Question: Why do I have a high percentage of reads mapping to introns?
Answer: It is common to have reads map to intronic regions. We observe the percentage of reads mapping to intronic regions varies by sample type and sample preparation (single cells vs. nuclei). For example, we observe a higher percentage of reads mapping to introns in certain sample types such as PBMCs (~20-25%), and also in nuclear datasets (46%):
For nuclei samples, you could evaluate the use of a "pre-mRNA" transcriptome reference for cellranger count if you would like to include reads mapping to intronic regions in the UMI count. For more information on creating a pre-mRNA reference package, please see the support site: Making pre-MRNA references.