Journal: Genomics, Proteomics & Bioinformatics

Article Title: LEGEND: Identifying Co-expressed Genes in Multimodal Transcriptomic Sequencing Data

doi: 10.1093/gpbjnl/qzaf056

Figure Lengend Snippet: Spatial expression patterns of LEGEND-identified gene clusters in mouse brain and human DLPFC Gene clusters identified by LEGEND from the SRT datasets of mouse brain and human DLPFC are divided into groups of high, medium, and low co-expression quality based on their ICQ values. Rows display a gene cluster randomly selected from each of these groups, respectively. The denoised spatial expression patterns of four genes randomly chosen from the cluster are visualized in each row. A . Mouse brain (the mBrain-SRT dataset). B . Human DLPFC (the hDLPFC-SRT dataset). ICQ, intra-cluster co-expression quality; DLPFC, dorsolateral prefrontal cortex.

Article Snippet: Three adult mouse brain datasets were downloaded from the 10X Genomics official website: mBrain-SRT ( https://support.10xgenomics.com/spatial-gene-expression/datasets/1.1.0/V1_Adult_Mouse_Brain ); mouse brain-Formalin-Fixed Paraffin-Embedded (mBrain-FFPE; https://www.10xgenomics.com/datasets/adult-mouse-brain-if-stained-ffpe-1-standard-1-3-0 ); and mouse brain-high definition (mBrain-HD; https://www.10xgenomics.com/datasets/visium-hd-cytassist-gene-expression-libraries-of-mouse-brain-he ).

Techniques: Expressing

Journal: Genomics, Proteomics & Bioinformatics

Article Title: LEGEND: Identifying Co-expressed Genes in Multimodal Transcriptomic Sequencing Data

doi: 10.1093/gpbjnl/qzaf056

Figure Lengend Snippet: Performance comparison between LEGEND and competing methods for identifying gene co-expression groups in scRNA-seq and SRT datasets from mouse and human brain A . The overall quality of gene co-expression groups identified in each of the four scRNA-seq datasets is quantified using co-expression DB index (Y-axis), where a lower score indicates superior clustering quality. B . The overall quality of gene co-expression groups identified in each of the 15 SRT datasets is quantified using spatial coherence DB index (Y-axis), where a lower score indicates more effective clustering. In dataset names, the prefix “m” represents “mouse”, the prefix “h” represents “human”, and the suffix “sc” represents “scRNA-seq”. DB, Davies–Bouldin; AD, Alzheimer’s disease; MTG, middle temporal gyrus; CS-CORE, cell-type-specific co-expressions; COTAN, CO-expression Tables ANalysis; CNN-PReg, convolutional neural network with protein–protein interaction-graph regularization; SPARK, spatial pattern recognition via kernels.

Article Snippet: Three adult mouse brain datasets were downloaded from the 10X Genomics official website: mBrain-SRT ( https://support.10xgenomics.com/spatial-gene-expression/datasets/1.1.0/V1_Adult_Mouse_Brain ); mouse brain-Formalin-Fixed Paraffin-Embedded (mBrain-FFPE; https://www.10xgenomics.com/datasets/adult-mouse-brain-if-stained-ffpe-1-standard-1-3-0 ); and mouse brain-high definition (mBrain-HD; https://www.10xgenomics.com/datasets/visium-hd-cytassist-gene-expression-libraries-of-mouse-brain-he ).

Techniques: Comparison, Expressing

Journal: Genomics, Proteomics & Bioinformatics

Article Title: LEGEND: Identifying Co-expressed Genes in Multimodal Transcriptomic Sequencing Data

doi: 10.1093/gpbjnl/qzaf056

Figure Lengend Snippet: LEGEND improves both single-cell and spatial clustering performance A . SpaGCN is employed for spatial clustering across 13 SRT datasets, while Seurat v5 for single-cell clustering in one scRNA-seq dataset (the rectangle-enclosed panel). Both methods utilize feature gene sets selected by LEGEND or six competing methods. The X-axis displays the ARI changes (+, gain; −, loss) compared to baseline performance achieved using the complete gene set (red numbers). The number of genes selected by each method is noted on their bars. B . Enhanced tissue domain detection within mouse brain using LEGEND-selected feature gene set. The leftmost panel displays the ground-truth domain labels of the mBrain-SRT dataset, with two adjacent tissue domains, Cortex_4 and Cortex_5, outlined by a black rectangle. The right panels show magnified views of the Cortex_4 and Cortex_5 domains, comparing domain labeling by SpaGCN using feature genes selected by LEGEND and competing methods. Notably, only scGeneClust and LEGEND effectively distinguish Cortex_5 from Cortex_4, as indicated by yellow circles. The accuracy of spatial domain detection is measured using ARI, shown above each right panel. ARI, Adjusted Rand Index; SPARK-X, SPARK-eXpedited; BinSpect, Binary Spatial Extraction; VST, variance stabilizing transformation; M3Drop, fitting Michaelis-Menten function to the relationship between mean expression and dropout-rate.

