introduction to biological network analysis and visualization with cytoscape part1

Introduction to Biological Network Analysis and Visualization with Cytoscape

Keiichiro OnoCytoscape Core Developer TeamUC, San Diego Trey Ideker Lab / National Resource for Network Biology

5/10/2016 The Scripps Research Institute

Lecture 1: Basics

Keiichiro Ono

Cytoscape Core Developer since 2005 @UCSD Trey Ideker Lab

Area of Interest:Biological Data Integration & Visualization

Agenda

• Lecture 1 (Today): Introduction to Biological Network Analysis and Visualization

• What is the benefits of biological network analysis and visualization?

• Introduction to Cytoscape

• Preview of Lecture 2: cyREST

• Lecture 2:Reproducible Analysis & Visualization

• Introduction to Jupyter Notebook

• Create a reproducible network visualization workflows with Python

All documents, data, and code are available here:

https://github.com/idekerlab/tsri-lecture

https://github.com/idekerlab/tsri-lecture

Why Network Analysis?

Networks?

EP300

PPARG

SMARCD3

STMN1

SMARCA4

OPTN

ATP6V1C1

PSMD1

HTT

PRNP

HNRNPUL1

CCDC88A

CLU

HSP90AB1

SMARCD3

MAP4K4

MIF4GDUSP11

MARCH6TUBB

EDF1 CHD8

Protein-Protein Interactions

Human-Curated Pathways

KEGG Pathway (TCA Cycle) visualized by Cytoscape KGMLReader

Interactomes

Human Interactome data from BioGRID visualized with Cytoscape

Social Networks

Network extracted from Panama Papers data set

Node 1 - Edge Type - Node 2

Protein 1 - Y2H - Protein 2

Networks vs Pathways

Networks Pathways

Networks Pathways

Collection of binary interactions Human-curated / detailed

Large Scale Small Scale

Generated from omics-data Constructed from literature

Networks / Pathways = Graphs

Benefits of Network Analysis


- You have list of N genes from your screening

- Now you want to know: - Relationships among those genes

- Functions - etc.

