A Novel Grid-based Visualization Approach for Metabolic Networks with Advanced Focus&Context View Markus Rohrschneider 1 , Christian Heine 1 , Andr´ e Reichenbach 1 , Andreas Kerren 2 , and Gerik Scheuermann 1 1 University of Leipzig, Department of Computer Science, Germany 2 V¨ axj¨ o University, School of Mathematics and Systems Engineering, Sweden Abstract. The universe of biochemical reactions in metabolic pathways can be modeled as a complex network structure augmented with domain specific annotations. Based on the functional properties of the involved reactions, metabolic networks are often clustered into so-called pathways inferred from expert knowledge. To support the domain expert in the exploration and analysis process, we follow the well-known Table Lens metaphor with the possibility to select multiple foci. In this paper, we introduce a novel approach to generate an interactive layout of such a metabolic network taking its hierarchical structure into account and present methods for navigation and exploration that pre- serve the mental map. The layout places the network nodes on a fixed rectilinear grid and routes the edges orthogonally between the node po- sitions. Our approach supports bundled edge routes heuristically mini- mizing a given cost function based on the number of bends, the number of edge crossings and the density of edges within a bundle. 1 Introduction To fully comprehend and appreciate the existing knowledge on chemical pro- cesses in living organisms it is essential to develop suitable tools to explore and navigate through vast amounts of information stored in biological databases. In biochemistry, complex networks defined by interactions and relations between different chemical compounds are considered as pathways, such as regulatory pathways controlling gene activity or metabolic pathways comprising chemical reactions for synthesis, transformation and degradation of organic substances in biological systems. In this work, we combine and apply information visualization techniques to present the complete set of biochemical reactions of metabolic pathways in a eucaryotic cell supplying means of exploration and navigation. Although the emphasis of this paper is placed on biochemical network data, the presented application is not limited to this area. Instead, it can handle any large graph carrying arbitrary annotational information by mapping given data properties to attributes being visualized by the software. To capture the complex chemical interactions of such a reaction network, metabolic pathways may be modeled as
12
Embed
A Novel Grid-based Visualization Approach for Metabolic …homepage.lnu.se/staff/akemsi/pubs/kerren-gd09b-postprint.pdf · A Novel Grid-based Visualization Approach for Metabolic
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
A Novel Grid-based Visualization Approachfor Metabolic Networks
with Advanced Focus&Context View
Markus Rohrschneider1, Christian Heine1, Andre Reichenbach1,Andreas Kerren2, and Gerik Scheuermann1
1 University of Leipzig, Department of Computer Science, Germany2 Vaxjo University, School of Mathematics and Systems Engineering, Sweden
Abstract. The universe of biochemical reactions in metabolic pathwayscan be modeled as a complex network structure augmented with domainspecific annotations. Based on the functional properties of the involvedreactions, metabolic networks are often clustered into so-called pathwaysinferred from expert knowledge. To support the domain expert in theexploration and analysis process, we follow the well-known Table Lensmetaphor with the possibility to select multiple foci.In this paper, we introduce a novel approach to generate an interactivelayout of such a metabolic network taking its hierarchical structure intoaccount and present methods for navigation and exploration that pre-serve the mental map. The layout places the network nodes on a fixedrectilinear grid and routes the edges orthogonally between the node po-sitions. Our approach supports bundled edge routes heuristically mini-mizing a given cost function based on the number of bends, the numberof edge crossings and the density of edges within a bundle.
1 Introduction
To fully comprehend and appreciate the existing knowledge on chemical pro-cesses in living organisms it is essential to develop suitable tools to explore andnavigate through vast amounts of information stored in biological databases. Inbiochemistry, complex networks defined by interactions and relations betweendifferent chemical compounds are considered as pathways, such as regulatorypathways controlling gene activity or metabolic pathways comprising chemicalreactions for synthesis, transformation and degradation of organic substances inbiological systems.
In this work, we combine and apply information visualization techniques topresent the complete set of biochemical reactions of metabolic pathways in aeucaryotic cell supplying means of exploration and navigation. Although theemphasis of this paper is placed on biochemical network data, the presentedapplication is not limited to this area. Instead, it can handle any large graphcarrying arbitrary annotational information by mapping given data propertiesto attributes being visualized by the software. To capture the complex chemicalinteractions of such a reaction network, metabolic pathways may be modeled as
Fig. 1. A hyperedge depicting a transaminase reaction, which converts amino acidsinto corresponding alpha-keto acids and vice versa. In this example, the enzyme As-partate Aminotransferase converts the substrates L-Aspartate (amino acid) and 2-Ketoglutarate (alpha-keto acid) [input nodes] into the products Oxaloacetate (alpha-keto acid) and L-Glutamate (amino acid) [output nodes]. Many reactions are reversible,so the direction of the hyperedge simply gives a hint on the reaction’s chemical equi-librium.
