Question: Which reads are considered for UMI counting by Cell Ranger?
Answer: When counting UMIs, Cell Ranger only considers reads which meet all of the following criteria:
- Has a valid UMI
- Has a valid 10x barcode
- Has a MAPQ of 255
- Confidently assigned to one gene (as shown in the GX tag of the BAM file alignment record)
Starting in Cell Ranger 7.0, by default, Cell Ranger includes exonic and intronic reads for whole transcriptome gene expression analysis. Consequently, the web summary metrics "Reads Mapped Confidently to Transcriptome" and "Reads Mapped Antisense to Gene" will reflect reads mapped confidently to exonic regions as well as intronic regions. For more information please see the Algorithms Overview.
Prior to Cell Ranger v7.0, by default, intronic reads were excluded from final results, but could be added manually using the
--include-introns argument. For more information about analyzing data with intronic reads, please see our technical notes on interpreting introns and antisense reads and interpreting gene expression data with and without intronic reads.
Reads that contribute to UMI counting have an
xf:i:25 tag in the BAM file. Please see our Analysis Guide tutorial, which walks through the process of identifying records in the BAM file that contribute to UMI counting: Navigating 10x Genomics Barcoded BAM Files.