Screening 1

PPARGTCF7L2RETNIRS1HNF1AHNF4AKCNJ11GCKLIPCPTPN1ABCC8ENPP1HNF1B

ENSG00000167780ENSG00000255974

EBI-9992455

ENSG00000241119 ANP32B

EBI-10055098

EBI-10055672

EBI-9871829

NR3C1

EBI-10096648

STK16

SNCG

EBI-10034984

EBI-9973444

RAD50

EBI-9980932

ENSG00000070019

EBI-9871836

ENSG00000105398

EBI-9992700

ENSG00000207778

ENSG00000143819

EBI-9980935

ENSG00000125730

ENSG00000180432

ENSG00000197249

EXT2

EBI-5333164

EBI-10051521

OPTN

EBI-10050213 PPARGC1A

EBI-10039585

MAPK8

HNF4A EDF1

SFPQ

ENSG00000110245

PPARG EP300

SMARCD3

PRNPHNRNPUL1

ATP6V1C1

HTT

EBI-10050241

EBI-10039564

ENSG00000118137

PABPC4

EBI-10050232

EBI-10051518

EBI-9871840

EBI-10096653

ENSG00000095596

MARCH6

EBI-3946155

EBI-5333185

BCAR3IRS1 PIK3CA

LRRK2

Irs1

PIK3R1

PIK3R1

Grb2

Phb

Grin1

H1f0

Rps3

Rpn2

Ipo9

Scml2

Cand1

Eif2ak2

Ipo8

P p fi b p 1

S d p r

Poldip3

Tenm3Rars

Ppp2r1a

Vim

Rfc3

Hsp90ab1

R q c d 1

Rplp0

Hnrnpu

Irs1

Snd1Hspa8

Ung

Tp53bp2

GRB2

TP53BP2

YWHAB

IGF1R

DDR1

SMAD3

EBI-1108795

Ctnnd1

Ran

Ywhag

Rab6b

Ybx1

Epha2

Grin2b

Flot2

Aff3

Ptpn1

PELI1

EIF2AK2

INSR

YWHAE

NCOR1

Pik3r1

N U M B

PRKCQ

TP53BP2

PIK3R3

RAP2A Mink1 CHEBI:39112

Ywhab

CHEBI:39079

CDK5

STMN1

PPARG USP11

CHEBI:45783

CHEBI:18319

HSP90AB1

MAP4K4 MIF4GD

CHEBI:64310

CHEBI:49840

TUBBCLU

CHD8

CCDC88A

SMARCA4

SMARCD3

PSMD1

MAP4K2 MINK1

Map3k1MAP3K1

PKM

MAP3K11

CHEBI:49375

GOLIM4

ARHGAP39

GOSR1

BCOR

PTRF

PCNPKIDINS220

TMEM216

PHB

ABCD3

1C

ATP6V0D1

FLOT2

RNF213

SMARCD2

LTN1

PIAS1

Sumo1

ATP6V0A1WRAP53

EVC2

ACTN1

GALNT2

dhbF

B4E2V5

TCTN3

GHR

EGFRATP6V1B2

CDH2

TMEM17

RMDN3

PHB2 PTPN1 Cdh2

TCTN2

PDGFRB

ATF2

CHEBI:17440

mviM1MET

MVP

MSN

MREG

COL5A1

FLOT1

ASS1

CHEBI:17283

uvrB

FOXM1

t y r B

UBE2I

glnA

Ctnnb1

PSMA3

ENSG00000138795

CTNNB1

RUNX3

PSEN1

JUP

DAXX

TCF7L2

hmwP2

hemL2

vgrG7

f a d B

FBLN1

XRCC6

XRCC5

PARP1

Q99IB8-PRO_0000045599

pagA

Psen1

ENSG00000168646

ENSG00000065361

p p s C

Q99IB8-PRO_0000045596

Q8CLD5

TGIF1

YPO2975

p g i

pyrE

ENSG00000110092

ENSG00000136997


- You can see the relationships among the group of biological entities - Find drag targets - Overrepresented functions AND their

connections

Gene List to Network to Biological Insight

- How? - You need to search, integrate, and visualize

multiple data sources This is what you will learn in this lecture

What is Cytoscape?

An Open Source Platform for Biological Network Data Integration, Analysis and Visualization

Cytoscape

Cytoscape 3.4.0 (Latest Release)

Cytoscape- Open Source (LGPL)

- Free for both commercial and academic use - Developed and maintained by universities,

companies, and research institutions - UC, San Diego - University of Toronto - UC, San Francisco - ISB - And collaborators world-wide

Cytoscape- De-facto standard software in biological network

research community - Large User and Developer Community

- Expandable by Apps- This is why Cytoscape is a Platform, not a

simple desktop application

C. Elegans Interactome from BioGRID Database

?

Biological Networks

- Tell us anything by themselves - Just a big hairball…

Module 1

Module 2

In other words…

Module 1

Need a tool to extract meaningful biological modules

Basic Use Case

Networks

Public Interaction Databases

List of Genes

Other Data

Network Data Analysis

Analysis

Graph Analysis

NetworkX

igraph

Cytoscape

Python

Pandas

NumPy

SciPy

Excel

Visualization

Desktop

Gephi

Cytoscape

matplotlib

Web

Cytoscape.js

sigma.js

d3

NDV3

d3.chart

Google Charts

Data Storage

Graph

Neo4j

GraphXDocument

MongoDB

Relational

MySQL

IPython

3rd Party Apps

NetworkAnalyzer

Three Basic Steps for Data Visualization with Cytoscape

<?xml version="1.0" encoding="UTF-8"?> <graphml xmlns="http://graphml.graphdrawing.org/xmlns" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://graphml.graphdrawing.org/xmlns http://graphml.graphdrawing.org/xmlns/1.0/graphml.xsd">  <key id="degree" for="node" attr.name="degree" attr.type="double"/> <key id="betweenness" for="node" attr.name="betweenness"

attr.type="double"/> <graph id="G" edgedefault="directed"> <node id="n0"> <data key="degree">79</data> <data key="betweenness">0</data> </node> <node id="n1"> <data key="degree">9</data> <data key="betweenness">167</data> </node> <node id="n2"> <data key="degree">18</data> <data key="betweenness">75</data> </node> <node id="n3"> <data key="degree">8</data> <data key="betweenness">12</data> </node> <node id="n4"> <data key="degree">26</data> <data key="betweenness">210</data> </node> <node id="n5"> <data key="degree">29</data> <data key="betweenness">320</data> </node>

Data Integration

http://graphml.graphdrawing.org/xmlns

http://www.w3.org/2001/XMLSchema-instance

http://graphml.graphdrawing.org/xmlns

http://graphml.graphdrawing.org/xmlns/1.0/graphml.xsd

Analysis

Visualization

Drawing Biological Networks

VS

Drawing Tools

You need to specify color of each node, width of each edge, shape of nodes, etc.

