Plotting Molecules and Cells

Author: Clarence Mah | Last Updated: 6/6/2022

We will showcase the useful subcellular spatial capabilities of bento using the included seqFISH+ dataset.

Note

There are several options available for visualization that are light wrappers building on their seaborn package counterparts. Using the bento.pl.cellplot function, you may create:

• 2D histogram (default): kind = 'hist'

  • https://seaborn.pydata.org/generated/seaborn.histplot.html

• Scatter plot: kind = 'scatter'

  • https://seaborn.pydata.org/generated/seaborn.scatterplot.html

• Hexbin plots: kind = 'hex'

  • https://matplotlib.org/stable/api/_as_gen/matplotlib.pyplot.hexbin.html

Load Libraries

import bento

Load Data

Let’s use the seqFISH+ dataset.

adata = bento.datasets.load_dataset("seqfish")
adata

AnnData object with n_obs × n_vars = 179 × 3726
    obs: 'cell_shape', 'nucleus_shape', 'batch', 'cell_radius', 'nucleus_inner_edge_shape', 'cell_inner_edge_shape', 'nucleus_outer_edge_shape', 'n_detected', 'fraction_detected', 'cell_edge_count', 'cell_edge_fraction', 'cytoplasmic_count', 'cytoplasmic_fraction', 'nuclear_count', 'nuclear_fraction', 'nuclear_edge_count', 'nuclear_edge_fraction', 'none_count', 'none_fraction', 'td_cluster'
    var: 'n_detected', 'fraction_detected', 'cell_edge_count', 'cell_edge_fraction', 'cytoplasmic_count', 'cytoplasmic_fraction', 'nuclear_count', 'nuclear_fraction', 'nuclear_edge_count', 'nuclear_edge_fraction', 'none_count', 'none_fraction', 'td_cluster'
    uns: 'points', 'tensor', 'tensor_labels', 'tensor_loadings'
    layers: 'cell_edge', 'cell_edge_p', 'cell_inner_asymmetry', 'cell_inner_proximity', 'cell_outer_asymmetry', 'cell_outer_proximity', 'cytoplasmic', 'cytoplasmic_p', 'l_half_radius', 'l_max', 'l_max_gradient', 'l_min_gradient', 'l_monotony', 'none', 'none_p', 'nuclear', 'nuclear_edge', 'nuclear_edge_p', 'nuclear_p', 'nucleus_dispersion', 'nucleus_inner_asymmetry', 'nucleus_inner_edge_enrichment', 'nucleus_inner_proximity', 'nucleus_outer_asymmetry', 'nucleus_outer_edge_enrichment', 'nucleus_outer_proximity', 'point_dispersion', 'spliced', 'unspliced'

Plot Molecule Densities

The default cellplot is a 2D histogram limited to the first field of view.

bento.pl.cellplot(adata)

Hexbin

The RNA density can also be viewed as a hexbin plot. The colormap and other underlying parameters can be changed (see matplotlib.pyplot.hexbin for underlying parameters).

bento.pl.cellplot(adata, kind="hex", cmap="magma")

Multiple Subplots

Multiple fields of view can be plotted at once by specifying col="fov", as separate subplots. Field of view is annotated as a column (by default "batch") in adata.obs.

bento.pl.cellplot(adata, fovs="all", col="fov", col_wrap=4, height=4)

Individual Molecules and Cells

Finally, we can plot specific cells and genes by subsetting and slicing the adata object. Here we set kind="scatter" for a scatterplot (see sns.scatterplot for additional parameters) and hue="gene" to color each gene uniquely.

# Randomly choose 10 genes from the top 50 most highly expressed genes
some_genes = adata.to_df().mean().sort_values().tail(50).sample(10).index.tolist()

bento.pl.cellplot(
    adata[:8, some_genes],
    kind="scatter",
    hue="gene",
    col="cell",
    col_wrap=4,
    height=3,
    s=10,
)