hypergraphs, where unlike a regular graph, each edge can connect an arbitrarynumber of nodes. In this hypergraph model, each substance is represented bya node of the graph, and each reaction by a (directed) hyperedge connectingthe input node set—substrates—with the output node set—products— of thechemical reaction (see Fig. 1). To obtain a hierarchical graph, each metabolicpathway is represented by a node at the top level, where the pathway’s reactionnetwork constitutes the nested graph at the bottom level. The division intoseparate pathways, although based on expert knowledge, is somewhat arbitraryand may not be a strict partition of the graph. Nevertheless, we consider theclustering of the node and edge set as a partition to obtain a strictly confinedhierarchy on the graph. Compound nodes and reactions belonging to more thanone pathway are simply duplicated for the sake of simplicity of the resultinggraph. This step has two benefits: layouting the graph will be a much simplertask, and we can use the hierarchy to explore the network in a top-down mannerby examining the top-level graph at first and adding additional information onpathways of interest by expanding nodes.
2 Related Work
The visualization of large and complex biological networks is one of the key anal-ysis techniques to cope with this enormous amount of data. Here, the layout ofnetworks should be in agreement with biological drawing conventions and drawattention to relevant system properties that might remain hidden otherwise [14,13]. Further important issues are the preservation of the so-called mental map [1]when applying small changes to the graph and the possibility of clustering nodes.Depending on the concrete network drawing, there are further important visualrepresentation and interaction techniques that play important roles, e.g., nav-igation in the complete network, focusing on parts of the network, or gradualdifferentiability of nodes with less importance (side metabolites) [12]. However,only little research has been done in the past to solve the special layout andvisualization problems arising in this area. A lot of the most used software sys-tems for the visual analysis of generic biological networks, i.e., different kinds of
Visualizing Metabolic Networks using Advanced Focus&Context Techniques 3
networks like regulatory networks or protein-protein interactions, only provideimplementations of standard graph drawing algorithms, such as force-directedor hierarchical approaches [8].
Cytoscape [5] is one of the most popular tools for generic biochemical net-work visualization and supports a number of standard graph layout algorithms.Filtering functions are provided to reduce network complexity. For instance, theuser can select nodes and edges according to their name and other attributes.This system allows a simple mapping of data attributes to visual elements ofnodes and edges. VisANT [4] is another system designed to visualize genericbiochemical networks. In addition to the features of Cytoscape, it provides sta-tistical analysis tools, e.g., based on node degrees or the distribution of clusteringcoefficients. Their results are displayed in separate views, such as scatter plots.
Especially for metabolic networks, large and hand-drawn posters were pro-duced in the past, for example, Nicholson’s pathway map [9] or the widely-used metabolic pathway poster published by Roche Applied Science [18]. Otherprojects have created graphical representations of metabolic networks and offerthem via web pages (e.g., the BioCyc collection [7]). The widespread pathwaydrawings of the Kyoto Encyclopedia of Genes and Genomes (KEGG) database[16], see also Section 3, were also produced by hand. These drawings are con-nected via links, but real interaction is not available. Because of their manualgeneration, they are well readable and can thus serve as an example in terms ofquality and user conventions. Moreover, the availability of these representationshas established a de facto standard for metabolic network drawings: it featuresnear-orthogonal drawings where, for example, important paths are aligned orrelevant subgraphs are placed close to the center of the drawing [14, 13].
Newer approaches are based on a close interdisciplinary work between re-searchers in visualization and biochemistry. An example is the Caleydo frame-work [21] that extends the standard pathways of KEGG into 2.5D, similar to thereport of Kerren [12] and the work of Brandes et al. [15], combined with brush-ing, highlighting, focus&context, and detail on demand. In this way, it supportsthe interactive exploration and navigation between several interconnected (butstatic!) networks.
Saraiya et al. [19] discussed the requirements of metabolic network visualiza-tion collected from interviews with biologists. They observed five requirementsthat are important for biologists working on pathway analysis, but still not com-pletely realized in existing visualization systems (adapted from [10]):
1. automated construction and updating of pathways by searching literaturedatabases;
2. overlaying information on pathways in a biologically relevant format;3. linking pathways to multi-dimensional data from high-throughput experi-
ments, such as microarrays;4. overlooking multiple pathways simultaneously with interconnections between
them; and5. scaling pathways to higher levels of abstraction to analyze effects of complex
molecular interactions at higher levels of biological organization.
4 M. Rohrschneider et al.
Currently, our approach addresses several of the aforementioned requirementsand improves the most previous work by using of interaction techniques frominformation visualization. Our new, interactive layouts are based on the KEGGdata (Req. 1), and we provided the visualization with an intuitive focus&contextview. In this way, we can handle, for example, the complete metabolism of ageneralized eucaryotic cell (Req. 4) by following Shneiderman’s mantra [20]:overview first, zoom and filter, details on demand. If the user explores the path-ways interactively, the visualization approach preserves the mental map. To thebest of our knowledge, no other system can provide that to this extent. Oursystem is also able to embed textual information into the drawings and to useglyphs/icons for the representation of lower-level subgraphs if needed, similar tothe Pathway tools [3]. The integration of more complicated attributes as wellas biological patterns regarding topological substructures are still missing. Here,other tools, such as BioPath [6], still have an advantage to be fully accepted bybiologists.
