I am trying to learn single cell RNA-Seq data analysis. Right now I am following Seurat and Monocle documentation. I am aware of the fact that we have to create a single cell experiment object, which basically acts as a container for the scRNA data. But, I want make one these container from scratch so as to understand how the expression data and metadata are stored.
Most of the tutorials use .RDS or .MTX files to create SingleCellExperiement object. It seems like a 'black-box' for me because I cannot open and see the data as in CSV format. What I want to do is to a create CSV or TSV file for expression data and assay description and then combine these files to make a SingleCellExperiment object. Can anybody help to create a dummy SingleCellExperiment object and how metadata is added to this objects later ? Any link to the tutorials would also work.
See https://bioconductor.org/packages/devel/bioc/vignettes/SingleCellExperiment/inst/doc/intro.html
It is basically a SummarizedExperiments container for which a lot of documentation is available.
A good amount of datasets available in Cellranger repository to start with.
https://support.10xgenomics.com/single-cell-gene-expression/datasets
The 10x raw data directory looks something like following.
.RDS is an R object, so it could be anything. .MTX is a file format for storing sparse matrices. It's used by 10x Genomics, which is probably the most common platform now, so a lot of tutorials use that as a starting point. You can read more about it on 10x Genomics support website where they discuss it in depth and include code examples.