Skip to main content

Package for analysing SMLM data

Project description

locpix

locpix is a Python library for analysing point cloud data from SMLM.

This project is under active development

This is a short ReadMe just containing a QuickStart guide. For more comprehensive documentation please see https://oubino.github.io/locpix/

locpix includes the following functionality in order they are used in a normal workflow:

  1. Preprocess : Initialises project and converts .csv files representing SMLM data (point cloud) into .parquet files, necessary for this software

  2. Annotate : Generating histograms from the SMLM data and manually annotating these histograms to extract relevant localisations & labelling with seeds for watershed algorithm

  3. Segmentation:

    1. Classic segmentation : Use classic method to segment histograms to extract relevant localisations

    2. Cellpose segmentation : Use Cellpose method to segment histograms to extract relevant localisations

    3. Ilastik segmentation : Use Ilastik method to segment histograms to extract relevant localisations

    4. UNET segmentation : Use UNET to segment histograms to extract relevant localisations

  4. Membrane performance : Performance metrics calculation based on the localisations (not the histograms!)

Project Structure

We assume your input SMLM data are .csv files.

This input data must first be preprocessed into a user chosen project directory, using the Preprocess script. We strongly suggest this project directory is located outside the locpix folder.

The input and output of all further scripts will remain located inside the project directory, the input data folder will not be accessed again!

Usage configuration

Each script needs a configuration file (.yaml file), which should be specified using the -c flag.

Each configuration used will be saved in the project directory.

The templates for the configuration files can be found in the templates folder.

For instructions for the parameters in the configuration files see https://oubino.github.io/locpix/user_guide/templates/index.html

Quickstart

Installation

Prerequisites

You will need anaconda or miniconda or mamba. We recommend mamba

Install

Create an environment and install via pypi

conda create -n locpix-env python==3.10
conda activate locpix-env
pip install locpix

Preprocessing

Preprocess

This script preprocesses the input .csv data for later use AND must be run first.

This script will take in .csv files, and convert them to .parquet files, while also wrangling the data into our data format.

To run the script -i -c and -o flags should be specified

preprocess -i path/to/input/data -c path/to/config/file -o path/to/project/directory

Optionally we can add:

  • -s flag to check the correct files are selected before preprocessing them

  • -p flag if our files are .parquet files not .csv files.

Annotate

This script allows for manual segmentation of the localisations and adding markers to the images as seeds for the watershed algorithm.

To run the script -i and -c flags should be specified

annotate -i path/to/project/directory -c path/to/config/file

Optionally we can add:

  • -m flag to check the metadata of the project before running

  • -r flag to relabel histograms, will assume some labels are already present

Image segmentation

Classic segmentation

Perform classic segmentation on our localisation dataset.

To run the script -i and -c flags should be specified

classic -i path/to/project/directory -c path/to/config/file

Optionally we can add:

  • -m flag to check the metadata of the project before running

Cellpose segmentation

Need to activate extra requirements - these are big and not included in initial install.

Note that if you have a GPU this will speed this up.

Note we modified Cellpose to fit in with our analysis, therefore you need to install our forked repository - note below will clone the Cellpose repository to wherever you are located

If you have a GPU

pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu117
git clone https://github.com/oubino/cellpose
cd cellpose
pip install .

If you don’t have a GPU

pip install pytorch
git clone https://github.com/oubino/cellpose
cd cellpose
pip install .

To evaluate Cellpose model on the localisation dataset without any retraining on your dataset run the script with -i and -c flags specified

cellpose_eval -i path/to/project/directory -c path/to/config/file

Optionally we can add:

  • -m flag to check the metadata of the project before running

  • -o flag to specify folder in project dir to save output (defaults to cellpose_no_train)

  • -u flag to specify a user model to load in

To retrain first then evaluate we instead

Prepare data for training

train_prep -i path/to/project/directory -c path/to/config/file

Optionally we can add:

  • -m flag to check the metadata of the project before running

Train cellpose

cellpose_train -i path/to/project/directory -ct path/to/config/train_file -ce path/to/config/eval_file

Optionally we can add:

  • -m flag to check the metadata of the project before running

UNET segmentation

Need to activate extra requirements - these are big and not included in initial install.

Note that if you have a GPU this will speed this up.

