Mowgli / Git / Diff of /README.md

Models:

AlyssaS/

Mowgli

Downloads: 1

Diff of /README.md [061d85] .. [28d9d2]

Switch to unified view


# Mowgli: Multi Omics Wasserstein inteGrative anaLysIs
[![Tests](https://github.com/gjhuizing/Mowgli/actions/workflows/main.yml/badge.svg)](https://github.com/gjhuizing/Mowgli/actions/workflows/main.yml)
[![codecov](https://codecov.io/gh/cantinilab/Mowgli/branch/main/graph/badge.svg?token=UBUJF7098Q)](https://codecov.io/gh/cantinilab/Mowgli)
[![Documentation Status](https://readthedocs.org/projects/mowgli/badge/?version=latest)](https://mowgli.readthedocs.io/en/latest/?badge=latest)
[![PyPI version](https://img.shields.io/pypi/v/mowgli?color=blue)](https://img.shields.io/pypi/v/mowgli?color=blue)
[![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
[![DOI](https://zenodo.org/badge/391909874.svg)](https://zenodo.org/badge/latestdoi/391909874)

Mowgli is a novel method for the integration of paired multi-omics data with any type and number of omics, combining integrative Nonnegative Matrix Factorization and Optimal Transport. [Read the paper!](https://www.nature.com/articles/s41467-023-43019-2)



## Install the package

Mowgli is implemented as a Python package seamlessly integrated within the scverse ecosystem, in particular Muon and Scanpy.

### via PyPI (recommended)

On all operating systems, the easiest way to install Mowgli is via PyPI. Installation should typically take a minute and is continuously tested with Python 3.10 on an Ubuntu virtual machine.

```bash
pip install mowgli
```

### via GitHub (development version)

```bash
git clone git@github.com:cantinilab/Mowgli.git
pip install ./Mowgli/
```

### Test your installation (optional)

```bash
pytest .
```

## Getting started

Mowgli takes as an input a Muon object and populates its `obsm` and `uns` fields with the embeddings and dictionaries, respectively. Visit [mowgli.rtfd.io](https://mowgli.rtfd.io/) for more documentation and tutorials.

You may download a preprocessed 10X Multiome demo dataset [here](https://figshare.com/s/4c8e72cbb188d8e1cce8).

A GPU is not required for small datasets, but is strongly recommended above 1,000 cells. On CPU, the [cell lines demo](https://mowgli.readthedocs.io/en/latest/vignettes/Liu%20cell%20lines.html) (206 cells) should run in under 5 minutes and the [PBMC demo](https://mowgli.readthedocs.io/en/latest/vignettes/PBMC.html) (500 cells) should run in under 10 minutes (tested on a Ubuntu 20.04 machine with an 11th gen i7 processor).

```python
import mowgli
import mudata as md
import scanpy as sc

# Load data into a Muon object.
mdata = md.read_h5mu("my_data.h5mu")

# Initialize and train the model.
model = mowgli.models.MowgliModel(latent_dim=15)
model.train(mdata)

# Visualize the embedding with UMAP.
sc.pp.neighbors(mdata, use_rep="W_OT")
sc.tl.umap(mdata)
sc.pl.umap(mdata)
```

## Publication

```bibtex
@article{huizing2023paired,
  title={Paired single-cell multi-omics data integration with Mowgli},
  author={Huizing, Geert-Jan and Deutschmann, Ina Maria and Peyr{\'e}, Gabriel and Cantini, Laura},
  journal={Nature Communications},
  volume={14},
  number={1},
  pages={7711},
  year={2023},
  publisher={Nature Publishing Group UK London}
}
```

