1openTSNE
2========
3
4|Build Status| |ReadTheDocs Badge| |License Badge|
5
6openTSNE is a modular Python implementation of t-Distributed Stochasitc Neighbor Embedding (t-SNE) [1]_, a popular dimensionality-reduction algorithm for visualizing high-dimensional data sets. openTSNE incorporates the latest improvements to the t-SNE algorithm, including the ability to add new data points to existing embeddings [2]_, massive speed improvements [3]_ [4]_, enabling t-SNE to scale to millions of data points and various tricks to improve global alignment of the resulting visualizations [5]_.
7
8.. figure:: docs/source/images/macosko_2015.png
9 :alt: Macosko 2015 mouse retina t-SNE embedding
10 :align: center
11
12 A visualization of 44,808 single cell transcriptomes obtained from the mouse retina [6]_ embedded using the multiscale kernel trick to better preserve the global aligment of the clusters.
13
14- `Documentation <http://opentsne.readthedocs.io>`__
15- `User Guide and Tutorial <https://opentsne.readthedocs.io/en/latest/tsne_algorithm.html>`__
16- Examples: `basic <https://opentsne.readthedocs.io/en/latest/examples/01_simple_usage/01_simple_usage.html>`__, `advanced <https://opentsne.readthedocs.io/en/latest/examples/02_advanced_usage/02_advanced_usage.html>`__, `preserving global alignment <https://opentsne.readthedocs.io/en/latest/examples/03_preserving_global_structure/03_preserving_global_structure.html>`__, `embedding large data sets <https://opentsne.readthedocs.io/en/latest/examples/04_large_data_sets/04_large_data_sets.html>`__
17- `Speed benchmarks <https://opentsne.readthedocs.io/en/latest/benchmarks.html>`__
18
19Installation
20------------
21
22openTSNE requires Python 3.7 or higher in order to run.
23
24Conda
25~~~~~
26
27openTSNE can be easily installed from ``conda-forge`` with
28
29::
30
31 conda install --channel conda-forge opentsne
32
33`Conda package <https://anaconda.org/conda-forge/opentsne>`__
34
35PyPi
36~~~~
37
38openTSNE is also available through ``pip`` and can be installed with
39
40::
41
42 pip install opentsne
43
44`PyPi package <https://pypi.org/project/openTSNE>`__
45
46Installing from source
47~~~~~~~~~~~~~~~~~~~~~~
48
49If you wish to install openTSNE from source, please run
50
51::
52
53 python setup.py install
54
55
56in the root directory to install the appropriate dependencies and compile the necessary binary files.
57
58Please note that openTSNE requires a C/C++ compiler to be available on the system. Additionally,
59``numpy`` must be pre-installed in the active environment.
60
61In order for openTSNE to utilize multiple threads, the C/C++ compiler
62must support ``OpenMP``. In practice, almost all compilers
63implement this with the exception of older version of ``clang`` on OSX
64systems.
65
66To squeeze the most out of openTSNE, you may also consider installing
67FFTW3 prior to installation. FFTW3 implements the Fast Fourier
68Transform, which is heavily used in openTSNE. If FFTW3 is not available,
69openTSNE will use numpy’s implementation of the FFT, which is slightly
70slower than FFTW. The difference is only noticeable with large data sets
71containing millions of data points.
72
73A hello world example
74---------------------
75
76Getting started with openTSNE is very simple. First, we'll load up some data using scikit-learn
77
78.. code:: python
79
80 from sklearn import datasets
81
82 iris = datasets.load_iris()
83 x, y = iris["data"], iris["target"]
84
85then, we'll import and run
86
87.. code:: python
88
89 from openTSNE import TSNE
90
91 embedding = TSNE().fit(x)
92
93Citation
94--------
95
96If you make use of openTSNE for your work we would appreciate it if you would cite the paper
97
98.. code::
99
100 @article {Poli{\v c}ar731877,
101 author = {Poli{\v c}ar, Pavlin G. and Stra{\v z}ar, Martin and Zupan, Bla{\v z}},
102 title = {openTSNE: a modular Python library for t-SNE dimensionality reduction and embedding},
103 year = {2019},
104 doi = {10.1101/731877},
105 publisher = {Cold Spring Harbor Laboratory},
106 URL = {https://www.biorxiv.org/content/early/2019/08/13/731877},
107 eprint = {https://www.biorxiv.org/content/early/2019/08/13/731877.full.pdf},
108 journal = {bioRxiv}
109 }
110
111openTSNE implements two efficient algorithms for t-SNE. Please consider citing the original authors of the algorithm that you use. If you use FIt-SNE (default), then the citation is [4]_ below, but if you use Barnes-Hut the citation is [3]_.
112
113
114References
115----------
116
117.. [1] Van Der Maaten, Laurens, and Hinton, Geoffrey. `“Visualizing data using
118 t-SNE.” <http://www.jmlr.org/papers/volume9/vandermaaten08a/vandermaaten08a.pdf>`__
119 Journal of Machine Learning Research 9.Nov (2008): 2579-2605.
120.. [2] Poličar, Pavlin G., Martin Stražar, and Blaž Zupan. `“Embedding to Reference t-SNE Space Addresses Batch Effects in Single-Cell Classification.” <https://link.springer.com/article/10.1007/s10994-021-06043-1>`__ Machine Learning (2021): 1-20.
121.. [3] Van Der Maaten, Laurens. `“Accelerating t-SNE using tree-based algorithms.”
122 <http://www.jmlr.org/papers/volume15/vandermaaten14a/vandermaaten14a.pdf>`__
123 Journal of Machine Learning Research 15.1 (2014): 3221-3245.
124.. [4] Linderman, George C., et al. `"Fast interpolation-based t-SNE for improved
125 visualization of single-cell RNA-seq data." <https://www.nature.com/articles/s41592-018-0308-4>`__ Nature Methods 16.3 (2019): 243.
126.. [5] Kobak, Dmitry, and Berens, Philipp. `“The art of using t-SNE for single-cell transcriptomics.” <https://www.nature.com/articles/s41467-019-13056-x>`__
127 Nature Communications 10, 5416 (2019).
128.. [6] Macosko, Evan Z., et al. \ `“Highly parallel genome-wide expression profiling of
129 individual cells using nanoliter droplets.”
130 <https://www.sciencedirect.com/science/article/pii/S0092867415005498>`__
131 Cell 161.5 (2015): 1202-1214.
132
133.. |Build Status| image:: https://dev.azure.com/pavlingp/openTSNE/_apis/build/status/Test?branchName=master
134 :target: https://dev.azure.com/pavlingp/openTSNE/_build/latest?definitionId=1&branchName=master
135.. |ReadTheDocs Badge| image:: https://readthedocs.org/projects/opentsne/badge/?version=latest
136 :target: https://opentsne.readthedocs.io/en/latest/?badge=latest
137 :alt: Documentation Status
138.. |License Badge| image:: https://img.shields.io/badge/License-BSD%203--Clause-blue.svg
139 :target: https://opensource.org/licenses/BSD-3-Clause
140