The generation of the actual layout of the hierarchical pathway graph is mo-tivated by the style of the “official” KEGG diagrams to be consistent with thedomain experts expectations. The diagrams usually use the orthogonal style fordrawing edges. To avoid overlapping labels, we ensure a minimum separationof the diagrams elements by using a regular grid based approach. Algorithmsfor orthogonal grid drawing have been widely studied; we cannot provide an ex-tensive overview here and refer the reader to [8, 26] for an introduction. Theseapproaches often follow a topology-shape-metrics approach [25]: First, computea planar embedding of the input graph, possibly planarizing it by augmentingvertices at crossings, second, compute an orthogonal representation of the em-bedding, and finally generate coordinates by compaction of the orthogonal repre-sentation. Usually, edges are not allowed to run simultaneously on the same gridsegment, i.e., connection between two neighboring grid positions. The pathwaysof the KEGG database can be converted into graphs, but a planarization of themrequires an enormous amount of augmented vertices. If edges are not allowed torun on the same grid segment, their layout dominates the area of the drawingresulting in poor resolution. Furthermore, as a pathway constitutes a semanticentity, they should be presented as a unit and without diagram elements fromforeign pathways interfering. No existing orthogonal drawing algorithm was ableto take these constraints into account, therefore we developed our own that doesnot planarize the graph but keeps track of edge crossings and heuristically mini-mizes them, allows “edge bundling” although it penalizes it, shows pathways asunits and performs dynamic compactions based on the currently focused partsof the pathway hierarchy.
3 Network Data Source
The development of graph interaction techniques especially suited to fit bio-logical problems makes it necessary to experiment with realistic datasets. Togenerate artificial graph data is of course possible, but it is hard to estimate therequired complexity of such datasets to simulate realistic scenarios. The Kyoto
Visualizing Metabolic Networks using Advanced Focus&Context Techniques 5
Encyclopedia of Genes and Genomes (KEGG, [16]) System provides annotatedpathway data facilitating the construction of metabolic pathway graphs of differ-ent sizes. KEGG is one of the major bioinformatics resources publicly accessible.It integrates genomic, chemical, i.e., molecular, and systemic functional informa-tion describing cellular processes and organism behavior. It provides a knowledgebase for systematic analysis in bioinformatics research and the life sciences. Weextracted the hypergraph structure including semantic information as discussedin [17]. The constructed graph covers the complete metabolism of a general-ized eucaryotic cell and contains 4980 compound and 154 pathway nodes, 4943reactions and 1248 inter-pathway edges.
4 Hierarchical Orthogonal Grid Layout
A hypergraph H = (V,E) as an extension of the graph concept allows the ele-ments of E called hyperedges to connect multiple vertices. Conceptually, a chem-ical reaction can be described as a hyperedge between compounds that are mod-eled as vertices. This requires a mark whether a vertex is a substrate or product.We model the data in the KEGG database as a hierarchy of one top-level graphthat contains a vertex for each pathway and one hypergraph per pathway. Iftwo pathways exchange compounds according to KEGG, both a regular edgeexists between them in the top level graph as well as an edge between the twohypergraphs representing the two pathways.
The layout of the hierarchical KEGG hypergraph is generated by convertingthe hypergraphs of the hierarchy into their corresponding bipartite graphs andcomputing a layout of this graph hierarchy. The generated layout is orthogonalto match the style of the official KEGG diagrams. Furthermore, its vertices’positions lie on a grid to ensure both a minimum separation between labels and tomake the algorithm both simpler and faster. The layout algorithm allows multi-edges but no loops and proceeds recursively—parents before their children. Foreach graph of the hierarchy, the layout consists of three phases named: VertexPosition, Edge Routing, and Edge Bundling.
