Recently, the research group of Professor Qu Kun from the School of Life Sciences and Medicine of the University of Science and Technology of China systematically evaluated the performance of 16 spatial transcriptome and single-cell transcriptome data integration algorithms in predicting the spatial distribution of genes or cell types by designing a set of analytical procedures. The research results are entitled "Benchmarking spatial and single-cell transcriptomics integration methods for transcript distribution prediction. "and cell type deconvolution" was published online in the internationally renowned academic journal Nature Methods on May 16, 2022.

The spatial position of cells in tissues and organs is crucial for their specific functions. In recent years, researchers have developed a variety of spatial transcriptome technologies that can detect the expression of the whole transcriptome in cells while preserving the quasi-spatial localization of cells, so as to study the cell subsets and their molecular mechanisms that play a key role in the development or development of diseases. However, the existing spatial transcriptome technology has two shortcomings: 1. Sequence-based spatial transcriptome technology cannot achieve true single-cell resolution; 2. 2. Imaging spatial transcriptome-based techniques can detect limited gene fluxes. To overcome technical limitations, bioinformaticians have designed a variety of algorithms to integrate spatial transcriptome and single-cell transcriptome data to predict the spatial distribution of cell types and/or the complete transcriptome information of individual cells. These algorithms have greatly deepened our understanding of spatial transcriptomic data and related biological and pathological processes. However, due to the significant differences in the working principle and application scope of different algorithms, it is difficult for researchers to choose the best algorithm to predict the spatial distribution of cell types and gene expression.

Professor Qu Kun's research group has long been committed to the development of biological big data analysis algorithms and software. In this study, we collected 45 spatial transcriptome and single-cell transcriptome datasets and 32 simulation datasets from the same tissue source, and designed a variety of indicators to systematically evaluate the performance of 16 integration algorithms from multiple dimensions such as accuracy, robustness, and computing resource time.

图1.整合分析流程

The results show that Cell2location, SpatialDWLS and RCTD algorithms can predict the spatial distribution of cell types more accurately. Tangram, gimVI and SpaGE algorithms are the best algorithms to predict the spatial distribution of gene expression. Tangram, Seurat and LIGER have relatively high computational efficiency and are suitable for processing large data sets. This research work summarizes the properties, performance and applicability of each algorithm, summarizes the advantages of efficiency algorithms, and provides a reference for researchers to further improve the performance of algorithms. It also provides an analytical process on github that integrates spatial and single-cell transcriptome data to help researchers choose the best analytical tools for processing their own data.

Qu Kun, Professor, School of Life Sciences and Medicine, USTC, is the corresponding author of this paper. Li Bin, special associate researcher of the research group, Zhang Wen, PhD student, and Guo Chuang, special associate researcher, are the co-first authors of this paper. This work is supported by the Outstanding Young People Fund of the Foundation Committee, the National Key Research and Development Program, the Natural Science Fund of the Foundation Committee, the Basic Research Youth Team of the Chinese Academy of Sciences, and the major science and Technology projects of Anhui Province. The research group of Professor Xue Tian, Professor Chen Fa Lai and Professor Cheng Linzhao from the University of Science and Technology of China provided great support for the smooth development of this work.

Paper link：https://www.nature.com/articles/s41592-022-01480-9

(Department of Life Sciences and Medicine, Department of Scientific Research)
Source: HKUST News Network