smftools.plotting.qc_plotting

Functions

plot_read_qc_histograms(adata, outdir, ...)	Plot a grid of QC histograms: rows = samples (from sample_key), columns = obs_keys.
plot_reference_barcode_chimera_rate(obs, ...)	Summarize the deaminase PCR-chimera rate per reference and barcode.

smftools.plotting.qc_plotting.plot_read_qc_histograms(adata, outdir, obs_keys, sample_key, bins=60, clip_quantiles=(0.0, 0.995), min_non_nan=10, rows_per_fig=6, topn_categories=15, figsize_cell=(3.6, 2.6), dpi=150)

Plot a grid of QC histograms: rows = samples (from sample_key), columns = obs_keys.

Numeric columns -> histogram per sample. Categorical columns -> bar chart of top categories per sample.

Saves paginated PNGs to outdir.

Parameters

adataAnnData

AnnData object with the columns in obs_keys/sample_key in .obs.

outdirstr

Directory to write output figures into.

obs_keyslist[str]

Columns in adata.obs to plot, one column of panels per key.

sample_keystr

Column in adata.obs defining rows (samples/barcodes).

binsint

Histogram bins for numeric metrics.

clip_quantilestuple or None

Clip numeric data globally per metric for consistent axes, e.g. (0.0, 0.995).

min_non_nanint

Minimum finite values to plot a panel.

rows_per_figint

Number of samples per page.

topn_categoriesint

For categorical metrics, show top-N categories (per sample).

figsize_cell(float, float)

Size of each subplot cell (width, height).

dpiint

Figure resolution.

smftools.plotting.qc_plotting.plot_reference_barcode_chimera_rate(obs, outdir, *, barcode_column='barcode', reference_column='Reference_strand', min_events_per_span=3, min_segment_purity=0.9, max_single_strand_fraction=0.8, filename='reference_barcode_chimera_rate.png', dpi=150)

Summarize the deaminase PCR-chimera rate per reference and barcode.

Applies the same strand-switch thresholds used by the preprocess labeling step to the raw per-read metrics (ct_event_count, ga_event_count, strand_segment_purity), then plots a reference x barcode heatmap of the fraction of reads flagged as chimeric (cells annotated with rate and read count). A *.csv companion table is written alongside the PNG.

Parameters:

• obs -- DataFrame with the strand-switch metric columns plus barcode_column and reference_column (e.g. the raw spine obs or ragged frame).

• outdir -- Directory to write the plot and summary table into.

• barcode_column (default: 'barcode') -- Column holding the per-read barcode/sample label.

• reference_column (default: 'Reference_strand') -- Column holding the mapped reference (strand) label.

• min_events_per_span (default: 3) -- Minimum C->T and G->A events required on each side.

• min_segment_purity (default: 0.9) -- Minimum best two-segment purity to accept a clean switch.

• max_single_strand_fraction (default: 0.8) -- Maximum one-sidedness for a read to qualify.

• filename (default: 'reference_barcode_chimera_rate.png') -- Output PNG filename.

• dpi (default: 150) -- Figure resolution.

Returns:

Path to the written PNG, or None if required columns are missing or there is nothing to plot.

Return type:

str | None