Different number of cells from cell ranger vdj report and after read.csv?

[denvercal1234GitHub]

Hi Nick,

Thanks again for this great tool and the continued dev.

Could you help me understand the difference in the number of cells from cellranger vdj web summary and that from viewing the object after read.csv of the filtered_contig_annotations.csv file (setting stringsAsFactors = F)?

Thank you!

### Having 6189 rows?
> dim(VDJ_obj)
[1] 6189   18

[ncborcherding]

Hey Quang,

Nice to hear from you again.

cellranger vdj web summary is reporting the number of cells with VDJ sequence (n=2,736) and number of cells with productive VDJ sequences (n=2,061).

In contrast, filtered_contig_annotations.csv will contain productive and nonproductive VDJ sequences across all cells - including both immune receptor chains. Thus, you should expect about 2 x number of cells from the cellranger vdj web summary (e.g. - 2*2736). This is not a perfect calculation - there will be cells with only a single chain recovered, as well as cells with > 2 chains recovered.

Hope that helps and thanks for reaching out,
Nick