SEVtras delineates small extracellular vesicles at droplet resolution from single-cell transcriptomes
SEVtras stands for sEV-containing droplet identification in scRNA-seq data.
You can freely use SEVtras to explore sEV heterogeneity at single droplet, characterize cell type dynamics in light of sEV activity and unlock diagnostic potential of sEVs in concert with cells.
Overview of SEVtras.
"numpy", "pandas", "scipy", "umap",
"statsmodels", "gseapy", "scanpy"
pip install SEVtrasWe also suggest to use a separate conda environment for installing SEVtras.
conda create -y -n SEVtras_env python=3.7
source activate SEVtras_env
pip install SEVtrasInput for SEVtras is a cell-by-gene matrix. In the case of scRNA-seq dataset using 10X Genomics, we used raw_feature_bc_matrix directory generated by Cell Ranger as input. Output of SEVtras consists of the score of sEV signals and classification for each droplet. Such sEV information will be used for downstream analysis and as basis for the construction of the sEV secretion activity index (ESAI) for different cell types.
We implemented four functions for sEV recognizing and functional analyses. In SEVtras, sEV_recognizer recognizes sEV-containing droplets in the raw scRNA-seq data; ESAI_calculator calculates sEV secretion activity for samples and deconvolves these droplets to their original cell type and estimited corresponding sEV secretion activity; cellfree_simulator simulates transcriptional profile of cell free droplets in scRNA-seq; and sEV_enrichment performs GO enrichment for sEV data.
The pipeline of SEVtras only composed two parts: sEV_recognizer and ESAI_calculator.
Part I:
SEVtras.sEV_recognizer(input_path='./tests', sample_file='./tests/sample_file', out_path='./outputs', species='Homo')Part II:
SEVtras.ESAI_calculator(adata_ev_path='./outputs/sEVs_SEVtras.h5ad', adata_cell_path='./outputs/adata_cell.h5ad', out_path='./outputs', Xraw=False, OBSsample='batch', OBScelltype='celltype')Further tutorials please refer to https://SEVtras.readthedocs.io/.