Article Snippet: Three adult mouse brain datasets were downloaded from the 10X Genomics official website: mBrain-SRT ( https://support.10xgenomics.com/spatial-gene-expression/datasets/1.1.0/V1_Adult_Mouse_Brain ); mouse brain-Formalin-Fixed Paraffin-Embedded (mBrain-FFPE; https://www.10xgenomics.com/datasets/adult-mouse-brain-if-stained-ffpe-1-standard-1-3-0 ); and mouse brain-high definition (mBrain-HD; https://www.10xgenomics.com/datasets/visium-hd-cytassist-gene-expression-libraries-of-mouse-brain-he ).

Techniques: Labeling, Extraction, Transformation Assay, Expressing

Journal: Nucleic Acids Research

Article Title: CellMap: precision mapping of cellular landscape in spatial transcriptomics

doi: 10.1093/nar/gkaf1484

Figure Lengend Snippet: Benchmark CellMap on the Visium HD data from human CRC. ( A )The UMAP layout depicting the clustering space of human CRC scRNA-seq data (B cells, Endothelial, Fibroblast, Intestinal Epithelial,Myeloid, Neuronal, Smooth Muscle, T cells, and Tumor). The cell types are color-coded, with each dot representing an individual cell. ( B ) Spatial structure of human CRC reconstructed using CellMap. ( C ) Spatial heat maps showing the spatial distribution of nine cell types predicted by CellMap in the Visium HD ST data, with each cell type highlighted in a different color. ( D ) Spatial heat maps showing cell type signature genes score calculated using AddModuleScore in Seurat. The colors from blue to red indicate the scores from low to high. ( E ) Benchmark of CellMap’s performance with different methods. The box plot reflects the overall distribution of Pearson’s correlation calculated for each spot by various method.

Article Snippet: Dataset3 (mouse kidney): 10X Visium, https://www.10xgenomics.com/resources/datasets?query=&page=1&configure%5BhitsPerPage%5D=50&configure%5BmaxValuesPerFacet%5D=1000 ; 10X Visium, GSE171406 in the GEO database [ ]; 10X Chromium, GSE129798 in the GEO database [ ].

Techniques:

Journal: bioRxiv

Article Title: SpaNorm: spatially-aware normalisation for spatial transcriptomics data

doi: 10.1101/2024.05.31.596908

Figure Lengend Snippet: (A) Estimated proportion of genes with region-specific library size effects. On average, CosMx and STOmics datasets have the highest proportion of genes exhibiting region-specific effects, followed by Xenium. Visium datasets have the lowest proportion. (B) Adjusted Rand Index of clusters identified using differently normalised data vs annotated spatial regions. Boxplots show the summary by platform. The coloured bars above each group of boxplots indicate the best-performing method for each dataset that makes up the group, based on maximum (darker-shade) and median (lighter-shade) statistics.

Article Snippet: The emergence of both spot-based spatial transcriptomics technologies (ST) such as 10x Genomics Visium [ ] as well as subcellular spatial transcriptomics (SST) technologies, such as 10x Genomics Xenium [ ], NanoString CosMx [ ], BGI Stereo-seq [ ], and Vizgen MERSCOPE [ ], hold the promise to address previously inaccessible biological questions and enhance our understanding of intercellular communication by preserving tissue architecture.

Techniques:

Journal: bioRxiv

Article Title: SpaNorm: spatially-aware normalisation for spatial transcriptomics data

doi: 10.1101/2024.05.31.596908

Figure Lengend Snippet: Within-set Spearman’s correlations of MERINGUE’s SVG statistic. (left) Visium Human DLPFC sets, (right) CosMx Human Lung, Xenium Human Breast Cancer and Xenium Mouse Brain sets. Higher Spearman’s correlation indicates more consistent SVG rankings among replicates of the same set. (* Breast cancer dataset with replicates of the same section)

Article Snippet: The emergence of both spot-based spatial transcriptomics technologies (ST) such as 10x Genomics Visium [ ] as well as subcellular spatial transcriptomics (SST) technologies, such as 10x Genomics Xenium [ ], NanoString CosMx [ ], BGI Stereo-seq [ ], and Vizgen MERSCOPE [ ], hold the promise to address previously inaccessible biological questions and enhance our understanding of intercellular communication by preserving tissue architecture.

Techniques: