Analyse coexpression with MCODE

Yasir Demirtaş, Kozo Nishida, Barry Demchak

py4cytoscape 0.0.10

Prerequisites

In addition to this package (py4cytoscape), you will need:

• Cytoscape 3.8 or greater, which can be downloaded from https://cytoscape.org/download.html. Simply follow the installation instructions on screen.

• Complete installation wizard

• Launch Cytoscape

• If your Cytoscape is 3.8.2 or earlier, install FileTransfer App (Follow here to do it.)

NOTE: To run this notebook, you must manually start Cytoscape first – don’t proceed until you have started Cytoscape.

Setup required only in a remote notebook environment

If you’re using a remote Jupyter Notebook environment such as Google Colab, run the cell below. (If you’re running a local Jupyter Notebook server on the desktop machine same with Cytoscape, you don’t need to do that.)

(Note that to use the current py4cytoscape release (instead of v0.0.10), remove the _PY4CYTOSCAPE= line in the snippet below.)

_PY4CYTOSCAPE = 'git+https://github.com/cytoscape/py4cytoscape@0.0.10'
import requests
exec(requests.get("https://raw.githubusercontent.com/cytoscape/jupyter-bridge/master/client/p4c_init.py").text)
IPython.display.Javascript(_PY4CYTOSCAPE_BROWSER_CLIENT_JS) # Start browser client

!pip install concurrent-log-handler
!pip install openpyxl

Requirement already satisfied: concurrent-log-handler in c:\users\hoge\miniforge3\lib\site-packages (0.9.19)
Requirement already satisfied: portalocker>=1.4.0 in c:\users\hoge\miniforge3\lib\site-packages (from concurrent-log-handler) (1.7.0)
Requirement already satisfied: pywin32!=226 in c:\users\hoge\miniforge3\lib\site-packages (from portalocker>=1.4.0->concurrent-log-handler) (301)

import py4cytoscape as p4c
import pandas as pd
import os
import uuid
from logging import getLogger, INFO
from concurrent_log_handler import ConcurrentRotatingFileHandler
from IPython.display import Image
from time import sleep
%config InlineBackend.figure_format = 'svg'

p4c.install_app('MCODE')

{}

{}

Clear all networks at the begining

p4c.session.close_session(False)

{}

Import Co-Exp Data and Rename columns

results = pd.read_excel('https://github.com/cytoscape/py4cytoscape/raw/0.0.10/doc/tutorials/input_data/co_exp_result.xlsx')

results

	gene1	gene2	score
0	ACSM3	COL6A2	-0.518721
1	ACSM3	DPYSL3	-0.504664
2	ACSM3	THY1	-0.494217
3	ADAM28	AL928768.3 /// Ig alpha 1-[alpha]2m /// IGH	0.553792
4	ADH1B	BUB1B	-0.565257
...	...	...	...
1117	THBS2	THY1	0.691359
1118	THBS2	VCAN	0.630551
1119	THBS4	THY1	0.530471
1120	THY1	VCAN	0.510101
1121	TOP2A	UBE2C	0.656060

1122 rows × 3 columns

results = results.rename(columns={'gene1': 'source', 'gene2': 'target'})

results

	source	target	score
0	ACSM3	COL6A2	-0.518721
1	ACSM3	DPYSL3	-0.504664
2	ACSM3	THY1	-0.494217
3	ADAM28	AL928768.3 /// Ig alpha 1-[alpha]2m /// IGH	0.553792
4	ADH1B	BUB1B	-0.565257
...	...	...	...
1117	THBS2	THY1	0.691359
1118	THBS2	VCAN	0.630551
1119	THBS4	THY1	0.530471
1120	THY1	VCAN	0.510101
1121	TOP2A	UBE2C	0.656060

1122 rows × 3 columns

Create network with co-exp data and give random network name

network_name = uuid.uuid4().hex
network_suid = p4c.networks.create_network_from_data_frames(
    edges=results,
    title= network_name,
    collection= network_name + '_collection'
)

Applying default style...
Applying preferred layout

p4c.notebook_show_image()

Run analyze command so your network will be one way directed

analyze = p4c.commands.commands_post('analyzer/analyze')

analyze

{'networkTitle': 'a4e2797fa98546b6b878e4f5099e1fdc (undirected)',
 'nodeCount': '197',
 'edgeCount': '1122',
 'avNeighbors': '11.86096256684492',
 'diameter': '8',
 'radius': '5',
 'avSpl': '3.271634753608188',
 'cc': '0.5249889291420875',
 'density': '0.06376861595077914',
 'heterogeneity': '0.8353654292197396',
 'centralization': '0.16378959604766055',
 'ncc': '4',
 'time': '0.082'}

