Synthetic Analyses of Single-Cell Transcriptomes from Multiple Brain Organoids and Fetal Brain
Description
Human brain organoid systems offer unprecedented opportunities to investigate both neurodevelopmental and neurological disease. Single cell-based transcriptomics or epigenomics have dissected the cellular and molecular heterogeneity in the brain organoids, revealing a complex organization. Similar but distinct protocols from different labs have been applied to generate brain organoids, providing a large resource to perform a comparative analysis of brain developmental processes. Here we take a systematic approach to compare the single cell transcriptomes of various human cortical brain organoids together with fetal brain to define the identity of specific cell types and differentiation routes in each method. Importantly, we identify unique developmental programs in each protocol, which will be a critical benchmark for the utility of human brain organoids in the future.
Files
Steps to reproduce
The scRNA-seq datasets derived from microdroplet platforms were retrieved and collected from NCBI Short Read Archive. For 10x Genomics platform, either SRA or BAM-formatted files were downloaded and converted into fastq files by fastq-dump (v2.9.4) or bamtofastq (v1.1.2), respectively. Raw sequence files were mapped to the hg19 human genome and the latest Ensembl gene annotation by count function of CellRanger with default parameters (v2.1.0). For the other platforms of single-cell transcriptomes, gene-cell count matrices were obtained from NCBI Gene Expression Omnibus (GEO). We used PGP1 scRNA-seq datasets from Velasco et al., because PGP1 datasets include multiple batch replicates. Single-cell transcriptome profile from human fetal cortex was also obtained from NCBI GEO (Zhong et al., 2018). To read the count matrix data into R, please do as follow: (Method. 1) Read data from matrix file 1. Make a directory for matrix files. ($ mkdir matrix) 2. Download "barcodes.tsv.gz", "features.tsv.gz" and "combined_Matrix_v3_6.mtx.gz" into "matrix" directory. 3. Rename "combined_Matrix_v3_6.mtx.gz" into "matrix.mtx.gz". 4. start R ($ R) and type as follow: > library(Seurat) #Please use Seurat version 3 > mat <- Read10X("matrix/", gene.column=1) (Method 2) Read data from R object 1. Download and decompress "raw_count.tar.gz" ($ tar xvfz raw_count.tar.gz) 2. start R ($R) and type as follow: > load("combined_Matrix_v3_6.dat")