Question: How does
cellranger aggr normalize for sequencing depth among multiple libraries?
Answer: When aggregating data from different libraries,
cellranger aggr normalizes for sequencing depth by downsampling the read counts across libraries.
cellranger aggr computes the downsampling rate for each library based on the mean number of confidently-mapped-to-transcriptome, valid-barcode reads per cell for each library. Libraries other than the one with lowest values are downsampled. In the example below, all libraries are normalized to the depth of Sample 4:
|Sample||Confidently Mapped to Transcriptome Reads per Cell||Fraction of Reads Kept|
While the downsampling rate is by default computed using "Confidently Mapped to Transcriptome Reads", the sampling is done from all reads with valid barcodes and valid UMIs.
It is also possible to normalize by downsampling based on raw reads per cell, or to not normalize at all. For more details, please see the "Depth Normalization" section at the bottom of this page.