpatialDWLS: accurate deconvolution of spatial transcriptomic data

SpatialDE2: Fast and localized variance component analysis of spatial transcriptomics

SpatialCTD: A Large-Scale Tumor Microenvironment Spatial Transcriptomic Dataset to Evaluate Cell Type Deconvolution for Immuno-Oncology

AntiSplodge: a neural-network-based RNA-profile deconvolution pipeline designed for spatial transcriptomics

Robust decomposition of cell type mixtures in spatial transcriptomics

Advances in mixed cell deconvolution enable quantification of cell types in spatial transcriptomic data

Benchmarking and integration of methods for deconvoluting spatial transcriptomic data

Spatial transcriptomics deconvolution at single-cell resolution by Redeconve

Spatial mutual nearest neighbors for spatial transcriptomics data

Advances in mixed cell deconvolution enable quantification of cell types in spatially-resolved gene expression data

EnDecon: cell type deconvolution of spatially resolved transcriptomics data via ensemble learning

Reference-free cell type deconvolution of multi-cellular pixel-resolution spatially resolved transcriptomics data

A systematic evaluation of state-of-the-art deconvolution methods in spatial transcriptomics: insights from cardiovascular disease and chronic kidney disease

Masked adversarial neural network for cell type deconvolution in spatial transcriptomics

SpaDecon: cell-type deconvolution in spatial transcriptomics with semi-supervised learning

SpatialPrompt: spatially aware scalable and accurate tool for spot deconvolution and domain identification in spatial transcriptomics

Deconvolution algorithms for inference of the cell-type composition of the spatial transcriptome

SD2: Spatially resolved transcriptomics deconvolution through integration of dropout and spatial information

SpatialDeX is a Reference-Free Method for Cell Type Deconvolution of Spatial Transcriptomics Data in Solid Tumors