There is one huge difference between Cytoscape and Illustrator…

In Cytoscape, Your Data Controls View

Creating Visualizations in Cytoscape

Name Type

BRCA1 gene

MAP2K1 gene

C05981 compound

• Mapping from Type to Node Shape • Mapping from Type to Node Color

C05981

BRCA1

MAP2K1

Creating mappings from data points to Visual Properties

Network Data

Annotated Networks

Attributes

Analyzed Data

Cytoscape Apps- Extension programs to

add new features to Cytoscape

- formerly called Plugins

- Large App developer/user community - This is why Cytoscape

is so successful in life science community!

Example Apps

ClueGO Creates and visualizes a functionally grouped network of terms/pathways

ReactomeFIPlugIn Explore Reactome pathways and search for diseases related pathways and network patterns using the Reactome functional interaction network

KEGGScape KEGG Pathway Importer for Cytoscape

clusterMaker2 Multi-algorithm clustering app for Cytoscape

cyREST (Now Part of The Core!) RESTful API for Cytoscape

cyREST RESTful API for Cytoscape

APPS.CYTOSCAPE.ORG

http://apps.cytoscape.org

Overview of App Ecosystem

A travel guide to Cytoscape plugins Rintaro Saito, Michael E Smoot, Keiichiro Ono, Johannes Ruscheinski, Peng-Liang Wang, Samad Lotia, Alexander R Pico, Gary D Bader, Trey Ideker (2012) Nature Methods 9 (11) p. 1069-1076

http://www.mendeley.com/library/showrecent/document/5072598951/#

Tips for Learning New Tools

Choose a Right Tool

Choose a Right ToolAnalysis VisualizationData Preparation

Data Visualization Tools

http://selection.datavisualization.ch/

http://selection.datavisualization.ch/

Tools

• In some cases, you can get exact same result using different tools

• Example: Data preparation (cleansing / cleansing)

• But if you choose right tools, you can do it 100x faster than others.

• ex: Re-formatting complex data sets

• Excel vs Python Script

• Some recommendations:

• R/Bioconductor, Python/Pandas, Git/GitHub/Gist

Learning Tools = Saving Your Time

Hands-on:

Introduction to Network Visualization with Cytoscape

Sample Data Files:

http://cl.ly/VTJsSame data is available in lecture1/dataset

http://cl.ly/VTJs

Data Visualization

- Goal: Help others to understand your data

- Emphasize what you want to tell

- Use color, shape,

size of objects effectively!

- Excellent resource for data visualization

- Tamara Munzner’s Web Site: http://www.cs.ubc.ca/~tmm/

Data Visualization

http://www.cs.ubc.ca/~tmm/

Today’s Goal

Story:

