Daehee Hwang

Department of New Biology, DGIST

Systems approaches in

biology and medicine

2040 Visions of Process Systems Engineering

Daehee@MIT

Daehee@MIT

Systems Biology in Early Times

• Elements (genes, proteins, etc.)- nodes

• Interactions between the elements- edges

• Elements and their interactions are affected

by the Context of other systems within-

cells and organisms

• Interactions between/among elements give

rise to the system’s Emergent properties

Systems biology: dynamic network

Non-Diseased Diseased

Disease Arises from Disease-Perturbed Networks

System Perturbation

Generation of comprehensive global data

Identification of key molecules

Network modeling

Generation of hypotheses

Validation of hypotheses

Current Systems Biology

Resources and tools for systems approaches

Choi et al., Molecular Plant (2014)

System Perturbation

Generation of comprehensive global data

Identification of key molecules

Network modeling

Generation of hypotheses

Validation of hypotheses

Current Systems Biology

Analysis of genomic, proteomic & metabolomics data & data integration.

Network modeling and analysis for ID of key regulators and pathways.

Identification of differentially expressed molecules.

Computational and systems

biological tools

Pointillist

MS-BID/ Prequips

Mathematical modeling

tools

Principal network analysis

iNID

TEMPI

Integration of multi-omics data

PNAS, 2005a & b

Integration,

visualization, &

analysis of LC-

MS/MS

proteomics data

Bioinformatics,

2008a & b

Functions of key

temporal & spatial

regulatory motifs

Science, 2009;

Developmental

Cell, 2013

ID of subnetworks

representing major dynamics

Bioinformatics, 2011

Network analysis

for understanding

interplays among

signaling pathways

Molecular Plant,

2014

Probabilistic

modeling of

dynamic networks

with multi-

information

Bioinformatics,

2014

Development of tools for systems biology

iNID database: transcriptome & interactome

1,171,417 interactions for 27,103 molecules

Developmental signaling pathways

Interplays among developmental pathways

Interplays of hormone and pathogen pathways

Interplays among developmental pathways

Network models showing Interplays among

developmental pathways

Computational and systems

biological tools

Pointillist

MS-BID/ Prequips

Mathematical modeling

tools

Principal network analysis

iNID

TEMPI

Integration of multi-omics data

PNAS, 2005a & b

Integration,

visualization, &

analysis of LC-

MS/MS

proteomics data

Bioinformatics,

2008a & b

Functions of key

temporal & spatial

regulatory motifs

Science, 2009;

Developmental

Cell, 2013

ID of subnetworks

representing major dynamics

Bioinformatics, 2011

Network analysis

for understanding

interplays among

signaling pathways

Molecular Plant,

2014

Probabilistic

modeling of

dynamic networks

with multi-

information

Bioinformatics,

2014

Development of tools for systems biology

GI(GIGANTEA), an Important Modulator of the Core Oscillator of Arabidopsis

Harmer, Annu.Rev.Plant.Biol.(2009)

Kim et al., Nature (2007)

A model of the plant circadian clock

Mathematical modeling predicts network structure

Input 1Input 2

GI_cGI_n

LHY

expre

ssio

n

time

Difference?

Log10(probability)

Structure 3 yielded the highest likelihood of generating the wild type oscillatory expression of LHY

wild type

Mathematical modeling predicts network structure

Experimental confirmation by transgenic plants

Nuclear and cytosolic GI form incoherent feedforward loop to regulate LHY

Application of systems approaches

Application of these tools for various studies

The utility of these tools has been demonstrated in various systems biology studies.

Dynamic disease-perturbed networks during the course of prion disease

progression Molecular Systems Biology, 2009 & age-related networks of

Arabidopsis leaves along the lifespan Science, 2009; Plant Physiology, 2016.

Construction of RA-associated networks PLOS One, 2012; PNAS, 2015

describing pannus formation, and selection of key regulators of the networks

Arthritis and Rheumatism, 2011; JEM, 2012; PNAS, 2014

Construction of T2D-associated networks, selection of key regulators of the

networks Cell Metabolism, 2013; Diabetes, 2015, and elucidation of a retrograde

signaling mechanism from the networks Science Signaling, 2012

Construction of cancer-associated networks Nature Reviews Cancer, 2011 & 2012

and selection of key regulators of the networks JCI, 2012; Oncogene, 2014;

Cancer Research, 2016; Nature Communications, 2015; JI, 2015

Construction of network models related to other diseases, such as AD Cell Death

and Differentiation, 2014, PD Nature Communications, 2013, and Interstitial

Cystitis Molecular & Cellular Proteomics, 2011, and a metabolic network

underlying pharmacokinetics of tacrolimus CPT, 2010.

