Reply

Merge datasets from different tissues sequenced with 10X

Posted By: mBrio, on Nov 15, 2018 at 11:56 PM

Dear 10X community,

 

I have a dataset composed by 4 mouse brain tumors, I would like to merge it with a published dataset of normal mouse brain (containing multiple cell types). Both datasets are obtained with 10X technology. Which are the computational steps to perform, in terms of normalization? Is it possible to assume that a linear regression of the variable "original.identity" (as e.g. performed in Seurat, with the function scale.data) is enough to correct for batch effects?

 

Thanks a lot for your support!

 

Highlighted

Re: Merge datasets from different tissues sequenced with 10X

Posted By: shauna-10x, on Nov 19, 2018 at 9:28 AM

Hi,

 

After reviewing your question, our support team suggested I open a support ticket on your behalf.  They will be able to provide the best support via their ticketing system.  You should have received an email from support@10xgenomics.com.

 

Thanks,

Shauna

Re: Merge datasets from different tissues sequenced with 10X

Posted By: agottscho, on Nov 19, 2018 at 1:37 PM

Hi mBrio,

 

This is Andrew Gottscho with 10x Genomics support.

 

The analysis workflow for this situation is as follows. The diagrams on this page are a good place to start: https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/what-is-cell-r...

 

1. Run cellranger count on each library separately. A library corresponds to a single GEM well (Chromium chip channel). Please see: https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/using/count 

2. Run cellranger aggr to combine the libraries. There are three different normalization options, the default is to subsample reads from higher-depth GEM wells until they all have an equal number of confidently mapped reads per cell. Please see: https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/using/aggregat...

 

Regarding seurat, since this is a third party tool that is not supported by 10x, I recommend contacting seuratpackage@gmail.com with any questions.

 

Hope this helps. If you have any follow up questions for us please email support@10xgenomics.com.

 

Best, Andrew