Single-Cell RNA

Barcode Rank Plot

The barcode rank plot (often referred to as the “knee plot”), orders barcodes based on the number of molecules associated with them. Typically, the cell barcodes are concentrated at the top of the rank plot, whereas the background barcodes are concentrated in the lower portion of the plot. The purpose of the cell calling is to find a point in the first “knee” area that separates the cells from the background.

A more negative slope and a poorly defined knee can be indicative of lower quality samples. It is not recommended to call cells too far down into the transition zone (area between first knee and second knee) as this risks adding compromised cells and background noise.

• Homogenous sample types and cells should be called towards the upper part of the transition zone

• More heterogenous sample types should be called ⅓ or ½ way down the transition zone

Single Cell RNA Metrics

Metric	Description
Total Input Reads	Total number of reads successfully sequences and demultiplexed for the sample. This includes reads missing barcodes, reads with exactly matching barcodes, reads with corrected barcodes, and reads with non-matching barcodes.
% Mapped	Percentage of reads that are mapped to the genome. DRAGEN 4.4.4 and 4.4.5 count all reads that are mapped to the reference to determine % Mapped. This includes unique exon matching reads unique intron matching reads filtered antisense reads filtered ambiguously matching reads filtered low MAPQ reads mitochondrial reads Filtered reads above are discarded from downstream analysis (barcode matching, counting, etc.)
Total Barcoded Reads	Number of reads with barcodes that match the whitelist. This includes reads with exactly matching barcodes and reads with corrected barcodes (either an exact match or within 1 edit distance away for each of the four barcode tiers)
Passing Cells	Total unique barcodes that pass the count threshold for a passing cell
% Reads in Passing Cells	Number of reads in passing barcodes divided by the total number of barcoded reads in all cells
Mean Reads per Cell*	Mean reads per cell (Total gene reads / Passing cells). Generally, should fall within range of 20,000-40,000 reads per (called) cell at minimum to ensure sufficient read depth. If mean reads per cell is lower than 20,000, it might be necessary to sequence at a higher read depth.
Median Reads per Cell	The median number of reads per passing barcode
Median Molecules per Cell	The median number of captured RNAs per passing barcode. A molecule represents an individual transcript that was captured onto a PIP, was converted into a cDNA library molecule (with one or more daughter molecules during fragmentation), sequenced, then properly barcoded, mapped, and IMI-corrected.
Median Genes per Cell	Median number of unique gene assignments observed across all molecules in the passing cell fraction
% Sequencing Saturation	Represents the probability that adding additional sequenced read will have already been observed. The higher the sequencing saturation, the less benefit of uncovering new information with additional sequencing reads.

*Only available for analyses with PIPseq CRISPR mode enabled

Extended Single-Cell RNA Metrics

This section is only available for analyses using DRAGEN 4.4.4+. with PIPseq CRISPR mode enabled.

Metric	Description
Input Gene Expression Reads*	Total input gene expression reads
Input Feature Reads*	Total input feature reads
Barcoded Gene Expression Reads*	Total number of gene expression reads with barcode matching whitelist after error-correction. This includes reads with exactly matching barcodes and reads with corrected barcodes (either an exact match or within 1 edit distance away for each of the four barcode tiers)
Barcoded Feature Reads*	Total number of input feature reads with barcode matching whitelist after error-correction. This includes reads with exactly matching barcodes and reads with corrected barcodes (either an exact match or within 1 edit distance away for each of the four barcode tiers)
Unique Exon Reads	Number of reads that are mapped to exons in genes.
Unique Intron Reads	Number of reads that are mapped to introns in genes
Filtered Antisense Reads	Number of filtered reads that are mapped to the opposite strand of a gene
Filtered Ambiguous Reads	Number of filtered reads that map to multiple genes equally well
Filtered Low MAPQ Reads	Number of filtered reads below the MAPQ threshold (default: 4)
Filtered Non-Matching Reads	Number of filtered reads that do not map to any genes
Mitochondrial Reads	Number of reads that map to mitochondria
Total Gene Reads	Total number of gene expression reads. Sum of the exon, intron, antisense, ambiguously matching, low MAPQ, mitochondrial, and non-matching reads
Counted Gene Reads	Total number of gene reads counted
Molecules	Total molecules
Genes Detected	Total genes detected

*Only available for analyses with PIPseq CRISPR mode enabled

Barcode Summary Metrics

Metric	Description
Cell Type	Barcode classification category. “PASS” for passing cells or “LOW” for background cells.
Total Gene Reads	Total number of genic reads in the category
Molecules	Total number of molecules in the category
Genes	Total number of unique genes detected for each cell in the category
Mitochondrial Reads	Total number of mitochondrial reads in the category. This is based on gene names that include prefixes like “MT-” and may not work with every mapping reference.
Feature Molecules*	Total number of CRISPR or other target molecules in the category
Features*	Total number of unique CRISPR sequences or other targets detected in the category

*Only available for analyses with PIPseq CRISPR mode enabled

Feature Metrics

This section is only available for analyses with PIPseq CRISPR mode enabled.

Metric	Description
Input Feature Reads	Total number of input feature reads
Barcoded Feature Reads	Total number of input feature reads with barcode matching whitelist after error-correction. This includes reads with exactly matching barcodes and reads with corrected barcodes (either an exact match or within 1 edit distance away for each of the four barcode tiers)
Feature Matching Reads	Number of reads that match to a feature barcode reference sequence
Feature Reads	Total Feature Reads.
Feature Molecules	Number of barcoded feature reads across all barcodes in the dataset (cells and background). If features do not have MIs, then this will equal the number of feature matching reads (a.k.a. counted feature reads)
Median Feature Reads per Cell	Median number of feature matching reads in passing cells
Feature Molecules per Cell	Median feature molecules per passing cell
Features per Cell	Median of the number of features per passing cell over all of the passing cells
Features Detected	Total number of different features detected in all passing cells
CRISPR Reads	Total number of CRISPR reads with barcode matching whitelist after error-correction
% Mapped Features	Percentage of CRISPR tags mapped
% Features in Cells	Percentage of valid CRISPR tags in cells
% Cells with Features	Percentage passing cells with CRISPR reads