Integration more than one million cell

Question

[Andy]

Hello everyone,

I have encountered a dataset with over one million cells, and I have tried using scVI, scANVI, and scanorama for integration. Unfortunately, these methods did not perform as well as I had hoped. Therefore, I would like to kindly ask if any of you are aware of other packages that are suitable for integrating such a large dataset with over one million cells. Your suggestions and insights would be greatly appreciated.

Thank you in advance for your help. Best Andy

[Radu Tanasa]

What do you mean by "did not perform as well as I had hoped"? Could you please provide more details? Do the cells still cluster by sample as opposed to expected type (eg: marker gene scores).

[Kevin Blighe]

Could try the integration manually via an abacus, but it may take longer than the life-time of the Universe. Are you hitting a memory limit? Please share any error messages. Thanks Andy.

Answer 1

[cmdcolin]

scanpy's tagline is "scales to >1M cells" https://github.com/scverse/scanpy