In my scRNA seq experiment, single-cell libraries were generated using the GemCode Single-Cell Instrument and Single Cell 3′ Library & Gel Bead Kit v2 and Chip Kit (10x Genomics) according to the manufacturer’s protocol. About 8700 cells were added to each channel with a targeted cell recovery estimate of 5000 cells.

However, after processing my raw data with CellRanger software, I get 737,280 cells (barcodes) and about 31,000 genes. I cannot understand why I get so many cells and why this barcode whitelist is not up to 10,000 cells. After filtering this data retaining genes expressed in >= 5 cells as well as retaining cells with at least 200 detected genes, I get around 2200 cells.



1 Answer 1


The raw data from cell ranger contains all of the barcodes detected in the experiment. These raw barcodes are filtered by cell ranger to identify the barcodes that likely represent cells rather than empty droplets / dead cells. You can find the filtered data in the filtered feature-barcode matrices. You can also examine the Run summary HTML file to see a summary of the cell ranger barcode filtering.

The reason why there are so many raw barcodes is an important optimization by 10X to increase the cell capture rate compared to techniques such as drop-seq. In a 10X experiment the number of gel beads is much higher than the number of cells which leads to a higher probability of a cell and a bead ending up in the same droplet. This also leads to a large number of empty droplets with just a bead and some ambient RNA or a dead cell. These all end up being sequenced and the empty drops tend to have a small number of unique molecules (the majority less than 10 UMIs / barcode).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.