FlowSOM

This repository contains a Python implementation of FlowSOM algorithm for clustering and visualizing a mass cytometry data set.

For more details about the algorithm, please check (En|中)

Installation

Just use pip

pip install FlowSom

Or download this repository to a directory of your choice and then run:

pip install -r requirements.txt

How to use it

<dl> <dt>Read Files</dt> <dd>In order to use FlowSOM you need your data saved as a .csv file or a .fcs file.</dd> </dl>

file = r'flowmetry.fcs'

<dl> <dd>Or</dd> </dl>

file = 'flowmetry.csv'

<dl> <dt>Import Package</dt> <dd>Then you need to import the package.</dd> <dd>If you install the package via pip, then you should run</dd> </dl>

from flowsom import flowsom

<dl> <dd>If you download the repository, you should run</dd> </dl>

from flowsom import *

<dl> <dt>Play Around</dt> <dd>Then you can run FlowSOM just as follows:</dd> </dl>

fsom = flowsom(file) # read the data
fsom.som_mapping(50, 50, 31, sigma=2.5, 
                 learning_rate=0.1, batch_size=100)  # trains SOM with 100 iterations
fsom.meta_clustering(AgglomerativeClustering, min_n=40, 
                     max_n=45, 
                     iter_n=3) # train the meta clustering for cluster in range(40,45)       

Use the trained output

After the training, you will be able to:

• Get the weights of SOM with method fsom.map_som
• Get the best number of clustering with method fsom.bestk
• Get the prediction dataframe with method fsom.df and fsom.tf_df
• Visualize the final clustering outcome with methodfsom.vis

Examples

The demo code could be found here.

The distance map of SOM trained from a sample flow cytometry data:

<img src="https://github.com/Hatchin/FlowSOM/blob/master/img/som.png" alt="Flow example">

The visualization example after meta-clustering using Minimal Spanning Tree (MST): <img src="https://github.com/Hatchin/FlowSOM/blob/master/img/mst.png" alt="MST example">

FlowSOM Algorithm

FlowSOM analyzes flow or mass cytometry data using a self-Organizing Map (SOM). Using a two-level clustering and star charts, FlowSOM helps to obtain a clear overview of how all markers are behaving on all cells, and to detect subsets that might be missed otherwise.

The algorithm consists of four steps:

• reading the data
• building a Self-Organizing Map
• building a minimal spanning tree
• computing a meta-clustering

Self-Organizing Map

SOM is a type of unsupervised Artificial Neural Network able to convert complex, nonlinear statistical relationships between high-dimensional data items into simple geometric relationships on a low-dimensional display. Introduction

Minimum Spanning Tree

A minimum spanning tree (MST) or minimum weight spanning tree is a subset of the edges of a connected, edge-weighted undirected graph that connects all the vertices together, without any cycles and with the minimum possible total edge weight.

Meta-clustering

The meta-clustering technique conducted on the SOM is hierarchical consensus meta-clustering, which clusters the weights of trained SOM into different groups.

Acknowledge

FlowSOM is built based on FlowCytometryTools, MiniSom and Consensus Clustering.

Update pypi: source