peak discrepancy about sierra

[rj-patrick]

Hi @cag1343,

Apologies for the slow response. Regarding the cell numbers, do you mean that there were originally 170,706 cells, and this is retained with NewPeakSeurat, but after running the PeakSeuratFromTransfer you are left with 56902? I'm assuming you are working with multiple runs/experiments that have been combined together? The likely explanation for the 'missing' cells is that there is a discrepancy with the cell barcode suffix. Sierra by default follows the CellRanger strategy of appending '-1', '-2' etc. to cell barcodes according to the order of datasets that are provided to the AggregatePeakCounts function. If this results in suffixes that are different to what is in your Seurat object those cells will be excluded.

Regarding peak numbers, 394 peaks is definitely unexpected. Are you using the same reference file that was used for the alignment of the BAM files and that generated the 17,473 features?

Cheers,
Ralph