import matplotlib
matplotlib.use("Agg") # forces matplotlib to not launch X11 window
import matplotlib.pyplot as plt
import matplotlib as mpl
import numpy as np
import seaborn as sns
import pandas as pd
import cv2
from scipy import stats
from matplotlib.patches import Patch
from matplotlib.lines import Line2D
__all__ = ["focus_plot"]
def make_scatter(twas_df):
"""
Make a scatterplot of zscore values with gene names as xtick labels.
:param twas_df: pandas.DataFrame containing at least zscores and gene-names
:return: numpy.ndarray (RGB) formatted scatterplot of zscores
"""
mpl.rcParams["figure.figsize"] = [6.4, 4.8]
fig, ax = plt.subplots()
size_arr = []
color_arr = []
custom_palette = ["#e4f1fe", "#bdd7e7", "#6baed6", "#3182bd", "#08519c"]
size_palette = [2, 4, 8, 10, 12]
for i, row in twas_df.iterrows():
pip = row["pip"]
if pip < 0.2:
color_arr.append(custom_palette[0])
size_arr.append(size_palette[0])
elif 0.2 <= pip < 0.40:
color_arr.append(custom_palette[1])
size_arr.append(size_palette[1])
elif 0.40 <= pip < 0.60:
color_arr.append(custom_palette[2])
size_arr.append(size_palette[2])
elif 0.60 <= pip < 0.80:
color_arr.append(custom_palette[3])
size_arr.append(size_palette[3])
else:
color_arr.append(custom_palette[4])
size_arr.append(size_palette[4])
n_rows = len(twas_df.index)
x_values = np.arange(1, n_rows + 1)
ax.scatter(x=x_values, y=twas_df["logp"].values, s=size_arr, c=color_arr, edgecolor="black")
# create legend
legend_elements = [
Line2D([0], [0], marker="o", color="w", label="[0.0, 0.2)",
markerfacecolor=custom_palette[0], markersize=size_palette[0], markeredgecolor="k"),
Line2D([1], [1], marker="o", color="w", label="[0.2, 0.4)",
markerfacecolor=custom_palette[1], markersize=size_palette[1], markeredgecolor="k"),
Line2D([2], [2], marker="o", color="w", label="[0.4, 0.6)",
markerfacecolor=custom_palette[2], markersize=size_palette[2], markeredgecolor="k"),
Line2D([3], [3], marker="o", color="w", label="[0.6, 0.8)",
markerfacecolor=custom_palette[3], markersize=size_palette[3], markeredgecolor="k"),
Line2D([4], [4], marker="o", color="w", label="[0.8, 1.0]",
markerfacecolor=custom_palette[4], markersize=size_palette[4], markeredgecolor="k")]
plt.legend(handles=legend_elements, loc="best", title="PIP")
n_rows = len(twas_df.index)
gene_names = twas_df["mol_name"].values
plt.xticks(np.arange(1, n_rows + 1, 1.0))
plt.xticks(x_values, labels=gene_names, rotation="vertical")
plt.subplots_adjust(bottom=0.25) # make room for xlabel
plt.ylabel("$-\log_{10}(p)$", fontsize=18)
plt.xlabel("")
# drop right/top axis bars
ax.spines["right"].set_visible(False)
ax.spines["top"].set_visible(False)
fig.tight_layout()
fig.canvas.draw()
# save as numpy.ndarray format
data = np.fromstring(fig.canvas.tostring_rgb(), dtype=np.uint8, sep="")
scatter_plot = data.reshape(fig.canvas.get_width_height()[::-1] + (3,))
return scatter_plot
def heatmap(wcor):
"""
Make a scatterplot of zscore values with gene names as xtick labels.
