Integrating -Omics Brent D. Foy, Ph.D. Associate Professor Department of Physics Wright State University Dayton, OH.

Integrating -Omics

Brent D. Foy, Ph.D.

Associate Professor

Department of Physics

Wright State University

Dayton, OH

2

Overview

• Combining Genomic Data with Proteomic Data

– Which gene makes which protein?

– If mRNA level goes up, does the protein level go up?

• Biomolecular Network Modeling

– Issues

– State of the Field

– Our work

3

Gene to Protein Identification

Partial table from Affymetrix rat gene tox chip

The ‘J02722’ is the GenBank nucleotide ID for this gene.

4

Gene to Protein Identification

• A Search for ‘J02722’ on GenBank (http://www.ncbi.nlm.nih.gov/Genbank/) or EBI (http://www.ebi.ac.uk/cgi-bin/emblfetch) brings up gene information page.

• Scroll down for protein id. GenBank gives link for ‘AA41346.1’. EMBL gives links for EPD: ‘EP31003’ and Swiss-Prot: ‘P06762’. Clicking on link takes to information page on protein.

• Match up Affymetrix gene id with protein id provided by proteomics experiment.

• Can do reverse, given protein id, find gene id.

5

Gene to Protein Identification

• Since we have ~150 identified proteins from proteomics, and ~1000 genes on Affymetrix gene chip, we did the reverse approach (given protein, find mRNA), and found 21 genes corresponding to 16 proteins that were present in both.

• Discrepancy?

– AFFY and GenBank # M25157 – Rat Cu, Zn superoxide dismutase, from Sprague Dawley, lung cell line, 601 base pairs

– AFFY and GenBank # Y00404 - Rat mRNA for copper-zinc-containing superoxide dismutase, from Sprague Dawley, liver, 650 base pairs

– Errors in public databases, or just incomplete knowledge of mRNA or protein varieties

6

Change in mRNA Expression vs Change in Protein Expression

Ratio of expression in absence of galactose to expression in presence of galactose

Ideker T, et al., Science, 292: 929-934, 2001.

7

mRNA Expression vs. Protein Level

Protein level vs gene expression,ratio 75 mM to 0 mM, different times

gene expression, ratio 75 mM to 0 mM, t = 0

0 1 2

pro

tein

leve

l, ra

tio

75

mM

to

0 m

M, t

= 3

0

1

2

Control caseNo hydrazine exposure

gene expression level

1 10 100 1000 10000 100000

pro

tein

leve

l

1

10

100

1000

10000

100000

8

Time Course – mRNA and Protein Levels

50 mM Hydrazine-exposed Hepatocytes

Immunoglobulin Heavy Chain Binding Protein

0

5000

10000

15000

20000

25000

-5 0 5 10 15 20 25 30

Protein

mRNA

N-hydroxy-2-acetylaminofluorene; sulfotransferase

0

500

1000

1500

2000

2500

-5 0 5 10 15 20 25 30

Protein

mRNA

Heme Oxygenase (HSP32)

0

2000

4000

6000

8000

10000

12000

-5 0 5 10 15 20 25 30

Protein

mRNA

Soluble Cytochrome b5

01000200030004000500060007000

-5 0 5 10 15 20 25 30

Protein

mRNA

9

Biomolecular Network Modeling

TranscriptomeAnalysis

ProteomeAnalysis

MetabolomeAnalysis

Protein - Pia

Protein - ProteinInteractions - Pia...

Metabolic Pathways

mRNAia

ProteinModifications - Pia1

Cellular Metabolites

Genei*

rRNA

tRNA

GenomeAnalysis

Pre-mRNAi

Genome

mRNAib

snRNA

mRNAij

Sk Mk

Action Pathways

Control Pathways

10

Metabolic Network Modeling -Tracer studies

• Quantify activities of biochemical pathways

• For example, C-13 NMR analysis of TCA cycle and gluconeogenesis in liver

lactate

pyruvate acetyl-CoA

oxaloacetate

F1 F6

F2

F3

a-ketoglutarate glutamate

fumarate

F4

F5F7

glucose

F8

F9

plasmalactate

ILOL

plasmaglucose

OG

IG

plasmaglutamate

ITOT

lipid + acetate

IAOA

F10

11

Genetic Regulation

• Genes expressed in distinct domains, precisely delineated by time, state of cell, and level of response.

• This control is exerted by regulatory elements in the promoter and enhancer regions of genes.

• Field still young, but some quantitative results are appearing.

A B A C D mRNA sequence

DNA

Regulatory factors

• Feedback with other genes

12

Biomolecular Network Modeling – Issues

• Compared to standard modeling of kinetic processes, challenges include:

– Stochastic reaction behavior due to random diffusion processes and small numbers of molecules

– Multiple protein-protein, protein-mRNA, etc. interactions

– computational efficiency, parallelized code for operation on multiple CPUs

– Can you separate out the model for a pathway from the whole cell?