In the Vertex Position phase, we try to find a unique integer positionfor each vertex that minimizes the stress: the amount of error that takes placeby the projection of the “high-dimensional” graph-theoretic distances to thegeometric distances between the vertices positions. As vertices and edges arelaid out on a regular grid, the Manhattan distance is used as geometric distance.When an edge leads outside the graph, its hierarchy parent has already beenlaid out. Thus, the direction from which the edge enters the graph is known.For each of the four orientations, we temporarily add a port vertex to the graphand connect the edges to foreign graphs to that port. Unlike the other graph’svertices the position of ports is fixed on the boundary of the grid in the upcomingoptimization phase.We implemented a stress minimization algorithm inspired by Kamada and Kawai[22]: starting from an inital random integer positioning of vertices, we select avertex with high local stress and find a continuous position for that vertex whereits local stress becomes minimal using the Newton-Raphson method. We insert
6 M. Rohrschneider et al.
it then at the closest integer position not taken by any vertex. We comparedthis method with two different approaches. A brute force version picks a randomvertex and tests all integer positions in a vincinity of its position and insert it atthe best position. A simulated annealing [23] variant picks a random vertex andnew position in the vincinity of the old position, but performs the insertion onlyon improvement and deterioration with decreasing likelihood. We found that thebrute force method optimized quality as it is difficult to trap this method in alocal minimum. The simulated annealing is very fast, but does not provide nearlythe same quality. All heuristics terminate after a fixed number of iterations thatis proportional to the number of vertices.The Edge Routing phase computes a combinatorial description of an edgerouting along the edges of the regular grid. The vertices’ positions are not alteredby this phase. The combinatorical description is computed one edge at a timeand after all edges have been processed once, an iterative process removes singleedges and adds them again optimizing on the global cost of the layout. Given acombinatorial description of the current edge routing, we construct a route graphthat consists of the original graph’s vertices and the grid’s edges as vertices andedges for valid transitions between these elements. Given this representation,we are able to compute the optimal routing of an edge by solving a single-pair-shortest-path instance on the route graph. The optimality is given by a costfunction that takes the number of crossings, the number of bends, the lengthof an edge, and the “density” of edges on a grid segment into account. Notethat the quality of the resulting configuration depends both on the originalvertices’ positions and the actual order of edge insertions. Good performancewas achieved when inserting the edges in the order of increasing distances oftheir incident vertices. To reduce runtime and memory consumption, we usethe A∗ search algorithm [24] to solve the SPSP instance using the Manhattandistance as heuristic.
The Edge Bundling phase shifts segments of edges’ routes orthogonal tothe grid segments they lie on to remove overlaps. It preserves the edges’ relativeordering and straightens them in the process. This problem can be solved for eachrow and each column separately. We generate for each row and each column adirected acyclic graph that contains line segments as vertices and edges betweenthese lines, if they are ordered in the combinatorial edge routing. Any topologicalnumbering of this graph gives a displacement that avoids occlusions betweenedge routes of the same column/row, and using the topological numbering ofminimum weight packs the edge bundles nicely together.
5 Graph Interaction
The graph interaction and exploration methods described in this section haveall been implemented in our visualization software. The grid layout algorithmis the central component of the adapted Table Lens method to explore hierar-chical graphs. We firstly present this technique with supplementary search andhighlighting operations and explain later how the graphical user interface letsthe user apply these methods to interact with the metabolic network graph.
Visualizing Metabolic Networks using Advanced Focus&Context Techniques 7
5.1 Exploration Techniques
Two fundamental navigation operations on hierarchical graphs are node expan-sion to reveal the node’s nested graph and collapse. For 2D graph representa-tions, it is natural and desirable to present a flat graph at all times regardlessthe graph’s expansion state. This means that the expansion of a node requiresit to be hidden and replaced by its nested graph. The inverse operation replacesthe nested graph by its parent. The well-known Table Lens metaphor [2] appliedto hierarchical graph exploration fulfills this requirement. It is an establishedfocus&context method to give an overview on large tabular datasets to examineobvious patterns and to provide detailed view on specific items at the same time.In our application, pathway nodes at the top level are placed in the center of acell, edges are routed along the cell borders as intended result of the previouslypresented layout algorithm. When a node is expanded, the row and the columnare enlarged in which the node is situated. Edges leading to and from one of thefour ports (see Sec. 4 and Fig. 4 for example) of the pathway node are elongatedwhile the remaining elements keep their relative position.This approach followsBen Shneiderman’s mantra of visual information-seeking: overview first, zoomand filter, details on demand [20]. Our application supports this concept in thefollowing ways:Overview first. The grid layout algorithm positions top-level nodes on a regulargrid where each grid position can be regarded as a cell in a table. The user startswith examining the completely collapsed graph, i.e., only top-level nodes are vis-ible. The application allows to display a node simply by showing the associatedpathway’s name as caption (see Fig. 2) or by creating an iconized view of thenode’s nested graph.Zoom and Filter. We have implemented semantic zooming to display labels oncea certain threshold is reached. Tool tips add additional information on eachpathway node. If enabled, icons in top-level nodes depicting the nested graphgive a quick hint on the pathway’s size and complexity.Details on Demand. The user can expand selected pathway nodes to explore thedetailed network of chemical reactions. In contrast to the established Table Lensmethod, an arbitrary number of cells (pathways) can be enlarged (multiple foci)and examined in detail (see Fig. 3 and 4). Advanced selection and highlightingtechniques facilitate and support the exploration process: selecting a pathwaynode highlights all objects belonging to that cluster. Selecting a specific reac-tion node highlights all edges to the associated substrate and product nodes.Selecting a compound node highlights all reactions this compound is involvedincluding its connections to adjacent pathways.
