Render
This module performs the extraction of useful metrics, and uses them to generate a series of summary plots, in PNG, HTML or PDF format.
Usage
The minimal command is cl render --input-path <PATH> --output-path <PATH>
, where input-path
is the path to the AVITI data folder and output-path is the directory where the plots will be placed.
Options
--stats-json
: Path to the cell2stats 'RunStats.json' file. Required only if different from<INPUT_PATH>/RunStats.json
.--panel-json
: Path to the cell2stats 'Panel.json' file. Required only if different from<INPUT_PATH>/Panel.json
.--raw-parquet
: Path to the cell2stats 'RawCellStats.parquet' file. Required only if different from<INPUT_PATH>/RawCellStats.parquet
.--wells
: Comma-separated list of wells to include in the analysis. By default, all wells will be used.--batches
: Comma-separated list of batches to include in the analysis. By default, all batches will be used.--plot-list
: Comma-separated list of plots to generate. Accepted values include 'All', 'BatchWell', 'Well', 'Count', 'Correlation', 'UMAP'. Default: All.--format
: It can be one of 'html', 'png' or 'pdf'. Default is 'png'.
Outputs
Barcoding performance metrics
Groups of barcodes are sequenced in serial batches, where each batch is defined by a specific sequencing primer.
Metric | Description | Expected Value |
---|---|---|
PercentAssignedReads | Of all polonies, percentage assigned to an expected barcode | > 70% |
PercentMismatch | Of all polonies assigned to a barcode, percentage assigned with a mismatch | < 35% |
PercentAssignedReads.*
: Percentage of assigned reads.PercentMismatch.*
: Percentage of
Cell segmentation performance metrics
Cell segmentation is performed based on the cell paint images for the cell membrane, nucleus and actin. The metrics below summarize the results of the segmentation process for each well in the flowcell.
Metric | Description | Expected Value |
---|---|---|
PercentConfluency | Fraction of well area occupied by cells | 25-50% (variable based on cell seeding) |
CellCount | Number of objects detected during segmentation | >10,000 (variable based on cell seeding) |
MedianCellDiameter | Approximate median diameter of cells in microns | ~35 um |
PercentNucleatedCells | Fraction of cells with segmented nucleus | > 97% |
PercentConfluency.*
: Wells confluency.CellCount.*
: Cell counts.MedianCellDiameter.*
: Cell diameters.
Cell assignment metrics
After barcoding and cell segmentation is complete, individual barcodes are assigned to cells. The metrics below summarize this process.
Metric | Description | Expected Value |
---|---|---|
AssignedCountsPerMM2 | Number of assigned polonies per mm2 of cell area | ~150,000 (protein), 200,000-300,000 (RNA) |
AssignedCountsPerMM2.*
: Counts per mm^2.
Correlation metrics
For each pair of wells, we can calculate the correlation of log-transformed average counts as a measure of reproducibility. For both RNA and protein data types, replicates wells should have R2 > 0.95.
RNADistCorr.*
: RNA counts distance correlation matrix.ProteinDistCorr.*
: Protein counts distance correlation matrix.
UMAP projection and cell state
The single-cell data can be used to generate a UMAP projection. Across the entire flowcell, the UMAP projection should separate the data into two well-defined clusters, with the two different cell types comprising each cluster. Within each cluster, a circular pattern should exist, corresponding to the annotated cell state.
WellLabelUMAP.*
: UMAP coloured by labels.phaseUMAP.*
: UMAP coloured by cell state.