Tutorial about transforming LocData

Locan provides methods for transforming localization data sets into new LocData objects.

from pathlib import Path

%matplotlib inline

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

import locan as lc

/tmp/ipykernel_2055/1692805095.py:6: DeprecationWarning: 
Pyarrow will become a required dependency of pandas in the next major release of pandas (pandas 3.0),
(to allow more performant data types, such as the Arrow string type, and better interoperability with other libraries)
but was not found to be installed on your system.
If this would cause problems for you,
please provide us feedback at https://github.com/pandas-dev/pandas/issues/54466
        
  import pandas as pd

lc.show_versions(system=False, dependencies=False, verbose=False)

Locan:
   version: 0.20.0.dev41+g755b969

Python:
   version: 3.11.6

Spatially randomize a structured set of localizations

Assume that localizations are somehow structured throughout a region. Often it is helpful to compare analysis results to a similar dataset in which localizations are homogeneously Poisson distributed. A LocData object with this kind of data can be provided by the randomize function.

rng = np.random.default_rng(seed=1)

locdata = lc.simulate_Thomas(parent_intensity=1e-5, region=((0, 1000), (0, 1000)), cluster_mu=100, cluster_std=10, seed=rng)
locdata.print_summary()

Jupyter environment detected. Enabling Open3D WebVisualizer.
[Open3D INFO] WebRTC GUI backend enabled.
[Open3D INFO] WebRTCWindowSystem: HTTP handshake server disabled.
identifier: "1"
comment: ""
source: SIMULATION
state: RAW
element_count: 705
frame_count: 0
creation_time {
  2024-03-14T11:10:11.039092Z
}

locdata_random = lc.randomize(locdata, hull_region='bb', seed=rng)

fig, ax = plt.subplots(nrows=1, ncols=2)
locdata.data.plot.scatter(x='position_x', y='position_y', ax=ax[0], color='Blue', label='locdata')
locdata_random.data.plot.scatter(x='position_x', y='position_y', ax=ax[1], color='Blue', label='locdata')
plt.tight_layout()
plt.show()

print('Area of bounding box for structured data: {:.0f}'.format(locdata.properties['region_measure_bb']))
print('Area of bounding box for randomized data: {:.0f}'.format(locdata_random.properties['region_measure_bb']))
print('Ratio: {:.4f}'.format(locdata_random.properties['region_measure_bb'] / locdata.properties['region_measure_bb']))

Area of bounding box for structured data: 595088
Area of bounding box for randomized data: 592451
Ratio: 0.9956

Regions other from bounding box can be specified as RoiRegion instance.

region = lc.ConvexHull(locdata.coordinates).region
locdata_random = lc.randomize(locdata, hull_region=region)

fig, ax = plt.subplots(nrows=1, ncols=2)
locdata.data.plot.scatter(x='position_x', y='position_y', ax=ax[0], color='Blue', label='locdata')
locdata_random.data.plot.scatter(x='position_x', y='position_y', ax=ax[1], color='Blue', label='locdata')
plt.tight_layout()
plt.show()

print('Area of bounding box for structured data: {:.0f}'.format(locdata.properties['region_measure_bb']))
print('Area of bounding box for randomized data: {:.0f}'.format(locdata_random.properties['region_measure_bb']))
print('Ratio: {:.4f}'.format(locdata_random.properties['region_measure_bb'] / locdata.properties['region_measure_bb']))

Area of bounding box for structured data: 595088
Area of bounding box for randomized data: 558256
Ratio: 0.9381

Apply an affine transformation to localization coordinates

A wrapper function provides affine transformations based on either numpy or open3d methods.

matrix = ((-1, 0), (0, -1))
offset = (10, 10)
pre_translation = (100, 100)

locdata_transformed = lc.transform_affine(locdata, matrix, offset, pre_translation, method='numpy')

fig, ax = plt.subplots()
locdata.data.plot.scatter(x='position_x', y='position_y',color='Blue', label='locdata', ax=ax)
locdata_transformed.data.plot.scatter(x='position_x', y='position_y', color='Red', label='locdata', ax=ax);

Apply a BunwarpJ transformation to localization coordinates

Often a transformation matrix was computed using ImageJ. The bunwarp function allows applying a transformation from the raw matrix of the ImageJ/Fiji plugin BunwarpJ. Here we show a very small region with a single fluorescent bead that is recorded on a red and a green dSTORM channel.

matrix_path = lc.ROOT_DIR / 'tests/test_data/transform/BunwarpJ_transformation_raw_green.txt'
locdata_green = lc.load_asdf_file(path=lc.ROOT_DIR /
                                 'tests/test_data/transform/rapidSTORM_beads_green.asdf')
locdata_red = lc.load_asdf_file(path=lc.ROOT_DIR /
                                 'tests/test_data/transform/rapidSTORM_beads_red.asdf')

locdata_green_transformed = lc.bunwarp(locdata=locdata_green, matrix_path=matrix_path, pixel_size=(10, 10), flip=True)

fig, ax = plt.subplots()
locdata_red.data.plot.scatter(x='position_x', y='position_y',color='Red', label='locdata_red', alpha=0.5, ax=ax)
locdata_green_transformed.data.plot.scatter(x='position_x', y='position_y', color='Green', label='locdata_green_transformed', alpha=0.5, ax=ax)
locdata_green.data.plot.scatter(x='position_x', y='position_y',color='Blue', label='locdata_green', alpha=0.5, ax=ax);

fig, ax = plt.subplots(figsize=(10, 8))
lc. render_2d_rgb_mpl([locdata_red, locdata_green_transformed, locdata_green], bin_size=5, bin_range=((200, 1000), (600, 1400)), rescale='equal', ax=ax);