5.2 Design of the Graphical User Interface
The GUI of the visualization software basically consists of three components,see Fig. 6 and 7.
The Graph View at the left hand side of the window renders the graph andprovides an interface to interact with or edit the topology of the graph directly.Each graphical object can be individually selected, and applicable properties canbe assigned via a context menu. The integrated graph editing capability allows
8 M. Rohrschneider et al.
Aminosugars
metabolism
ositol phospha
metabolism
Inositol
metabolism
Gly oxy late and
dicarboxy late
metabolism
Butanoate
metabolism
Fructose and
mannose
metabolism
Propanoate
metabolism
Nucleotide
sugars
metabolism
entose phospha
pathway
Citrate cy cle
(TCA cy cle)
Py ruv ate
metabolism
Pentose and
glucuronate
nterconv ersions
Ascorbate and
aldarate
metabolism
Gly cosami
nogly can
degradation
Peptidogly can
biosy nthesis
D-Glutamine
and D-glutamate
metabolism
map01051
Purine
metabolism
Nitrogen
metabolism
Vitamin
B6 metabolism
Glutamate
metabolism
Gly cine, serine
and threonine
metabolism
Histidine
metabolism
Urea cy cle and
metabolism
of amino groups
Alanine and
aspartate
metabolism
Py rimidine
metabolism
Fatty acid
biosy nthesis
Fatty acid
elongation
n mitochondria
Nicotinate and
nicotinamide
metabolism
Gly coly sis /
Gluconeogenesi
Starch and
sucrose
metabolism
C5-Branched
dibasic acid
metabolism
Galactose
metabolism
Carbon f ixation
n photosy ntheti
organisms
Lipopoly sa
ccharide
biosy nthesis
Gly ceropho
spholipid
metabolism
Retinol
metabolism
Sy nthesis and
degradation
f ketone bodie
Valine, leucine
and isoleucine
degradation
map01056
3-Chloroac
ry lic acid
degradation
Pantothenate
and CoA
biosy nthesis
map00523
beta-Alanine
metabolism
Methane
metabolism
Fatty acid
metabolism
Ty rosine
metabolism
Valine, leucine
and isoleucine
biosy nthesis
Ribof lav in
metabolism
Ly sine
biosy nthesis
Fig. 2. The top-level graph of the ”Car-bohydrate Metabolism” (bright) and re-lated pathways (dark). The highlightednodes can be expanded to reveal the de-tailed reaction network.
Fig. 3. Detailed view of the expandednode ”Citrate Cycle (TCA)”. The high-lighted compound node ”Acetyl-CoA”plays a central role within this pathwayand establishes several connections to ad-jacent pathways.
the user to manually construct pathway graphs or to modify a given layouteither generated by the algorithm or loaded from file. Expanding or collapsingindividual nodes can be performed by either double-clicking a node or selectingthe operation via the associated entry in the context menu.
The Data Browser displays the hierarchical structure of the grap (explorerlayout) and grants access to textual or numerical attributes of each graph ele-ment. Generic graph element properties, e.g., edge width, node size and shape,color or transparency, can be manipulated, and the effects will be directly dis-played in the graph view. A simple search function among the textual attributescan be used as a filter to highlight and select a group of graph elements. Thisis an intuitive way to state queries like ”Select all pathways containing the com-pound Pyruvate” (see Fig. 5). Highlighting elements matching a given searchpattern is also propagated to the top-level.
The Algorithm Info Area at the bottom-right hand side displays textual out-put giving feedback on the progress of invoked graph or layout algorithms andto present search results, cf. Fig. 7.