If you're looking for the repository with code to reproduce the experiments in our preprint, [here is is!](https://github.com/cantinilab/mowgli_reproducibility)

	a/README.md		b/README.md
1	# Mowgli: Multi Omics Wasserstein inteGrative anaLysIs	1	# Mowgli: Multi Omics Wasserstein inteGrative anaLysIs
2	[![Tests](https://github.com/gjhuizing/Mowgli/actions/workflows/main.yml/badge.svg)](https://github.com/gjhuizing/Mowgli/actions/workflows/main.yml)	2	[![Tests](https://github.com/gjhuizing/Mowgli/actions/workflows/main.yml/badge.svg)](https://github.com/gjhuizing/Mowgli/actions/workflows/main.yml)
3	[![codecov](https://codecov.io/gh/cantinilab/Mowgli/branch/main/graph/badge.svg?token=UBUJF7098Q)](https://codecov.io/gh/cantinilab/Mowgli)	3	[![codecov](https://codecov.io/gh/cantinilab/Mowgli/branch/main/graph/badge.svg?token=UBUJF7098Q)](https://codecov.io/gh/cantinilab/Mowgli)
4	[![Documentation Status](https://readthedocs.org/projects/mowgli/badge/?version=latest)](https://mowgli.readthedocs.io/en/latest/?badge=latest)	4	[![Documentation Status](https://readthedocs.org/projects/mowgli/badge/?version=latest)](https://mowgli.readthedocs.io/en/latest/?badge=latest)
5	[![PyPI version](https://img.shields.io/pypi/v/mowgli?color=blue)](https://img.shields.io/pypi/v/mowgli?color=blue)	5	[![PyPI version](https://img.shields.io/pypi/v/mowgli?color=blue)](https://img.shields.io/pypi/v/mowgli?color=blue)
6	[![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)	6	[![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
7	[![DOI](https://zenodo.org/badge/391909874.svg)](https://zenodo.org/badge/latestdoi/391909874)	7	[![DOI](https://zenodo.org/badge/391909874.svg)](https://zenodo.org/badge/latestdoi/391909874)
8		8
9	Mowgli is a novel method for the integration of paired multi-omics data with any type and number of omics, combining integrative Nonnegative Matrix Factorization and Optimal Transport. [Read the paper!](https://www.nature.com/articles/s41467-023-43019-2)	9	Mowgli is a novel method for the integration of paired multi-omics data with any type and number of omics, combining integrative Nonnegative Matrix Factorization and Optimal Transport. [Read the paper!](https://www.nature.com/articles/s41467-023-43019-2)
10		10
11	![figure](figure.png)	11
12
13	## Install the package	12	## Install the package
14		13
15	Mowgli is implemented as a Python package seamlessly integrated within the scverse ecosystem, in particular Muon and Scanpy.	14	Mowgli is implemented as a Python package seamlessly integrated within the scverse ecosystem, in particular Muon and Scanpy.
16		15
17	### via PyPI (recommended)	16	### via PyPI (recommended)
18		17
19	On all operating systems, the easiest way to install Mowgli is via PyPI. Installation should typically take a minute and is continuously tested with Python 3.10 on an Ubuntu virtual machine.	18	On all operating systems, the easiest way to install Mowgli is via PyPI. Installation should typically take a minute and is continuously tested with Python 3.10 on an Ubuntu virtual machine.
20		19
21	```bash	20	```bash
22	pip install mowgli	21	pip install mowgli
23	```	22	```
24		23
25	### via GitHub (development version)	24	### via GitHub (development version)
26		25
27	```bash	26	```bash
28	git clone git@github.com:cantinilab/Mowgli.git	27	git clone git@github.com:cantinilab/Mowgli.git
29	pip install ./Mowgli/	28	pip install ./Mowgli/
30	```	29	```
31		30
32	### Test your installation (optional)	31	### Test your installation (optional)
33		32
34	```bash	33	```bash
35	pytest .	34	pytest .