Note this is only needed if haven’t done for cellpose above

If you have a GPU

pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu117

If you don’t have a GPU

pip install pytorch

To train UNET

unet -i path/to/project/directory -c path/to/config/file

Optionally we can add:

  • -m flag to check the metadata of the project before running

Ilastik segmentation

Need to prepare the data for Ilastik segmentation

ilastik_prep -i path/to/project/directory -c path/to/config/file

Optionally we can add:

  • -m flag to check the metadata of the project before running

Then run the data through the Ilastik GUI, which needs to be installed from Ilastik and to run it please see https://oubino.github.io/locpix/user_guide/usage.html#id7

Then convert the output of the Ilastik GUI back into our format

ilastik_output -i path/to/project/directory

Optionally we can add:

  • -m flag to check the metadata of the project before running

Membrane performance

To evaluate membrane performance for a particular method, run below, where method name needs to match where the segmentation files are

membrane_performance_method -i path/to/project/directory -c path/to/config/file -o method_name

Optionally we can add:

  • -m flag to check the metadata of the project before running

To evaluate performance of membrane segmentation from classic, cellpose and ilastik

membrane_performance -i path/to/project/directory -c path/to/config/file

Optionally we can add:

  • -m flag to check the metadata of the project before running

To aggregate the performance over the folds for methods classic, cellpose without training, cellpose with training and ilastik

agg_metrics -i path/to/project/directory

Licenses

Package

License

alabaster 0.7.12

UNKNOWN

app-model 0.1.1

BSD 3-Clause License

appdirs 1.4.4

MIT

arrow 1.2.3

Apache 2.0

astroid 2.12.13

LGPL-2.1-or-later

asttokens 2.2.0

Apache 2.0

attrs 22.1.0

MIT

Babel 2.11.0

BSD

backcall 0.2.0

UNKNOWN

backports.functools-lru-cache 1.6.4

UNKNOWN

beautifulsoup4 4.11.1

MIT

binaryornot 0.4.4

BSD

black 22.12.0

MIT

build 0.9.0

MIT

cachey 0.2.1

BSD

cellpose 2.1.2.dev26+g731fe4e

BSD

certifi 2022.9.24

MPL-2.0

cfgv 3.3.1

MIT

chardet 5.1.0

LGPL

charset-normalizer 2.1.1

MIT

click 8.1.3

BSD-3-Clause

cloudpickle 2.2.0

BSD 3-Clause License

colorama 0.4.6

UNKNOWN

comm 0.1.3

BSD 3-Clause License

commonmark 0.9.1

BSD-3-Clause

contourpy 1.0.6

BSD-3-Clause

cookiecutter 2.1.1

BSD

coverage 6.5.0

Apache 2.0

cycler 0.11.0

BSD

dask 2022.11.1

BSD

debugpy 1.6.4

MIT

decorator 5.1.1

new BSD License

distlib 0.3.6

Python license

docstr-coverage 2.2.0

MIT

docstring-parser 0.15

MIT

docutils 0.17.1

public domain, Python, 2-Clause BSD, GPL 3 (see COPYING.txt)