6 Performance Results
Our KEGG import routine is suitable to construct pathway graphs of differentsize and complexity. To implement, test and demonstrate the discussed tech-niques, we constructed two graphs. Images 2 through 5 were created from 17pathway files downloaded from the KEGG database covering the complete car-bohydrate metabolism. Additional non-expandable pathway nodes were createdwhen referenced in one of the input files. A graph with a total of 649 compound
Visualizing Metabolic Networks using Advanced Focus&Context Techniques 9
Aminosugars
metabolism
ositol phospha
metabolism
Inositol
metabolism
Gly oxy late and
dicarboxy late
metabolism
Butanoate
metabolism
Fructose and
mannose
metabolism
Propanoate
metabolism
Nucleotide
sugars
metabolism
D-Gly cera
ldehy de
3-phosphate
alpha-D
-Glucose 6
-phosphate
2-Amino
-2-deoxy -D
-gluconate
D-Glucose
2-Dehy dro-3
-deoxy -D
-gluconate
Py ruv ate
2-Dehy dro-3
-deoxy -6
-phospho-D
D-Ribose
5-Phospho
-alpha-D
-ribose 1
D-Sedohep
tulose 7
-phosphate
D-Gly cera
ldehy de
3-phosphate
beta-D
Fructose 1,6
bisphosphate
D-Ery th
rose 4
-phosphate
6-Phospho-2
-dehy dro-D
-gluconate
Deoxy
ribose2.7.1.15
D-Gly
cerate
D-Glucono-1,
-lactone 6
-phosphate
2-Deoxy
-D-ribose
1-phosphate
beta-D
-Fructose 6
-phosphate
D-Gly cer
aldehy de
D-Xy lulose 5
-phosphate
D-Xy lulose 5
-phosphate
Py ruv ate
metabolism
2.7.4.23
5.3.1.9
4.1.2.-
Succiny l
-CoA
Thiamin
diphosphate
S-succiny l
dihy droli
poy lly sine
2-(a-Hy dr
oxy ethy l)
thiamine
S-acety ld
ihy drolip
oy lly sine
Pentose and
glucuronate
nterconv ersions
N6-(lipo
y l)ly sine
1.3.99.1
-(dihy drolipo
ly sine
Ascorbate and
aldarate
metabolism
Gly cosami
nogly can
degradation
Peptidogly can
biosy nthesis
D-Glutamine
and D-glutamate
metabolism
map01051
Purine
metabolism
Nitrogen
metabolism
Vitamin
B6 metabolism
Glutamate
metabolism
Gly cine, serine
and threonine
metabolism
Histidine
metabolism
Urea cy cle and
metabolism
of amino groups
Alanine and
aspartate
metabolism
Py rimidine
metabolism
Fatty acid
biosy nthesis
Fatty acid
elongation
n mitochondria
Nicotinate and
nicotinamide
metabolism
5.4.2.7
4.1.2.4
1.1.1.44
3.1.3.11
3.1.3.37
2.2.1.1
5.3.1.6
4.1.2.13
3.1.1.17
1.2.7.5
1.1.5.2 1.1.99.3
Ribose 1
-phosp
1.2.7.3
Py ruv ate
1.8.1.4
1.3.5.1
4.2.1.3
4.1.3.6
4.2.1.3
2.3.3.1
1.1.1.376.4.1.1
1.2.4.1
4.1.1.49
Starch and
sucrose
metabolism
1.2.- 4.1.2.1
4.1.3.16
C5-Branched
dibasic acid
metabolism
Galactose
metabolism
Carbon f ixation
n photosy ntheti
organisms
Lipopoly sa
ccharide
biosy nthesis
Gly ceropho
spholipid
metabolism
Retinol
metabolism
Sy nthesis and
degradation
f ketone bodie
Valine, leucine
and isoleucine
degradation
map01056
3-Chloroac
ry lic acid
degradation
Pantothenate
and CoA
biosy nthesis
map00523
beta-Alanine
metabolism
Methane
metabolism
Fatty acid
metabolism
Ty rosine
metabolism
Valine, leucine
and isoleucine
biosy nthesis
Ribof lav in
metabolism
Ly sine
biosy nthesis
Thiamin
diphosphate
2.3.1.61
D-Ribose
1,5-bis
phosphate
D-Glucono
-1,5-lactone
D-Gluco
nic acid
D-Ribose
5-phosphate
6-Phospho
-D-gl
uconate
beta-D
-Glucose
