Generating one legend for the cell-type colours in subplots generated by scv.pl.scatter

[JesseRop]

Dear developers,

Many thanks for this great tool.

Kindly inform on how I can get one legend for colour coding of the several subplots generated in this tutorial example

kwargs = dict(frameon=False, size=10, linewidth=1.5,
              add_outline='Ngn3 high EP, Pre-endocrine, Beta')

scv.pl.scatter(adata, df['Ngn3 high EP'][:5], ylabel='Ngn3 high EP', **kwargs)
scv.pl.scatter(adata, df['Pre-endocrine'][:5], ylabel='Pre-endocrine', **kwargs)

thanks,
Jesse

[WeilerP]

@JesseRop, a bit hacky but how about

import matplotlib.pyplot as plt

kwargs = dict(frameon=False, size=10, linewidth=1.5, add_outline='Ngn3 high EP, Pre-endocrine, Beta')

fig, ax = plt.subplots(ncols=3)

scv.pl.scatter(adata, df['Ngn3 high EP'][0], ylabel='Ngn3 high EP', ax=ax[0], size=10, linewidth=1.5, add_outline='Ngn3 high EP, Pre-endocrine, Beta')
plt.setp(ax[0].spines.values(), visible=False)
ax[0].get_legend().remove()
ax[0].xaxis.set_visible(False)
ax[0].tick_params(left=False, labelleft=False, right=False)

scv.pl.scatter(adata, df['Ngn3 high EP'][1], ax=ax[1], **kwargs)
ax[1].get_legend().remove()

scv.pl.scatter(adata, df['Ngn3 high EP'][2], legend_loc='right', ax=ax[2], **kwargs)

[JesseRop]

Many thanks for sharing this solution. I have tried this but only the first subplot gets plotted and the other 2 subplots are blank. What I was aiming for is for the 3 subplots to only have one legend rather than repeating the legend for all the subplots. Reproducible code below.

Generating the velocities and the table with genes to plot

import scvelo as scv

adata = scv.datasets.pancreas()

scv.pp.filter_and_normalize(adata, min_shared_counts=20, n_top_genes=2000)
scv.pp.moments(adata, n_pcs=30, n_neighbors=30)

scv.tl.velocity(adata)

scv.tl.rank_velocity_genes(adata, groupby='clusters', min_corr=.3)

df = scv.DataFrame(adata.uns['rank_velocity_genes']['names'])

The plot that with the legend for the cluster colors (Here the legend is repeated for all the plots: My goal is to have only one legend for all the subplots)

scv.pl.scatter(adata, df['Ngn3 high EP'][:5], ylabel='Ngn3 high EP', **kwargs, color='clusters', legend_loc='right')

The solution that you have provided above which only produces the first subplot.

import matplotlib.pyplot as plt

kwargs = dict(frameon=False, size=10, linewidth=1.5, add_outline='Ngn3 high EP, Pre-endocrine, Beta')

fig, ax = plt.subplots(ncols=3)

scv.pl.scatter(adata, df['Ngn3 high EP'][0], ylabel='Ngn3 high EP', ax=ax[0], size=10, linewidth=1.5, add_outline='Ngn3 high EP, Pre-endocrine, Beta')
plt.setp(ax[0].spines.values(), visible=False)
ax[0].get_legend().remove()
ax[0].xaxis.set_visible(False)
ax[0].tick_params(left=False, labelleft=False, right=False)

scv.pl.scatter(adata, df['Ngn3 high EP'][1], ax=ax[1], **kwargs)
ax[1].get_legend().remove()

scv.pl.scatter(adata, df['Ngn3 high EP'][2], legend_loc='right', ax=ax[2], **kwargs)

Versions

scv.logging.print_versions()

scvelo==0.2.4  scanpy==1.8.2  anndata==0.7.8  loompy==3.0.6  numpy==1.21.2  scipy==1.7.1  matplotlib==3.4.3  sklearn==0.24.2  pandas==1.3.2  

Thanks

[WeilerP]

@JesseRop, would you please check that is isn't simply a displaying issue? You can check if all subplots are shown when passing show=False to the first two calls of scv.pl.scatter and/or save the figure using

fig.savefig("figure.png")

to check if all three subplots are displayed.

[JesseRop]

It works with your suggestions. Thanks!