...

An important consideration when analyzing UMI data are the errors introduced into the UMIs themselves during PCR amplification of the original molecule. If these errors are not accounted for and each sequenced UMI is considered to be representative of the original UMI, the number of unique molecules can be significantly overestimated. To account for this, the Deduplicate UMIs task uses an implementation of the UMI-tools algorithm described in Smith et al. 2017. Paired-end read support was further improved by incorporating components of the UMI deduplication tool Connor.  

...

This method is also similar to the default method in the Drop-seq cookbookAlignment Cookbook (Macosko et al. 2015), which collapses UMI barcodes with a Hamming distance of 1. 

This method may output more UMIs than the default behavior as only UMIs within an edit distance of 1 are summarized, whereas UMIs with a greater distance can be linked in the UMI-tools method. For a comparison of the performance of the two approaches, please see the Adjacency (CellRangerCell Ranger) and Directional (UMI-tools) methods described in Smith et al. 2017. 

...

Connor, University of Michigan BRCF Bioionformatics Core https://pypigithub.org/projectcom/umich-brcf-bioinf/Connor/

Cell Ranger Algorithms Overview, 10X Genomics https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/algorithms/overview

Drop-seq Alignment Cookbook v1.2 Jan 2016, James Nemesh, Steve McCarroll’s lab, Harvard Medical School  http://mccarrolllab.com/wp-content/uploads/2016/03/Drop-seqAlignmentCookbookv1.2Jan2016.pdf

Macosko E, Basu A, Satija R, et al. Highly Parallel Genome-wide Expression Profiling of Individual Cells Using Nanoliter Droplets. Cell 2015; 161(5):1202-1214. https://doi.org/10.1016/j.cell.2015.05.002