Question: How does Cell Ranger process and filter UMIs?
Answer: Since UMIs are random, there is no whitelist to check them against. However, Cell Ranger performs quality filtering and correction for UMI sequencing errors.
The following criteria are used for UMI filtering:
- Must not be a homopolymer, e.g. AAAAAAAAAA
- Must not contain N
- Must not contain bases with base quality < 10
UMIs that are 1 Hamming distance (substitution) away from a higher-count UMI are corrected to the higher count UMI if they share a cell barcode.