Where do I share a single-cell dataset?

Question

[Lluís R.]

Where is the best standard place to store single-cell datasets to be accessible for the community? Is GEO the right place? I looked up some single-cell datasets and the number of samples is quite low does each sample include all the cells fastq files?

• GSE57249: Has just 56 samples
• GSE57872: Has 587 samples
• GSE145926: Has 21 samples
• ?

Related to sharing the data set, is it worth to publish a Data Note of a single cell data set? (Tradeoff between salami-splicing publications and work on a new manuscript, citation/reference of the data set for multiple purposes...)

Answer 1

[Nitin Narwade]

I was running on the internet with the same query and I think yes GEO is the best place to deposit raw sequencing data. There are other single cell specific data repositories, for example, Broad Institute's Single Cell Portal, UCSC Cell Browser, PanglaoDB, Singel Cell Studies, and perhaps plenty of others. Usually, these portal has an advantage over GEO, where you can interactively access the scRNA-Seq data like visualizing cluster embeddings, visualizing gene expression, cluster-specific pathways distribution, etc. Nevertheless, you can upload the processed data (expression matrices, metadata, R objects, etc) to GEO, but the user needs to download and visualize it by themselves.

I have uploaded 10X genomics data to the GEO once and in my opinion, you can directly submit the samples (multiplexed; raw sequencing data) so that users can start processing your data from the scratch.

So in my case, we had 3 samples with approximately 14k cells in each. I submitted 3 samples with the processed data (by CellRanger, feature count matrix, metadata, cell barcodes, and Seurat object in .rds format).

I hope this would help you during your submission.

Cheers!!

;)

[Lluís R.]

Yes it helps a lot, it clears the misconception that I had that each cell would appear as a sample on GEO. The other links are very useful too. Thanks