How to bring in Processed GEO Data ((scRNA-seq) into R and convert into Seurat Object for analysis ?

Question

[Vigorous]

For my own project, I'm attempting to search through a published dataset to be used for analysis. I am a R/Seurat newbie so I'm quite lost.

From what I understand, getGEO can only be used for microarray data and thus can't be used. When attempted it returned all the samples but 0 feature data.

Instead, I read it in using the function read.delim and it did work but it's arranged in a way that I'm not sure how to proceed.

The file in question (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE103322) is a processed .txt file and as a result, there is no barcode/cell ranger output that can be used to create a Seurat Object.

Thanks!

[Ram]

Contact the author and check with them if they're willing to share raw output from 10X (or a Seurat/SingleCellExperiment object)

[GenoMax]

These samples (there is 5900+) are not 10x. They are SMART-seq2 protocol.

Following is what is provided in the processed data file:

tab-delimited text file containing TPM values ( transcript-per-million reads) for 23,686 analyzed genes (rows), across 5902 cells (columns, sample names indicated at the top row). Headers indicate the enzyme being used for reverse transcription, the sample site (Primary vs. lymph node), the classification into cancer and non-cancer cells and the inferred non-cancer cell types.

See: https://github.com/satijalab/seurat/issues/3431