Input: public 10x PBMC 3k dataset (example run) · generated by the free QC tool at omics.hstgenomics.com ·
the input file was deleted once this report was built.
2,700cells / barcodes
32,738genes
817median genes / cell
2,197median UMIs / cell
2.03%median mitochondrial
6.6%removed by suggestion
Suggested thresholds for this dataset
Filter
Suggested
Common tutorial default
Minimum genes per cell
232
200
Maximum genes per cell
1,402
2,500
Maximum mitochondrial %
5%
5%
Cells retained
2,523 of 2,700
Suggestions are median ± 3 MAD of this dataset's own distributions, floored at 200 genes.
They are a starting point; tissue biology beats any formula.
Things worth a second look
No structural problems detected in these metrics.
Expected doublets from loading: about 58 barcodes
(2.1%). A gene-count ceiling will not remove them — a doublet detector will.
Distributions
Per-cell distributions with the suggested cut-offs marked.UMIs against mitochondrial fraction — dying cells sit low-count / high-mito.Barcode rank plot: a sharp knee means cells and empty droplets separate cleanly.
Most abundant genes
The top 10 genes carry 14.2% of all counts.
Genes carrying the largest share of counts.
Gene
% of all counts
MALAT1
2.53%
TMSB4X
1.94%
B2M
1.9%
RPL10
1.39%
RPL13
1.21%
RPL13A
1.2%
FTL
1.17%
RPS2
1.02%
RPS6
0.98%
FTH1
0.9%
What this does not cover
Doublet detection, ambient-RNA correction, clustering and annotation are not part of this free report.
Medians hide distributions; a real assessment looks per sample and at the joint distributions.
