Metadata-Version: 2.1
Name: pciSeq
Version: 0.0.57.dev0
Summary: Probabilistic cell typing for spatial transcriptomics
Home-page: https://github.com/acycliq/pciSeq
Author: Dimitris Nicoloutsopoulos
Author-email: dimitris.nicoloutsopoulos@gmail.com
License: BSD
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: BSD License
Classifier: Operating System :: OS Independent
Description-Content-Type: text/markdown
License-File: LICENCE
Requires-Dist: numpy-groupies
Requires-Dist: pandas
Requires-Dist: scipy
Requires-Dist: streamlit
Requires-Dist: altair
Requires-Dist: scikit-image
Requires-Dist: scikit-learn
Requires-Dist: tqdm
Requires-Dist: flask
Requires-Dist: numexpr
Requires-Dist: diplib
Requires-Dist: pyvips
Requires-Dist: natsort
Requires-Dist: redis
Provides-Extra: interactive
Requires-Dist: matplotlib (>=2.2.0) ; extra == 'interactive'
Requires-Dist: jupyter ; extra == 'interactive'

# pciSeq: Probabilistic Cell typing by In situ Sequencing
A Python package that implements the cell calling algorithm as described in [Qian, X., et al. Nature Methods (2020)](https://www.nature.com/articles/s41592-019-0631-4)
<p align="center">
    <img src="https://github.com/acycliq/pciSeq/blob/master/assets/screencast_resized.gif?raw=true" alt="screenshot"/>
</p>

## Installation
```
python -m pip install pciSeq
```
Requirement: Python >= 3.8

If you want to work with the source code you can download the repo and then replicate the python environment by
```
conda env create -n pciSeq -f /path/to/environment.yml
```

That will create a conda environment with the name `pciSeq` containing all the necessary packages to run the algorithm. To activate it run 
```
conda activate pciSeq
```
or, if you open the project in your IDE, then in your project settings, switch your interpreter to the interpreter of the `pciSeq` env. 
## Usage
You need to create two `pandas dataframes` for the spots and the single cell data and a `coo_matrix` for the label image (which in 
most cases will be the output of some image segmentation application). Then you pass them into the `pciSeq.fit()` method as follows: 
```
import pciSeq

res = pciSeq.fit(spots=spots_df, coo=label_image, scRNAseq=scRNA_df)
```
See the demo below for a more detailed explanation about the arguments of  `pciSeq.fit()` and its return values.

There is also a fourth argument (optional) to override the default hyperparameter values which are initialised 
by the [config.py](https://github.com/acycliq/pciSeq/blob/master/pciSeq/config.py) module. To pass user-defined hyperparameter values, create a `dictionary` with `keys` the
hyperparameter names and `values` their new values. For example, to exclude all Npy and Vip spots you can do:

```
import pciSeq

opts = { 'exclude_genes': ['Npy', 'Vip'] }
res = pciSeq.fit(spots=spots_df, coo=label_image, scRNAseq=scRNA_df, opts=opts)
```

## Demo
You can run a pciSeq demo in google colab: [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/acycliq/pciSeq/blob/master/notebooks/1_pciSeq.ipynb)

## Viewer
An interactive viewer to explore the data runs on this [url](https://acycliq.github.io/visage/). Instructions about 
building this viewer with your own data are [here](https://github.com/acycliq/visage). \
If you have `v 0.0.49` or greater you can also launch the viewer automatically by 
setting `opts = {'launch_viewer': True}` and passing it to `pciSeq.fit()`, see [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/acycliq/pciSeq/blob/master/notebooks/2_viewer.ipynb)

## Diagnostics
Diagnostics will help you understand whether pciSeq has been misconfigured, the algorithm has taken the 
wrong path and will produce meaningless results when it finishes. You will need however to install redis (or Memurai if you are using Windows).

For linux do:
`sudo apt-get install redis-server redis-tools`
and then start the service:
`sudo service redis-server start`

You can get the free version of memurai from [here](https://www.memurai.com/get-memurai). Once installed, the service should start automatically but you can manually do that by:
`memurai.exe â€“service-start`

If redis (or memurai) is missing from your system, the call to launch the diagnostics dashboard will be 
ignored. If you are interested in this feature you may find this notebook [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/acycliq/pciSeq/blob/master/notebooks/4_diagnostics.ipynb)
 useful

## Change Log
### [0.0.56] - 2023-07-03
 - Diagnostics dashboard

 - Baselayers on the viewer. You can have multiple background images and switch between them.

### [0.0.50] - 2023-05-27
 - Single cell data are optional, more info can be found here [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/acycliq/pciSeq/blob/master/notebooks/3_pciSeq_without_singleCell_data.ipynb)

 - `pciSeq.fit()` takes keyword arguments


## References 
Qian, X., et al. (2020). Probabilistic cell typing enables fine mapping of closely related cell types in situ. Nat
Methods 17, 101 - 106.


