Import modules:

import ratioimage.ratio_images as ri
import ratioimage.plot_results as rp

Load in images and segmentations (label images):

C0_imgs = ri.create_img_dict_from_folder(load_pattern="data/PercevalHR_data/imgs/*C0.tiff")
C1_imgs = ri.create_img_dict_from_folder(load_pattern="data/PercevalHR_data/imgs/*C1.tiff")
lab_imgs = ri.create_img_dict_from_folder(
    load_pattern="data/PercevalHR_data/segs/*.tif", label=True
)

C0_imgs.keys()

dict_keys(['a1g01', 'a1g02', 'a1g03', 'a1g04', 'a2g01', 'a2g02', 'a2g03', 'a2g04'])

Display example pair of ratio images + segmentation using plotting function from clonedetective.

from clonedetective.utils import plot_new_images

img_name = "a1g01"

plot_new_images(
    [C0_imgs[img_name], C1_imgs[img_name], lab_imgs[img_name]],
    ["C0 channel", "C1 channel", "label image (segmentation)"],
    interpolation="none",
)

Measure regionprops of images and calculate ratio of intensities. Output is a tidy dataframe

df = ri.ratiometric_measure_region_props_to_tidy_df(C1_imgs, C0_imgs, lab_imgs)
df = ri.split_img_key_col_to_sample_id_and_replicates(
    df, sample_id_colname="genotype", split_regex=r"g"
)

df = ri.sample_id_to_categories(
    df, sample_id_colname="genotype", old_to_new_sample_ids={"a1": "ctrl", "a2": "mut"}
)

df.head()

	label	area	mean_intensity_num	centroid-0	centroid-1	img_key	mean_intensity_denom	ratio_mean_int	genotype	rep_id
0	1	342	0.017960	7.400585	295.084795	a1g01	0.010094	1.779309	ctrl	01
1	2	200	0.012017	35.585000	80.495000	a1g01	0.006473	1.856478	ctrl	01
2	3	269	0.013259	79.152416	174.858736	a1g01	0.027961	0.474173	ctrl	01
3	4	213	0.013988	76.201878	282.995305	a1g01	0.019588	0.714101	ctrl	01
4	5	546	0.025333	86.560440	322.816850	a1g01	0.013664	1.854012	ctrl	01

Plot results - aggregated per cell:

import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(1.75, 2.5))

rp.tidy_create_strip_box_plot(
    x="genotype",
    y="ratio_mean_int",
    data=df,
    ax=ax,
    y_axis_limit=4,
    y_label="PercevalHR 488/405\n(Relative ATP/ADP)",
    palette=["#95a5a6", "#95a5a6"],
)

Plot results - aggregated per image/replicate:

grouped_df = (
    df.groupby(["genotype", "rep_id"]).agg({"ratio_mean_int": "mean"}).reset_index()
)

fig, ax = plt.subplots(figsize=(1.75, 2.5))

rp.tidy_create_swarm_box_plot(
    x="genotype",
    y="ratio_mean_int",
    ax=ax,
    data=grouped_df,
    y_axis_limit=2.5,
    y_label="PercevalHR 488/405\n(Relative ATP/ADP)",
    palette=["#95a5a6", "#95a5a6"],
)

Create ratiometric images:

ratio_imgs = ri.create_dict_of_ratio_images(C1_imgs, C0_imgs, lab_imgs)

ratio_imgs.keys()

dict_keys(['a1g01', 'a1g02', 'a1g03', 'a1g04', 'a2g01', 'a2g02', 'a2g03', 'a2g04'])

plot_new_images(
    [ratio_imgs['a1g01'], ratio_imgs['a2g01']],
    ["a1g01 ratio image", "a2g01 ratio image"],
    img_cmap="magma",
    vmax=2.5,
    figure_shape=(1, 2),
    figure_size=(10, 5),
    colorbar=True,
    colorbar_title="PercevalHR 488/405"
)