alpha-D
-Ribose
1-phosphate
D-Ribul
ose 5
-phosphate
2-Deoxy
-D-ribose
5-phosphate
beta-D
-Glucose
6-phosphate
2-Dehy dro-D
-gluconate
2-Phospho-D
-gly cerate5.3.1.9 1.1.1.49
1.1.1.435.1.3.1
2.7.6.15.4.2.2
5.4.2.7
2.7.1.11
2.7.1.15
2.2.1.1
2.7.1.134.3.1.9
2.7.1.45
4.2.1.12
3.1.1.31
2.7.1.12
4.2.1.39
1.1.3.4
1.1.3.5
1.1.99.10
2.2.1.2
1.1.1.472.7.1.-
1.1.1.215
4.1.2.9
Oxalosu
ccinate
Fumarate (S)-Malate
Oxaloa
cetate
Acety l-
CoA
Isocitrate
cis-Ac
onitate
Citrate
Succinate
N6-(lipo
y l)ly sine
N6-(dihy drol
ipoy l)ly sine
2-Oxogl
utarate
3-Carboxy -1
-hy droxy p
ropy l-ThPP
1.2.4.2
Phosphoen
olpy ruv ate
1.2.4.2
6.2.1.4
6.2.1.5
6.2.1.4
6.2.1.5
2.3.3.8
1.1.1.42
1.1.1.41
1.1.1.42
4.2.1.2
1.2.4.1
1.8.1.4
2.3.1.12
1.2.7.1
4.1.1.32
D-Gly cera
ldehy de
3-phosphate
beta-D
-Fructose 6
-phosphate
5.3.1.1
5.1.3.3
1.3.13 5.4.2
5.4.2.4
1.2.1.12
1.2.1.59
Salicin
3.2.1.86
2.7.1.69 PTS
-Glc-EIIC,
beta-D
-Glucose
6-phosphate
alpha
-D-Glucose
2-Phospho-D
-gly cerate
Phosphoen
olpy ruv ate
3-Phospho-D
-gly cerate
Ethanol
3-Phospho-D
-gly ceroy l
phosphate
Acety l-CoA
1.2.1.3
Salicin
6-phosphate
beta-D
Fructose 1,6
bisphosphate
alpha-D
-Glucose 6
-phosphate
(S)-Lactate
1.1.1.2
1.2.1.5
1.2.4.1
4.1.1.1
Acetate
3.1.3.11
3.1.3.37
4.2.1.11 2.7.1.403.1.3.9
5.3.1.9
2.7.1.1
2.7.1.2
5.1.3.15
5.3.1.9
3.1.3.10
2.7.1.412.7.1.63
1.3.13 5.4.2
5.4.2.4
D-Glucose
D-Glucose 1
-phosphate
Gly cerone
phosphate
beta-D
-Glucose
Enzy me N6
-(lipoy
l)ly sine
2-(alpha
Hy droxy ethy
l)thiamine
y drolipoy lly s
-residue
ty ltransf eras
Thiamin
diphosphatePy ruv ate
Acetal
dehy de
Enzy me N6
-(dihy droli
poy l)ly sine
Arbutin
6-phosphate
Arbutin
2.7.2.3
Oxaloa
cetate
2,3-Bisp
hospho
-D-gly cerate
4.1.1.1
1.8.1.4
6.2.1.1
1.1.1.1
EUTG
1.2.4.1
4.1.2.132.7.1.11
5.3.1.9 2.3.1.121.1.1.27
1.3.13 5.4.2
5.4.2.4
5.4.2.2
2.7.1.1
2.7.1.2
2.7.1.63
2.7.1.69 PTS
-Arb-EIIC,
2.7.1.69
PTS-Asc
4.1.1.32
3.2.1.86
1.1.99.8
4.1.1.49
1.2.7.5
1.2.7.6
1.2.1.9
1.2.7.1
Fig. 4. Three expanded pathway nodes:”Citrate Cycle (TCA)”, ”Pentose Phos-phate Pathway”, and ”Glycolysis / Glu-coneogenesis”; the latter being high-lighted in red with the connections to ad-jacent pathway nodes.
D-Gly cera
ldehy de
3-phosphate
beta-D
-Fructose 6
-phosphate
5.3.1.1
5.1.3.3
N-Acety l
-alpha-D
-glucosamine
1.3.13 5.4.2
5.4.2.4
1.2.1.12
1.2.1.59
Salicin
3.2.1.86
2.7.1.69 PTS
-Glc-EIIC,
beta-D
-Glucose
6-phosphate
alpha
-D-Glucose
2-Phospho-D
-gly cerate
Phosphoen
olpy ruv ate
3-Phospho-D
-gly cerate
Ethanol
3-Phospho-D
-gly ceroy l
phosphate
Acety l-CoA
1.2.1.3
Salicin
6-phosphate
beta-D
Fructose 1,6
bisphosphate
alpha-D
-Glucose 6
-phosphate
(S)-Lactate
N-Palmitoy l
gly coprotein
N-Acety ln
euraminate
N-Acety l-D
-manno
samine
UDP-N-acety
-3-(1-carb
oxy v iny l)
CMP-N
-gly coloy ln
euraminate
UDP-N-acety
-D-gal
actosamine
UDP-N
-acety l-D
-mannosami
UDP-N
-acety l
muramate
Chitin
alpha-D
Glucosamine
-phosphate
2.7.1.1
2.7.1.2
N-Acety l-D
-gluco
samine
Ferrocy to
chrome b5
UDP-N-acety
-D-gal
actosamin
Chitosan
N-Acety l
-D-man
nosamine
N-Gly co
loy l-ne
uraminate
CMP-N
-acety lne
uraminate
1.1.1.-
2.5.1.56
N-Acety lmu