entrypoints 0.4

UNKNOWN

exceptiongroup 1.0.4

UNKNOWN

executing 1.2.0

MIT

fastremap 1.13.3

LGPLv3

filelock 3.9.0

UNKNOWN

flake8 6.0.0

MIT

fonttools 4.38.0

MIT

freetype-py 2.3.0

UNKNOWN

fsspec 2022.11.0

BSD

HeapDict 1.0.1

BSD

hsluv 5.0.3

MIT

identify 2.5.17

MIT

idna 3.4

UNKNOWN

imagecodecs 2022.9.26

BSD

imageio 2.22.4

BSD-2-Clause

imageio-ffmpeg 0.4.7

BSD-2-Clause

imagesize 1.4.1

MIT

importlib-metadata 6.6.0

UNKNOWN

in-n-out 0.1.6

BSD 3-Clause License

iniconfig 1.1.1

MIT License

ipykernel 6.17.1

UNKNOWN

ipython 8.13.2

BSD-3-Clause

ipython-genutils 0.2.0

BSD

jedi 0.18.2

MIT

Jinja2 3.1.2

BSD-3-Clause

jinja2-time 0.2.0

MIT

joblib 1.2.0

BSD

jsonschema 4.17.3

MIT

jupyter-client 7.4.7

UNKNOWN

jupyter-core 5.1.0

UNKNOWN

kiwisolver 1.4.4

UNKNOWN

lazy-object-proxy 1.8.0

BSD-2-Clause

line-profiler 4.0.2

BSD

llvmlite 0.39.1

BSD

locket 1.0.0

BSD-2-Clause

locpix 0.0.12.dev70+ga7833b4.d20230120

UNKNOWN

magicgui 0.6.1

MIT license

MarkupSafe 2.1.1

BSD-3-Clause

matplotlib 3.6.2

PSF

matplotlib-inline 0.1.6

BSD 3-Clause

mccabe 0.7.0

Expat license

mypy-extensions 0.4.3

MIT License

napari 0.4.17

BSD 3-Clause

napari-console 0.0.6

BSD 3-Clause

napari-locpix 0.0.3

MIT

napari-plugin-engine 0.2.0

MIT

napari-svg 0.1.6

BSD-3

natsort 8.2.0

MIT

nest-asyncio 1.5.6

BSD

networkx 2.8.8

UNKNOWN

nodeenv 1.7.0

BSD

npe2 0.6.1

BSD-3-Clause

numba 0.56.4

BSD

numpy 1.23.5

BSD

numpydoc 1.5.0

BSD

opencv-python-headless 4.6.0.66

MIT

packaging 21.3

BSD-2-Clause or Apache-2.0

pandas 1.5.2

BSD-3-Clause

parso 0.8.3

MIT

partd 1.3.0

BSD

pathspec 0.10.2

MPL 2.0

pep517 0.13.0

UNKNOWN

pickleshare 0.7.5

MIT

Pillow 9.3.0

HPND

Pint 0.20.1

BSD

pip 23.1.2

MIT

platformdirs 2.5.4

UNKNOWN

pluggy 1.0.0

MIT

polars 0.15.1

MIT

pre-commit 3.0.3

MIT

prettytable 3.8.0

BSD (3 clause)

prompt-toolkit 3.0.33

UNKNOWN

psutil 5.9.4

BSD-3-Clause

psygnal 0.6.1

BSD 3-Clause License

pure-eval 0.2.2

MIT

pyarrow 10.0.1

Apache License, Version 2.0

pycodestyle 2.10.0

Expat license

pydantic 1.10.2

MIT

pydata-sphinx-theme 0.12.0

BSD 3-Clause License

pyflakes 3.0.1

MIT

Pygments 2.13.0

BSD License

PyOpenGL 3.1.6

BSD

pyparsing 3.0.9

UNKNOWN

PyQt5 5.15.7

GPL v3

PyQt5-Qt5 5.15.2

LGPL v3

PyQt5-sip 12.11.0

SIP

pyrsistent 0.19.2

MIT

pytest 7.2.0

MIT

pytest-cov 4.0.0

MIT

python-dateutil 2.8.2

Dual License

python-dotenv 0.21.0

BSD-3-Clause

python-slugify 7.0.0

MIT

pytomlpp 1.0.11

UNKNOWN

pytz 2022.6

MIT

PyWavelets 1.4.1

MIT

pywin32 304

PSF

PyYAML 6.0

MIT

pyzmq 24.0.1

LGPL+BSD

qtconsole 5.4.0

BSD

QtPy 2.3.0

MIT

requests 2.28.1

Apache 2.0

rich 12.6.0

MIT

scikit-image 0.19.3

Modified BSD

scikit-learn 1.1.3

new BSD

scipy 1.9.3

UNKNOWN

seaborn 0.12.2

UNKNOWN

setuptools 67.7.2

UNKNOWN

setuptools-scm 7.0.5

MIT

six 1.16.0

MIT

snowballstemmer 2.2.0

BSD-3-Clause

soupsieve 2.3.2.post1

UNKNOWN

Sphinx 4.5.0

BSD

sphinx-autoapi 2.0.0

MIT

sphinxcontrib-applehelp 1.0.2

BSD

sphinxcontrib-devhelp 1.0.2

BSD

sphinxcontrib-htmlhelp 2.0.0

BSD

sphinxcontrib-jsmath 1.0.1

BSD

sphinxcontrib-qthelp 1.0.3

BSD

sphinxcontrib-serializinghtml 1.1.5

BSD

stack-data 0.6.2

MIT

superqt 0.4.1

BSD 3-Clause License

text-unidecode 1.3

Artistic License

threadpoolctl 3.1.0

BSD-3-Clause

tifffile 2022.10.10

BSD

tomli 2.0.1

UNKNOWN

toolz 0.12.0

BSD

torch 1.13.0+cu117

BSD-3

torchsummary 1.5.1

UNKNOWN

torchvision 0.14.0+cu117

BSD

tornado 6.2

http://www.apache.org/licenses/LICENSE-2.0

tqdm 4.64.1

MPLv2.0, MIT Licences

traitlets 5.6.0

UNKNOWN

typer 0.7.0

UNKNOWN

typing-extensions 4.4.0

UNKNOWN

Unidecode 1.3.6

GPL

urllib3 1.26.13

MIT

virtualenv 20.17.1

MIT

vispy 0.11.0

(new) BSD

wcwidth 0.2.5

MIT

wheel 0.40.0

MIT License

wrapt 1.14.1

BSD

zipp 3.15.0

UNKNOWN

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

locpix-0.1.0.3.tar.gz (2.6 MB view hashes)

Uploaded Source

Built Distribution

locpix-0.1.0.3-py3-none-any.whl (77.4 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page