Temporal disease-perturbed networks:

Prion disease

Hwang et al., Molecular Systems Biology (2009)

Time-evolving cell cycle networks

Kim et al., Bioinformatics (2014)

Application of these tools for various studies

The utility of these tools has been demonstrated in various systems biology studies.

Dynamic disease-perturbed networks during the course of prion disease

progression Molecular Systems Biology, 2009 & age-related networks of

Arabidopsis leaves along the lifespan Science, 2009; Plant Physiology, 2016.

Construction of RA-associated networks PLOS One, 2012; PNAS, 2015

describing pannus formation, and selection of key regulators of the networks

Arthritis and Rheumatism, 2011; JEM, 2012; PNAS, 2014

Construction of T2D-associated networks, selection of key regulators of the

networks Cell Metabolism, 2013; Diabetes, 2015, and elucidation of a retrograde

signaling mechanism from the networks Science Signaling, 2012

Construction of cancer-associated networks Nature Reviews Cancer, 2011 & 2012

and selection of key regulators of the networks JCI, 2012; Oncogene, 2014;

Cancer Research, 2016; Nature Communications, 2015; JI, 2015

Construction of network models related to other diseases, such as AD Cell Death

and Differentiation, 2014, PD Nature Communications, 2013, and Interstitial

Cystitis Molecular & Cellular Proteomics, 2011, and a metabolic network

underlying pharmacokinetics of tacrolimus CPT, 2010.

Mitochondrial disorders

Defects in mitochondrial DNA

Retrogradesignaling

Changes in nucleus programming

Development of diverse disease

phenotypes

A Cybrid model with mtDNA 3243 mutation

mtDNA 3243 A>G mutation

Causes 1% of Diabetes

Cybrid Model

Dr. Park (SNU)

System-level Analysis of Cybrid Models

High glucose

(4.5g/L)

Hetero

Mutant

Wild

Proteome analysis20 50 80 110 140 170

Transcriptomic analysis

Chae et al., Science Signaling (2013)

-3 -1 0 1 3

C1

C2

C3

C4C6C5

Mutation-load Dependent Differential Expression Patterns

Number of DEGs UP DOWN Total

OXPHOS 2 46 48

Translation 28 69 97

0 100 200 300 400 500 600

C6

C5

C4

C3

C2

C1

Number of Genes

239

304

523

112

68

90

0 20 40 60 80 100

C6

C5

C4

C3

C2

C1

Number of regulators

(TFs / Signaling molecules)

17

22

25

9

6

6

31

46

55

27

17

21

Number of translation

related genes

Number of OXPHOS

related genes

38

53

40

5 14

7

4

1

C1C2

C3C4

C5C6

What Regulates Changes in C3 ?

TF

Name

No. of

genes

Regulated

P-value

CREB1 124 < 1.0E-05

GABPA* 112 < 1.0E-05

ELF1 100 < 1.0E-05

NRF1* 65 < 1.0E-05

ELK1 63 < 1.0E-05

YY1* 60 < 1.0E-05

RXRA* 32 0.0003

ESRRA* 35 0.0031

ARNT 28 0.00668

RELA 29 0.00749

RXRA

PPARGC1α

NRF1

GABPA

OXPHOS Translation

9-cis RA

Future in systems biology & medicine

EBV+ Hypermutable Genome-stable High SCNA (Aneuploid)

Hypermutable

Multi-dimensional genomic analysis of cancers

Different size,

scale,

coverage, etc.

Complementary nature of different omics datasets

Single cell systems biology

전이 대장암: 생존율 감소

Stage 1

Stage 4

Stage 2

Stage 3

Cellular heterogeneity Cancer stem cells & drug resistance

Single cell transcriptomics

35

Cancer stem cells & drug resistance

Big data analysis

Finding relationships

Big data analysis

Finding rules

Other big data problems in proteomics

Finding co-occurring post-translational modifications in the same proteins and other

proteins in the same pathway???

Finding frequently co-varying proteins in different pathways (cross-talks between pathways) ???

Finding frequently co-varying network modules (pathways) that can be used as bases for

interpretation of a new dataset???

Network architecture

Edge: Correlation in peptide ratios

Edges: - Relationships based on PTMs- Co-occurrence rate

Edge: Protein-protein interactions

Happy Birthday !!

Happy Retirement !!