I want to show gene expression changes over time as a network diagram

YPL201C YPL211W YML007WYPL131W YOR327CYDR171W YCL067C

YCL032W YGL208WYER074WYBL050W YLR134WYPL149W

YDR050C YMR311CYGL134WYER102W YBR112CYKL101W YNL199C

YPL222W

YLR264W

YPL089C

YNL098C

YLL028W

YBR072W

YOR326W

YJR066W

YOR039W YNL135C YPR041WYDR174W YIL074C YKL028WYOR362C

YIL162W

YNL189W

YOR212W

YPR080W

YPR145W

YLL019C

YLR284CYPL031C YFR037CYML074C YPL240CYPR048W YBR274W YBR050C YML032C

YJR022WYBR248C YDR382W YER081WYIR009W YDR244W YOL016C

YER103W YGR058WYLR256WYAL003W YOR355WYIL061C YER111C YMR309C

YPL248CYOL127W YBR019CYLR362W YGL035CYPR167C YML123C

YBL026WYLL021W YNL091W YOR178C YIL113WYLR321C YML064C YMR117C YDL194WYNR007C

YOL058WYBR045CYER065CYNL167CYNL047C YGL097WYHR071W YDL078C YDL081CYDR354W

YER145C YGR136WYDR311W YPR119WYER112W YLR214WYCR012W

YER143W YBR043CYKL204W

YGR019WYEL041W YER133W

YOL149W YBR118WYAL038W YDR167WYMR058WYER079W YMR291W

YKL012W YDL113CYDR299W YDL075W YDL236W YGL229CYLR377C YNL145W

YNL236W YJL030W

YOL156W

YGL013C

YHR171W

YBL021CYMR021C

YHR174W

YFL038C

YER090WYPR062W YAR007C YNL307C YGL237CYML024WYDR335W YLR075W

YNL050CYGR046W YAL040CYLR191W YMR138WYIL045W YHR005C YNL301C

YKL211CYLR452C YPL075WYML051W YOL123W YGR088WYHR198C YMR300C

YJR060W YMR043WYPR124WYLR081W YLR319CYKL074C YOR036WYKL001C

YDR100W

YDR395W

YDR009W

YDR309C

YPR102C

YAL030W

YHR084W

YLR345W

YBR170C

YJL089WYFL026W YBR018C YGL115W YHR179WYDL215CYGR009C YOL120C

YFL017C YDR429C

YIL052C YGL073W

YGR108WYPR035W

YJL190CYOL086C YHR055CYBL005WYKR026C

YBR155W

YOR264W YKL109W

YOR167C

YDR070CYEL015W

YIL133C

YGL166WYHR030CYGL008C

YMR146C YBR160W

YOL136C

YOL051W

YBR020W YBR190WYDR323CYLR197W YFR014CYKL161C

YML054C YKR099WYLR340WYGL106W YBR093CYCL040W

YLR044C

YCR086WYDL130W

YJL203W

YEL009CYBR135W

YOR361C YGR085C

YER056CA

YNL216WYMR005W

YBR109C

YLR229C YER124C

YJL157C YDR461WYNL154CYLR117C YKR097W YIL069CYMR186W YJR109CYIL015W

YER040WYDR103W YGR074WYER052C YIL160CYOR290C YLR249W

YGL153WYOR215CYGR254W YLR432WYCR084CYOR089CYGR218W YOR303W

YGL161C YLR293CYDL030WYNL036W YHR135CYER179W YDR277CYDR184CYNL312W YML114C YFL039C YOL059WYER054C YER110CYLR109W YLR116WYNL214W YBL069W

YHR141CYER116CYJL219W YPL111WYDL023C YGL202WYER062C YMR183CYFR034CYGL122C

YIL105C YDL088CYPR010C YJR048W YIL070C YEL039CYDR412WYMR108W

YOR204W YMR255W YLR175W YHR115CYNL164C YJL013C YDL063C YNL117W

YIL143CYOR315W YDR146CYLR310CYGR014WYBR217W

YNR053C

YJL036W

YNL116W

YOR120W

YDL014W

YJL194W

YDL013W

YDR032C

YOR310C YPR113W

YLR153C

YGR048W

YGR203W

YNL113WYOR202W

YNR050C

YCL030C

YJL159W

YHR053CYPR110C?YLR258W YBL079W

YNL069C YNL311CYDR142C YGL044CYMR044W

What is Great Visualization…?

Design is complicated, because humans are complicated. Design is a process to avoid bad designs.

Mike Bostock (New York Times Visualization Team. Creator of D3.js)

It is hard to generalize the design process, but we can avoid pitfalls by following some basic rules.

Every pixel should carry information.

Edward Tufte

Avoid Data Overload

• Mapping too many attributes makes your visualization awful!

• It is hard to see the overall trend of your data sets if too many channels are used in a image

“Great Artists Steal…”

