SgsAnnData

SgsAnnData is an R package that facilitates the seamless conversion of single-cell analysis object from popular tools such as Seurat, Giotto, Signac, and ArchR into the AnnData. This format can be directly visualized in SGS which is an interactive browser for single-cell and spatial multimodal datasets.

Warning

Please note that the usage of SGS requires the prior installation of R packages such as Seurat, Giotto, Signac, and anndata. It is important to ensure that these packages are installed beforehand for smooth execution of SGS workflows.

Installation

Currently, SgsAnndata can only be installed through GitHub. However, we are actively working on providing a Bioconductor installation method in the near future.

devtools::install_github("bio-xtt/SgsAnnDataV2")

Usage

Seurat to AnnData

Single-Cell Transcriptome Object Conversion

When Seurat contains multiple assays, users can provide multi assay names they wish to export. This will automatically generates the h5ad for each assay type, like RNA.h5ad,integration.h5ad etc. We also offer the flexibility for users to output marker dataframes as a list, with the corresponding information stored in adata.uns['markers'].

library(Seurat)
library(anndata)
SeuratToAnndata(object=scRNA,
                outpath="/test_adata",
                assays=c("RNA", "SCT"),
                groups = NULL,
                reductions = c("tsne","umap"),
                markersDF = list("RNA"=rna_marker_df, "SCT"=sct_marker_df)) 

Spatial Transcriptome Object Conversion

When converting spatial transcriptomic data, we automatically determine the spatial data type based on the object's structure. We return the corresponding h5ad file based on the identified type. The spatial coordinates are stored within adata.obsm['spatial']. If the analyzed object includes spatially organized slice information, we store it in adata.uns['spatial']. Multiple slices are differentiated using different "library_id", ensuring clear distinction between them.

SeuratToAnndata(object=scRNA,
                outpath="/test_adata",
                assays=c("RNA", "Spatial"),
                groups = c("seurat_clusters", "region"),
                reductions = "tsne",
                markersDF = list("RNA"=rna_marker_df, "Spatial"=st_marker_df)) 

Signac to Anndata

We recommend exporting peak matrix, genescore matrix, and motif matrix from single-cell ATAC data analysis as h5ad files. If desired, these files can be further converted into Mudata, which offers a comprehensive and annotated multimodal dataset structure. Additionally, users can use export_pwm to output results from co-accessibility analysis and motif enrichment analysis, respectively. Users can access motif-related information via adata.uns.["motifs"].

SignacToAnndata(object=scATAC,
                outpath = "/test_adata",
                assays=c("RNA", "Peaks", "Motif"),
                assay.types=("gene","peak","motif"),
                markersDF = list("RNA"=gene_marker_df, "Peaks"=marker_peak_df, "Motif"=marker_motif_df),
                groups = c("clusters","sample", "age"),
                reductions = c("umap","tsne"),
                export_pwm = TRUE)   # to export motif pwm 

ArchR to AnnData

Similar to the Signac object conversion, users can convert ArchR objects and export relevant information based on their needs.

ArchrToAnndata(object=project5,
               outpath="/test_adata",
               assays=c("RNA", "Peaks", "Motif"),
               assay.types=("gene","peak","motif"),
               markersDF=list("RNA"=gene_marker_df, "Peaks"=marker_peak_df, "Motif"=marker_motif_df),
               groups = NULL, ## export all cell meta
               reductions = NULL, ## export all reduction coords
               export_pwm =TRUE) 

Giotto to AnnData

Similar to Seurat's conversion of spatial transcriptomic analysis results, we seamlessly incorporate tissue slices and spatial coordinates from Giotto object into the output h5ad object by default.

giottoToAnnData(object = giotto,
                outpath = "/test_adata",
                markerDF = NULL)

Citiation

Xia, T., Sun, J., Lu, F., Luo, Y., Mao, Y., Xu, L., & Wang, Y. (2024). Empowering Integrative and Collaborative Exploration of Single-Cell and Spatial Multimodal Data with SGS. bioRxiv, (), 2024.07.19.604227. Accessed July 23, 2024. https://doi.org/10.1101/2024.07.19.604227.