:param wcor: numpy.ndarray matrix of sample correlation structure for predicted expression
:return: numpy.ndarray (RGB) formatted heatmap of correlation structure
"""
mpl.rcParams["figure.figsize"] = [6.4, 6.4]
fig = plt.figure()
fig.subplots_adjust(bottom=0.20, left=0.28)
mask = np.zeros_like(wcor, dtype=np.bool)
mask[np.triu_indices_from(mask)] = True
ax = sns.heatmap(wcor, mask=mask, cmap="RdBu_r", square=True,
linewidths=0, cbar=False, xticklabels=False, yticklabels=False, ax=None,
vmin=-1, vmax=1)
ax.margins(2)
ax.set_aspect("equal", "box")
fig.canvas.draw()
# save image as numpy array
data = np.fromstring(fig.canvas.tostring_rgb(), dtype=np.uint8, sep="")
img = data.reshape(fig.canvas.get_width_height()[::-1] + (3,))
# rotate heatmap to make upside-down triangle shape
rows, cols, ch = img.shape
M = cv2.getRotationMatrix2D((cols / 2, rows / 2), 45, 1)
dst = cv2.warpAffine(img, M, (cols, rows), borderMode=cv2.BORDER_CONSTANT,
borderValue=(255, 255, 255))
# trim extra whitespace
crop_img = dst[int(dst.shape[0] / 2.5):int(dst.shape[0] / 1.1)]
return crop_img
def heatmap_colorbar():
"""
Make a colorbar legend for correlation heatmap for range [-1,1].
:param: None
:return: numpy.ndarray (RGB) formatted colorbar for range [-1,1]
"""
fig = plt.figure(figsize=(3.0, 1.0))
ax1 = fig.add_axes([0.05, 0.80, 0.9, 0.15])
norm = mpl.colors.Normalize(vmin=-1, vmax=1)
cmap = mpl.cm.get_cmap("RdBu_r")
mpl.colorbar.ColorbarBase(ax1, cmap=cmap, norm=norm, orientation="horizontal", ticks=[-1, -.50, 0, .50, 1])
# convert to numpy array format
fig.canvas.draw()
data = np.fromstring(fig.canvas.tostring_rgb(), dtype=np.uint8, sep="")
colorbar = data.reshape(fig.canvas.get_width_height()[::-1] + (3,))
# reshape and fill to match width of heatmap
new_size = (colorbar.shape[0], colorbar.shape[1])
desired_size_w = 640
desired_size_h = 100
delta_w = desired_size_w - new_size[1]
delta_h = desired_size_h - new_size[0]
top, bottom = delta_h // 2, delta_h - (delta_h // 2)
left, right = delta_w // 2, delta_w - (delta_w // 2)
left += 18 # make colorbar line up with heatmap
right -= 18
# fill in extra space with whitespace
color = [255, 255, 255]
colorbar = cv2.copyMakeBorder(colorbar, top, bottom, left, right, cv2.BORDER_CONSTANT, value=color)
return colorbar
def focus_plot(wcor, twas_df):
"""
Plot zscores and local correlation structure for TWAS fine-mapping.
:param twas_df: pandas.DataFrame containing at least zscores and gene-names
:param wcor: numpy.ndarray matrix of sample correlation structure for predicted expression
:return: matplotlib.figure.Figure object containing plot of zscores and local correlation heatmap
"""
# filter out the null model
twas_df = twas_df[twas_df["ens_gene_id"] != "NULL.MODEL"]
# add p-value to compute -log10 p
twas_df = twas_df.assign(logp=-stats.chi2.logsf(twas_df["twas_z"].values ** 2, 1))
scatter_plot = make_scatter(twas_df)
crop_img = heatmap(wcor)
colorbar = heatmap_colorbar()
# combine plots
numpy_vertical_concat = np.concatenate((scatter_plot, crop_img), axis=0)
numpy_vertical_concat = np.concatenate((numpy_vertical_concat, colorbar), axis=0)
numpy_vertical_concat = cv2.resize(numpy_vertical_concat, (0, 0), fx=2.5, fy=2.5)
fig = plt.figure()
plt.imshow(numpy_vertical_concat)
plt.title("")
plt.axis("off")
plot_arr = [fig]
plt.close("all")
return plot_arr