Run mcode command with options (you can change network settings)

mcode = p4c.commands.commands_post('mcode cluster degreeCutoff=2 fluff=false fluffNodeDensityCutoff=0.1 haircut=true includeLoops=false kCore=2 maxDepthFromStart=100 network=current nodeScoreCutoff=0.2 scope=NETWORK')
sleep(5)

mcode

{'id': 1,
 'parameters': {'scope': 'NETWORK',
  'includeLoops': False,
  'degreeCutoff': 2,
  'kCore': 2,
  'maxDepthFromStart': 100,
  'nodeScoreCutoff': 0.2,
  'haircut': True,
  'fluff': False,
  'fluffNodeDensityCutoff': 0.1,
  'selectedNodes': []},
 'clusters': [{'rank': 0,
   'name': 'Cluster 0',
   'score': 17.263157894736842,
   'seedNode': 138,
   'nodes': [64,
    96,
    98,
    99,
    100,
    196,
    102,
    103,
    105,
    106,
    138,
    81,
    83,
    179,
    116,
    86,
    89,
    90,
    93,
    95]},
  {'rank': 0,
   'name': 'Cluster 0',
   'score': 9.916666666666666,
   'seedNode': 143,
   'nodes': [68,
    69,
    70,
    74,
    79,
    143,
    80,
    82,
    84,
    87,
    88,
    91,
    92,
    94,
    104,
    108,
    172,
    173,
    174,
    176,
    177,
    114,
    178,
    62,
    126]},
  {'rank': 0,
   'name': 'Cluster 0',
   'score': 8.842105263157894,
   'seedNode': 180,
   'nodes': [161,
    162,
    230,
    73,
    144,
    145,
    146,
    147,
    148,
    180,
    149,
    150,
    151,
    153,
    217,
    155,
    156,
    157,
    158,
    159]},
  {'rank': 0,
   'name': 'Cluster 0',
   'score': 5.666666666666667,
   'seedNode': 227,
   'nodes': [227, 197, 134, 198, 199, 136, 200, 137, 201, 233, 202, 75, 152]},
  {'rank': 0,
   'name': 'Cluster 0',
   'score': 5.666666666666667,
   'seedNode': 123,
   'nodes': [112, 118, 121, 123, 124, 125, 110]},
  {'rank': 0,
   'name': 'Cluster 0',
   'score': 5.0,
   'seedNode': 242,
   'nodes': [187, 133, 213, 214, 215]},
  {'rank': 0,
   'name': 'Cluster 0',
   'score': 5.0,
   'seedNode': 142,
   'nodes': [66, 139, 140, 141, 142]},
  {'rank': 0,
   'name': 'Cluster 0',
   'score': 4.0,
   'seedNode': 115,
   'nodes': [97, 115, 101, 117, 182, 171, 63]},
  {'rank': 0,
   'name': 'Cluster 0',
   'score': 4.0,
   'seedNode': 163,
   'nodes': [163, 164, 165, 166]},
  {'rank': 0,
   'name': 'Cluster 0',
   'score': 3.0,
   'seedNode': 252,
   'nodes': [252, 249, 190]},
  {'rank': 0,
   'name': 'Cluster 0',
   'score': 3.0,
   'seedNode': 206,
   'nodes': [208, 209, 206]}],
 'stale': False,
 'lock': {}}

Mcode created clusters, we need big clusters which is node count bigger than 20 (you can change threshold)

big_clusters = [i for i in range(len(mcode['clusters'])) if len(mcode['clusters'][i]['nodes']) >= 20]

big_clusters

[0, 1, 2]

Create empty list and append cluster tables to in it

Create cluster images as a png and network file as a .cyjs so you are able to use in cytoscape.js

p4c.get_table_column_names()

['SUID',
 'shared name',
 'name',
 'selected',
 'id',
 'AverageShortestPathLength',
 'ClusteringCoefficient',
 'ClosenessCentrality',
 'IsSingleNode',
 'PartnerOfMultiEdgedNodePairs',
 'SelfLoops',
 'Eccentricity',
 'Stress',
 'Degree',
 'BetweennessCentrality',
 'NeighborhoodConnectivity',
 'NumberOfDirectedEdges',
 'NumberOfUndirectedEdges',
 'Radiality',
 'TopologicalCoefficient',
 'MCODE::Score (1)',
 'MCODE::Node Status (1)',
 'MCODE::Clusters (1)']

clusters = []
img_names = []
for i in range(len(big_clusters)):
    view_id = p4c.commands.commands_post('mcode view id=1 rank=' + str(i+1) )
    clusters.append(p4c.tables.get_table_columns() \
                  .drop(columns=['id','name','selected','MCODE::Clusters (1)','SUID']) \
                  .sort_values(by=['MCODE::Score (1)'], ascending=False) \
                 .reset_index(drop=True) )

    p4c.network_views.get_network_view_suid()
    img_name = uuid.uuid4().hex
    p4c.export_image(str(img_name))
    p4c.export_network(img_name,'cyjs')
    img_names.append(img_name+'.png')

img_names

['a2683fbd40324fcdb6c1fbab9b931a06.png',
 '59e2d4909be74f938fb045c1bec2b6c7.png',
 'a10d252d6cd346f5a948d526a29cf1bf.png']

p4c.notebook_show_image(img_names[0])

p4c.notebook_show_image(img_names[1])

p4c.notebook_show_image(img_names[2])