13

Biomolecular Network Modeling – Task

gene A mRNA A prot A rxn A1 A2

gene B mRNA B prot B rxn B1 B2

gene C mRNA C prot C

gene D mRNA D prot D

•Compounds other than genes are mobile

•Some of these mobile compounds affect many reactions (e.g. ATP, ions)

14

Biomolecular Network Modeling – Finding the Parameters

Use the simulation itself to narrow down on the possibilities

1. Optimize on stability

Parameter 1P

aram

eter

2

Stable regions

2. Optimize on something else:maximum energy efficiencyrapid cell division

15

Biomolecular Network Modeling - State of the Field

• E-Cell

• Virtual Cell

• Bio-Spice/Arkin

• Specific Laboratories – Institute for Systems Biology/Leroy Hood’s group

• Useful links page: http://www.cds.caltech.edu/erato/links.html

16

E-Cell

• From Laboratory for Bioinformatics, Keio University, Japan

• Attempt to integrate genes, RNA, proteins, and metabolites of entire cell in one simulation

• Freely available, http://www.e-cell.org/

17

E-Cell

• Used to simulate a “minimal cell” based on Mycoplasma genitalium

• 127 genes

• Integrate with online databases

• Many parameters estimated

• Substances modeled include small molecules, macromolecules, multi-protein complexes, protein-DNA complexes

• Multiple reaction types

18

Time Time

ATP Some mRNA levels

Remove glucose from culture medium

Tomita, M., et al.; Bioinformatics, Volume 15, Number 1, 72-84 (1999)

E-Cell, published results

19

Virtual Cell

• National Resource for Cell Analysis and Modeling (NRCAM), located at University of Connecticut Health Center

• Access via internet, http://www.nrcam.uchc.edu/

• Has a graphical, “biological users” interface

• Compared to E-Cell

– Includes 3-d spatial information within cell

– Has not been applied to gene->mRNA->protein->metabolites

20

Define physiology, with reactionsamong substances

Virtual Cell

21

Geometric results

Virtual Cell

22

Bio-Spice

• Initiated at Berkeley National Laboratory, http://gobi.lbl.gov/~aparkin/index.html

• Development of Bio-Spice is currently the subject of a DARPA project

• It will be a Simulation Program for Intra-Cell Evaluation, like SPICE for circuit design

• Intended to be a “user-friendly simulation tool that captures the network of molecular interactions including gene-gene, gene-protein, and protein-protein interactions.”

23

Institute for Systems Biology - Galactose in Yeast

Ideker T, et al., Science, 292: 929-934, 2001.

24

Circles are genes, yellow means product affects another gene’s transcription, blue means proteins interact. Grayscale of circles is mRNA change with galactose in medium.

ISB - physical interaction network

Ideker T, et al., Science, 292: 929-934, 2001.

25

Development of Quantitative Tools - Transcription

B A TATA mRNA sequence

DNA

Regulatory factors

RNAPolymerase

TFIIITF_ATF_BActivated

Nucleotides

26

Development of Quantitative Tools - Transcription (cont.)

State of Promoter kon for RNA Polymerase

TATA A Boff any any 1e-99 (M*s)-1

on off off 1e-30on on off 5e-23on off on 1e-99on on on 5e-23

27

Development of Quantitative Tools - Transcription (cont.)

B A TATA product = TF_A

Gene A

A TATA product = TF_B

Gene B

Plus a first-order process for degradation of TF_A and TF_B

28

Development of Quantitative Tools - Transcription (cont.)

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

x 104

0

5

10

15

20

25

30

35

40

45

Time (s)

# A

mR

NA

mo

lec

ule

s

Time course of number of TF_A

POLYMERASE1696 events29.53% on

TFIII3967 events99.62% on

TF_A5 events51.45% on

TF_B1852 events45.97% on

0.5 1 1.5 2 2.5 3 3.5 4 4.5

0

0.5

1

1.5

2

x 104

Time course of binding to gene A promoter

29

Biomolecular Network Modeling - Future Tasks

• Ultimate goal is to provide physiological insight on integrated genomic, proteomic, metabolic data sets in response to toxicity interventions

• Establish contact with online databases– Gene->protein->metabolite connections (KEGG, others)– protein-protein interactions (published list, Nature Biotech)– protein-DNA interactions (TRANSFAC, SCPD)

• Evaluate proper scale of modeling effort relevant to task. Scale in both the level of biological detail, and in terms of man-hours.

• Choose software and gain expertise with it, or create software as needed.

• One early goal - explore minimal cell and its stability in response to perturbation

30

Collaborators

AFRL

Dr. John Frazier

Dr. Charles Wang

Dr. Victor Chan

AFOSR

Dr. Walt Kozumbo

AFIT

Dr. Dennis Quinn

2Lt Matt Campbell

WSU

Dr. Tatiana Karpinets

31

Integrating -omics

Questions?