"Improving the description of metabolic networks: the TCA cycle as example". In:

Brend Wanders Paul van der Vet Maurice van Keulen

<b.wanders@utwente.nl>

PayDIBI Pay-as-you-go data integration for bioinformatics

ACO1 ACO2

IDH3A IDH3B IDH3G IDH1

IDH2

SUCLA2

SUCLG1 SUCLG2 SUCLA2P1

LOC283398 SDHA SDHB

SDHD SDHC FH

MDH1

MDH2 MDH1B

CS SLC35G3

citrate

cis-aconitate

D-threo-isocitrate

3-carboxy-1-hydroxypropyl-ThPP

succinyl-CoA

2-oxosuccinate

succinate fumarate

(S)-malate

oxaloacetate

ACO1 ACO2 IREB2

ACO1 ACO2

OGDH OGDHL

DLST

SUCLA2 SUCLG1 SUCLG2

SUCLA2

SUCLG1 SUCLG2

IDH1

IDH2 IDH2

IDH1 IDH2

IDH1 IDH2

OGDH OGDHL SDHA

SDHB

SDHD SDHA SDHC

SDHB

SDHD SDHC

H2O H2O

NAD

H⁺ NADH

NAD NADP

CO2

NADPH H⁺

NADH CO2

H⁺

ubiquinol

ubiquinone FADH2

FAD⁺

CoA ATP

ADP GDP

IDP ITP

acetyl-CoAH2O

H⁺ CoA

NAD

NADH H⁺

H2O

Pi CoA

Pi GTP

CoA

NADP

H⁺ NADPH

NAD CoA

NADH CO2

H⁺

CO2

ThPP

CO2

lipoamide

Pi

NADH

dihydrolipamide CoA

S⁸-succinyldihydrolipoamide

NAD H⁺

5-7 8-9

1 2-4

10 Legend

Number of databases that agree

pyruvate

CO2

NADH

NAD

ATP CoA acetyl-CoA

Pi ADP ACLY

acetate

CLYBL CO2

ITP PCK1 PCK2

GTP CO2

PCK1 PCK2 GDP phosphoenolpyruvate

dihydrolipoamide CoA S-acetyldihydrolipoamide pyruvate

ThPP CO2

2-hydroxyethyl-ThPP

ThPP lipoamide

NAD⁺ NADH PDHA1

PDHB PDHA2

PDHA1

PDHB PDHA2

DLD

pyruvate

ATP HCO3-

ADP ATP CO2

H2O Pi ADP

PC

DLAT

Pi PC

IDP

PDK1 PDK2 PDK3 PDK4

PDHA1 PDHB DLAT CoA

CoA

oxidized ferredoxin

CO2

H⁺ reduced ferredoxin

2-oxoglutarate

DLD succinate semialdehyde

CO2

H2O NAD or NADP

H⁺ NADH or NADPH

acetoacetate acetoacetyl-CoA

acceptor reduced acceptor

acceptor

reduced acceptor

DLD DLST

OGDH DHTKD1

OGDHL PDHX NNT

NAD H⁺ H⁺ NADH

ThPP

citrate Main compound:

Compound: H2O

ACO1 Gene encoding

for enzyme:

H⁺

Arrow indicates the direction of the reaction on which the majority

of the databases agrees.

A

B

C

D

E

F

G H

I

J

ThPP: thiamin diphosphate Pi: hydrogenphosphate

Image courtesy of M.D. Stobbe et al.

"Improving the description of metabolic networks: the TCA cycle as example". In:

The FASEB Journal 26.9 (2012), pp.

Union of the descriptions of the TCA cycle given by 10 pathway databases.

Colors indicate the level of agreement on a reaction and on the genes.

1 Combination of multiple data sources, each created with a specific purpose in mind. This first integration attempt will have inconsistencies and semantic

mismatches.

2 The bioinformatician analyses the combination

attempt to determine whether the current level of integration is good enough for their purposes.

3 Typical usage of integrated data by the bio-

informatician is exploring and querying to gain

insight into the structure and properties of the data.

The system can explain the results of a query.

4 Based on their analysis and exploration, the

bioinformatician can issue feedback and improved

integration rules to the system. These rules are then stored in the personal knowledge base.

5

6 When the integration level is

to the bioinformatician's liking, they can stop iterating over the

analyse-explore-feedback loop and use the integrated data. At

a later point, the bioinformat- ician can come back to step 5

to continue refining the integration.

A new combination attempt is created in the next iteration, allowing the bioinformatician to continue

refining their rules. This allows the bio-

informatician to spend effort only when necessary.

+ =

1

2

3

4

5

Personal KB

6