Oct 31, 2019

Letter from 10x Genomics

Olivia Habern

At 10x Genomics, we strive to communicate proactively with our customers on issues that may affect the interpretation of results generated using our products. Recently, the Buenrostro Lab identified a new artifact called Barcode Multiplets in data generated using our Single Cell ATAC Solution (Lareau et al. 2019 bioRxiv). For most customers this should not be a concern as it will not impact your research. For customers using our Single Cell ATAC Solution, we have identified the cause of the artifact and determined that with computational processing this small amount of data affected by Barcode Multiplets can be readily removed.

Briefly, the presence of more than one barcode in a single droplet can occur in two different ways:

  • Barcode Multiplets. This refers to a droplet that receives a gel bead carrying more than one barcode. The presence of more than one barcode can lead to what appears to be a copy of the original cell but with much lower sequencing coverage.
  • Gel Bead Multiplets. This is a known phenomenon that occurs when more than one barcoded gel bead is present in a droplet. This results from the fundamental tradeoff for maximizing the fraction of droplets with only one gel bead versus those with none. We have previously reported that less than 5% of all droplets generated contain more than one gel bead (Zheng et al. 2017, Nat Comm).

We have confirmed the presence of elevated Barcode Multiplets in our Single Cell ATAC data. In this product, the elevated levels are systematic, can be detected computationally and can be removed with software processing. For other 10x products, including Single Cell Gene Expression, Immune Profiling and CNV Solutions, we have confirmed that the rate of Barcode Multiplets is very low (<1%).

To address this, we have taken the following actions:

  • We have developed an algorithm to detect and remove data arising from Barcode Multiplets. In our testing, removal of this data results in the loss of a small percentage (1-5%) of the total sequencing reads in the experiment as the vast majority of cells being removed represent low-quality artifactual cells with the fewest number of sequencing reads.
  • Stand-alone software is available immediately to customers by contacting our Software Support team at [email protected].
  • Our next release of Cell Ranger ATAC (v1.2), available in late November 2019, will incorporate this algorithm into the pipeline.

At 10x, we strive to provide the highest quality data with our products. We greatly appreciate your support and collaboration. Should you have any questions regarding this notification, please feel free to contact your Account Executive, Field Application Scientist, or email our Support Team at [email protected].

Jonathan Schimmel
VP, Global Support and Sales Operations