MSL5

TEM1

PRP40

MUD2

HAP4HAP2

CYC1

GCY1

HAP3

YHR198C

ECI1

YEL015W

GAL1

GAL7

GAL80

GAL3

GAL11

GAL4

GAL2

MLS1

SIP4

FBP1

GAL10

SWI5

SUC2

MIG1

ADH1PGK1

CDC19

GCR1

CBF1ENO1

ENO2

MCK1

CYC7

HAP1

CTT1

NCE103

SSL2

TFB1YNL091W

TRP4

ARG1

GCN4

SKO1

HIS3

ADE4 ILV2

TIF35

TIF5 NIP1

GNA1

PRE10

PRT1

YDR070C

GPD2

RPS17A

BAS1

HIS7

RPS24B

MSL1

HIS4

PDC5

PHO84

PHO4

YNL047C YIL105C

MET16

RPL11BRPS8B

RPL10

RPL11A

CKS1

RPL31A

PHO13

PDC1

SXM1RPL34B

RPL16B

ATC1

CAR1

FCY1

RFA2

ICL1SRP1

TPI1RPL18B

RPL25

PHO5

RPS24ARPL18A

DMC1 RAP1

RPL16A

HSP42

MSL5

TEM1

PRP40

MUD2

HAP4HAP2

CYC1

GCY1

HAP3

YHR198C

ECI1

YEL015W

GAL1

GAL7

GAL80

GAL3

GAL11

GAL4

GAL2

MLS1

SIP4

FBP1

GAL10

SWI5

SUC2

MIG1

ADH1PGK1

CDC19

GCR1

CBF1ENO1

ENO2

MCK1

CYC7

HAP1

CTT1

NCE103

SSL2

TFB1YNL091W

TRP4

ARG1

GCN4

SKO1

HIS3

ADE4 ILV2

TIF35

TIF5 NIP1

GNA1

PRE10

PRT1

YDR070C

GPD2

RPS17A

BAS1

HIS7

RPS24B

MSL1

HIS4

PDC5

PHO84

PHO4

YNL047C YIL105C

MET16

RPL11BRPS8B

RPL10

RPL11A

CKS1

RPL31A

PHO13

PDC1

SXM1RPL34B

RPL16B

ATC1

CAR1

FCY1

RFA2

ICL1SRP1

TPI1RPL18B

RPL25

PHO5

RPS24ARPL18A

DMC1 RAP1

RPL16A

HSP42

Map gene expression values to color

Avoid using more colors in other components (edge/label)

If necessary, map other data into non-overlapping visual properties

(edge score to width)

Part 1: Session File and Basic Navigation

Cytoscape 3.4 Desktop

Toolbar

Network Panel

Bird’s Eve View

Table Browser

Network Views

Local ColumnTable Tabs

List Data(Values in [ ])

Shared Column

Session File

- Snapshot of your workspace - Networks - Tables - Visual Styles - System Properties

Open a Session

- Click folder icon - Or, File → Open

Exercise 1: Loading a session

Navigation- Pan: Drag - Zoom

- IN: Mouse Wheel UP - OUT: Mouse Wheel DOWN

- Selection: Shift + Drag - Fit to Window

- Selected region - Entire network

First Neighbor of Nodes

CTR+6

Create New Sub-Network From Selection

CTR+N

- Grid View

- Detached View

Part 2: Data Import

Network Data Formats- SIF - GML - XGMML - GraphML - BioPAX - PSI-MI - SBML - KGML (KEGG) - Excel - Text Table - CSV - Tab

NCBI Gene ID 672

On Chromosome 17

GO Terms DNA Repair Cell Cycle

DNA Binding

Ensemble ID ENSG00000012048

BRCA1

Data Tables for Cytoscape- Example:

- Numeric- Gene expression profiles - Network statistics calculated in other

applications, such as R - Confidence scores for edges

- Text (or categorical)- GO annotation for genes - List of genes related to disease X - Targets for FDA approved drugs - Genes on KEGG Pathway Y - Clusters / group / community calculated

in external programs - …

Your Data Sets- Anything saved as a table can be

loaded into Cytoscape - Excel - Tab Delimited Document - CSV

- As long as proper mapping key is available, Cytoscape can map them to your networks.

Mapping Key in the Network

Mapping Key in the Table

Exercise 2: Loading network and tables

Part 3: Visualization

Layouts

Force-Directed + Edge Bundling

Stacked-Node Layout + Default Edge Bend

Circular + Edge Bend

Automatic Layout

- Choose proper algorithm - Tree-like data - Hierarchical Layout - Scale-Free Network - Force-directed - Circular process - Circular Layout

- Tweak parameters if necessary

Manual Layout

- Tweak result from automatic layout - Scale - Align - Rotate

Exercise 3: Apply layouts

Visual Style

- Collection of mappings from Attributes to Visual Properties

Visual Styles

- Defaults + Mappings - Expression values to node color - Gene function to node shape - Interaction detection method to edge line

type - Confidence score to edge width

Core Idea: Data Controls The View

Data Controls The View• Photoshop / Illustrator

• You control the pixels and objects on the display

• Data Visualization Tools (including Cytoscape)

• Data points are mapped to visual properties

• Color

• Size

Data Controls The View

Expression Values To Node Colors

Discrete Mapping Editor

Continuous Mapping Editor

Exercise 4: Create New Visual Style

Preview of Lecture 2: Reproducible Workflow with Jupyter Notebook and cyREST

Cline, Melissa S., et al. "Integration of biological networks and gene expression data using Cytoscape." Nature protocols 2.10 (2007): 2366-2382.

Results

Sharing Results

😐

Sharing Results and Process

😃

Point & Click Operation is Easy, but not Reproducible…

Goal: Reproducible Science

Goal: Reproducible Science

REST

Tools You Need

REST

- GitHub - For source code sharing

- IPython (Jupyter) Notebook - Your electronic lab notebook

- cyREST - RESTful API module for Cytoscape

-

- Two Google Groups

- [email protected]

- [email protected]

- ANY question is OK!

Getting Help