36	```	35	```
37		36
38	## Getting started	37	## Getting started
39		38
40	Mowgli takes as an input a Muon object and populates its `obsm` and `uns` fields with the embeddings and dictionaries, respectively. Visit [mowgli.rtfd.io](https://mowgli.rtfd.io/) for more documentation and tutorials.	39	Mowgli takes as an input a Muon object and populates its `obsm` and `uns` fields with the embeddings and dictionaries, respectively. Visit [mowgli.rtfd.io](https://mowgli.rtfd.io/) for more documentation and tutorials.
41		40
42	You may download a preprocessed 10X Multiome demo dataset [here](https://figshare.com/s/4c8e72cbb188d8e1cce8).	41	You may download a preprocessed 10X Multiome demo dataset [here](https://figshare.com/s/4c8e72cbb188d8e1cce8).
43		42
44	A GPU is not required for small datasets, but is strongly recommended above 1,000 cells. On CPU, the [cell lines demo](https://mowgli.readthedocs.io/en/latest/vignettes/Liu%20cell%20lines.html) (206 cells) should run in under 5 minutes and the [PBMC demo](https://mowgli.readthedocs.io/en/latest/vignettes/PBMC.html) (500 cells) should run in under 10 minutes (tested on a Ubuntu 20.04 machine with an 11th gen i7 processor).	43	A GPU is not required for small datasets, but is strongly recommended above 1,000 cells. On CPU, the [cell lines demo](https://mowgli.readthedocs.io/en/latest/vignettes/Liu%20cell%20lines.html) (206 cells) should run in under 5 minutes and the [PBMC demo](https://mowgli.readthedocs.io/en/latest/vignettes/PBMC.html) (500 cells) should run in under 10 minutes (tested on a Ubuntu 20.04 machine with an 11th gen i7 processor).
45		44
46	```python	45	```python
47	import mowgli	46	import mowgli
48	import mudata as md	47	import mudata as md
49	import scanpy as sc	48	import scanpy as sc
50		49
51	# Load data into a Muon object.	50	# Load data into a Muon object.
52	mdata = md.read_h5mu("my_data.h5mu")	51	mdata = md.read_h5mu("my_data.h5mu")
53		52
54	# Initialize and train the model.	53	# Initialize and train the model.
55	model = mowgli.models.MowgliModel(latent_dim=15)	54	model = mowgli.models.MowgliModel(latent_dim=15)
56	model.train(mdata)	55	model.train(mdata)
57		56
58	# Visualize the embedding with UMAP.	57	# Visualize the embedding with UMAP.
59	sc.pp.neighbors(mdata, use_rep="W_OT")	58	sc.pp.neighbors(mdata, use_rep="W_OT")
60	sc.tl.umap(mdata)	59	sc.tl.umap(mdata)
61	sc.pl.umap(mdata)	60	sc.pl.umap(mdata)
62	```	61	```
63		62
64	## Publication	63	## Publication
65		64
66	```bibtex	65	```bibtex
67	@article{huizing2023paired,	66	@article{huizing2023paired,
68	title={Paired single-cell multi-omics data integration with Mowgli},	67	title={Paired single-cell multi-omics data integration with Mowgli},
69	author={Huizing, Geert-Jan and Deutschmann, Ina Maria and Peyr{\'e}, Gabriel and Cantini, Laura},	68	author={Huizing, Geert-Jan and Deutschmann, Ina Maria and Peyr{\'e}, Gabriel and Cantini, Laura},
70	journal={Nature Communications},	69	journal={Nature Communications},
71	volume={14},	70	volume={14},
72	number={1},	71	number={1},
73	pages={7711},	72	pages={7711},
74	year={2023},	73	year={2023},
75	publisher={Nature Publishing Group UK London}	74	publisher={Nature Publishing Group UK London}
76	}	75	}
77	```	76	```
78		77
79	If you're looking for the repository with code to reproduce the experiments in our preprint, [here is is!](https://github.com/cantinilab/mowgli_reproducibility)	78	If you're looking for the repository with code to reproduce the experiments in our preprint, [here is is!](https://github.com/cantinilab/mowgli_reproducibility)