Plotting heatmap overlays

Import dmslogo along with the other Python packages used in this example:

# NBVAL_IGNORE_OUTPUT

import matplotlib.pyplot as plt

import numpy

import palettable

import pandas as pd

import dmslogo
import dmslogo.colorschemes

We will make an overlay with two properties shown at a site level. Make a simple data frame that gives these site values (note how the overlay values are the same for each letter at each site, since the are site level overlays):

data = pd.DataFrame.from_records(
    data=[
        (1, "A", 1, 1, 0),
        (1, "C", 0.7, 1, 0),
        (2, "C", 0.1, 3, 2),
        (2, "D", 1.2, 3, 2),
        (5, "A", 0.4, 2, 4),
        (5, "K", 0.4, 2, 4),
    ],
    columns=["site", "letter", "height", "overlay_val_1", "overlay_val_2"],
)

data

	site	letter	height	overlay_val_1	overlay_val_2
0	1	A	1.0	1	0
1	1	C	0.7	1	0
2	2	C	0.1	3	2
3	2	D	1.2	3	2
4	5	A	0.4	2	4
5	5	K	0.4	2	4

First, we need to convert the numerical values in the data frame to colors. To do that, we create a dmslogo.colorschemes.ValueToColorMap for each overlay value. Note that you define your own matplotlib.colors.Colormap, or use the name of any pre-built matplotlib colormap or any in the larger set provided by palettable.

Note how below we define each color map, use it create a column in data giving the color, and then also draw the scale bar. This last step is necessary if you want to be able to get a matplotlib.Figure that you can use to save the scale bar to show along with the logo plot:

# NBVAL_IGNORE_OUTPUT

print(f"\nHere is scale bar for first overlay:")
map1 = dmslogo.colorschemes.ValueToColorMap(
    minvalue=data["overlay_val_1"].min(),
    maxvalue=data["overlay_val_1"].max(),
    cmap="viridis",
)
map1_fig, _ = map1.scale_bar(orientation="horizontal", label="overlay 1")
display(map1_fig)
plt.close(map1_fig)
data["overlay 1"] = map1.val_to_color(data["overlay_val_1"])

print(f"\nHere is scale bar for second overlay:")
map2 = dmslogo.colorschemes.ValueToColorMap(
    minvalue=data["overlay_val_2"].min(),
    maxvalue=data["overlay_val_2"].max(),
    cmap=palettable.cmocean.sequential.Dense_20.mpl_colormap,
)
map2_fig, _ = map2.scale_bar(orientation="horizontal", label="overlay 2")
display(map2_fig)
plt.close(map2_fig)
data["overlay 2"] = map2.val_to_color(data["overlay_val_2"])

data

Here is scale bar for first overlay:

Here is scale bar for second overlay:

	site	letter	height	overlay_val_1	overlay_val_2	overlay 1	overlay 2
0	1	A	1.0	1	0	#440154	#e6f1f1
1	1	C	0.7	1	0	#440154	#e6f1f1
2	2	C	0.1	3	2	#fde724	#7770d4
3	2	D	1.2	3	2	#fde724	#7770d4
4	5	A	0.4	2	4	#20908c	#360e24
5	5	K	0.4	2	4	#20908c	#360e24

Now first use dmslogo.logo.draw_logo to draw the logo plot without any overlays:

# NBVAL_IGNORE_OUTPUT

fig, axes = dmslogo.draw_logo(
    data=data,
    x_col="site",
    letter_col="letter",
    letter_height_col="height",
    title="no overlay",
)

Now use dmslogo.logo.draw_logo to draw the logo plot with the first overlay:

# NBVAL_IGNORE_OUTPUT

fig, ax = dmslogo.draw_logo(
    data=data,
    x_col="site",
    letter_col="letter",
    letter_height_col="height",
    title="one overlay",
    heatmap_overlays=["overlay 1"],
)

Draw with both overlays:

# NBVAL_IGNORE_OUTPUT

fig, axes = dmslogo.draw_logo(
    data=data,
    x_col="site",
    letter_col="letter",
    letter_height_col="height",
    title="two overlays",
    heatmap_overlays=["overlay 1", "overlay 2"],
)