Hello, While processing the ROSMAP single-cell RNA-seq data, I noticed a discrepancy between the cell barcode counts in the gene-expression files and the demultiplexing file. Using library B7A (191213-B7-A) as an example: Barcode file (syn51121968) has 24,695 unique cell barcodes Demultiplexing file (syn34572333) has 37,542 unique cell barcodes assigned to B7A Barcodes that match in both files: 19,207 Could you please clarify why the demultiplexing file contains substantially more cell barcodes than the gene-expression barcode file? Is this discrepancy due to quality-control filtering or differences in how the files were generated? Thank you for the help.

Created by Mykhailo Zaporozhets mishazap
Hi @mishazap, thanks for your question. The demultiplexing ID mapping file contains only those barcodes for cells that could be confidently mapped to a single individual. Therefore, cells that failed some QC, were ambiguous, or didn't map to an individual are not included. Conversely, barcodes that mapped confidently to a single individual but failed downstream RNA quality control might exist in the demultiplexing file but not in the barcode file. This should explain the discrepancy. Accordingly, please do use only barcodes matching across both files. Hope this helps! Erik Nelson

Discrepancy in the amount of cell barcodes between the count matrices and the demultiplexing file for ROSMAP, snRNAseq - DLPFC, Experiment 2 page is loading…