batch-effect correction

[WeiWei060512]

I'm trying running DEsingle on an integrated dataset. I was wondering if there is a way to correct the batch effect.

[miaozhun]

Hi, I think you could try the single-cell RNA-seq data batch effects correcting R package batchelor, or refer to this article.

[WeiWei060512]

Hi, thanks for your reply.
As far as I know, after correct the batch effect, those packages/methods generate normalised counts matrix. But DEsingle takes raw counts matrix, does DEsingle also work with normalised counts?
Thanks.

[miaozhun]

Sorry, I'm afraid it won't work with the normalised counts.

[pluto-the-lost]

I think you could try SCVI, which is a VAE model and also use ZINB distribution to model scRNA data. It has a batch indicator in the latent space so if you set the indicator to the same, the output should be batch corrected gene expression counts