single cell data analysis

Question

[vibes1002003]

Dear all,

I am working on single cell dataset from 10x . I am trying to integrate two datasets from two different experiment but both dataset from 10x. I would like to know when integrate two datasets, the name of barcodes have an effect on tsne clustering. For integtating two datasets, what would be the best approach?

Thanks, Rahul

[vibes1002003]

Thanks for the answer. I will try tools you mentioned above.

I also encounter some related to duplication of cell barcode in matrix file.
Have someone experience in this direction? Second, is it possible to combine umi tag with cellbarcode in matrix or count file fom 10x data ?

Answer 1

[swbarnes2]

Why is cellranger's aggr unsuitable for you?

[GenoMax]

Perhaps OP does not have access to cellranger and/or raw data? Someone else may have done the cellranger analysis.

Answer 2

[Haci]

If by "integration" you mean generating a single count table from two different datasets, you would need to add prefixes to your cell barcodes. I know that Seurat's Merge() (or sth like that) does it for you.

If you mean "aligning" datasets so as to remove batch effects, etc, you might need to check Seurat's CCA or scran's MNNCorrect. There are many more such algorithms.

Answer 3

[shoujun.gu]

No perfect merging methods. If you have to, maybe you can try MNN or scanorama.