ramic acid 6
-phosphate
5.1.3.14
5.4.2.3
2.6.1.16
-Phosphatidy
-1D-my o
-inositol 5
4.2.-.-
2.7.1.68
D-my o-Inosito
1,4,5
trisphosphate
-Phosphatidy
-1D-my o
-inositol 3
my o-Inos
itol 4
-phosphate
D-my o-Inosito
1,3
bisphosphate
1D-my o
-Inositol 1,4
bisphosphate
1D-my o
-Inositol 3
-phosphate
1,2-Diacy l
-sn-gly cerol
my o-Inositol
2-Deoxy -5
-keto-D
luconic acid
2.1.1.40
3-O-Met
hy l-my o
-inositol
2,4,6/3,5
-Pentah
y droxy cy
Acety l-CoA
D-Gly
cerate
my o-Inositol
5.3.1.1
iolB
Oxaloa
cetate
meso-Tart
aric acid
2-Hy droxy
-3-oxos
uccinate
D-Ribulose
1,5-bis
phosphate
2-Hy droxy
-3-oxop
ropanoate
2-Phospho
gly colate
Dihy drox
y f umarate
5,10-Meth
eny ltetra
hy drof olate
Hy droxy
py ruv ate
10-Formy
ltetrahy
drof olate
5,10-Methy
lenetetrah
y drof olate
1.1.1.79
Formy l
-CoA
2-Oxogl
utarate
N-Formy ld
eriv ativ es
1.5.1.5
1.1.1.79
6.2.1.8
beta-D
-Fructose
CDP-3,6
-dideoxy -D
-mannose
UDP-D
-xy lose
alpha-D
-Glucose 6
-phosphate
3-Keto
sucrose
D-Gluc
uronate
Sucrose
beta-D
-Fructose 6
-phosphate
alpha-D
-Glucose 6
-phosphate
alpha,alpha'
-Trehalose 6
-phosphate
beta-D
-Glucose
6-phosphate
D-Glucose
alpha
-D-Glucose
GDP-g
lucose
ADP-g
lucose
Amy lose
Sucrose
D-Glu
coside
alpha,alpha
-Trehalose
3.2.1.21
(2,6-beta
D-Fructosy l)
Cellot
etraose
3.2.1.21
3.2.1.122
3.2.1.93
2.4.1.13
3.2.1.37
1.17.1.1
2.7.1.69
PTS-Mal
2.4.1.1
3.2.1.10
5.4.99.16
2-Aceto
lactate
2-(alpha
Hy droxy ethy
l)thiamine
(S)-3
-Hy droxy but
anoy l-CoA
(S)-Acetoin
Butanoic
acid
2-Oxogl
utarate
2-Hy droxy
glutarate
2-Hy droxy g
lutary l-CoA
(R)-3-((R)-3
-Hy droxy
butanoy lox
4-Hy droxy bu
tanoic acid
L-Glu
tamate
Viny lac
ety l-CoA
2.3.1.19 Succinate
semialdehy de
4-Aminob
utanoate
6.2.1.2
D-Gly cera
ldehy de
3-phosphate
Mannitol
beta-D
-Fructose 2
-phosphate
beta-D
Fructose 2,6
bisphosphate
D-Mannose
Sorbitol
6-phosphate
GDP-L
-f ucose
L-Rha
mnose
L-Rhamno
f uranose
Gly cerone
phosphate
GDP-m
annose
L-Rham
nonate
GDP-D
mannuronate
L-Fucose
1-phosphate
GDP-4
-dehy dro-6
-deoxy -D
ADP-m
annose
beta-D
Fructose 1,6
bisphosphate
gouronide wit
deoxy -alpha-
ery thro-hex-4
GDP-6
-deoxy -D
-mannose
D-Sorbitol
D-Fructose
Lev an
D-Fuconate
L-Fucul
ose 1
-phosphate
L-Fuculose
1.1.99.21
L-Sorbose
3.1.3.54
1.1.1.138
2.7.1.11
3.1.3.46
(R)-Methy l
malony l-CoA
1.1.1.132
5.3.1.8
1.1.2.2
Propanoate
2-Oxobu
tanoate
L-Valine
beta
-Alanine
Malon
y l-CoA
3-Oxoprop
iony l-CoA
2-Propy n
-1-al
6.2.1.1
6.2.1.17
(S)-Methy l
malonate
semialdehy de
4.2.1.79
2.3.1.86.2.1.13
2.8.3.3
UDP-a
piose
Pectin
1-Phospho
-alpha-D
galacturonate
dTDP-6
-deoxy
-L-mannose
UDP-D
-galactose
1,4-beta-D
-Xy lan
dTDP-ga
lactose
dTDP-4
-dehy dro
-6-deoxy
UDP-D
-xy lose
dTDP-4
acetamido-4,
-dideoxy
D-Xy lose
UDP-L
-arabinose
GDP-6
-deoxy -L
-mannose
D-Galac
turonate
GDP-4
-dehy dro-6
-deoxy -L
2.7.7.32
D-Gly cera
ldehy de
3-phosphate
alpha-D
-Glucose 6
-phosphate
up:O69208
2.6.1.33
2-Amino
-2-deoxy -D
-gluconate
D-Glucose
2-Dehy dro-3
-deoxy -D
-gluconate
2-Dehy dro-3
-deoxy -6
-phospho-D
D-Ribose
5-Phospho
-alpha-D
-ribose 1
D-Sedohep
tulose 7
-phosphate
D-Gly cera
ldehy de
3-phosphate
beta-D
Fructose 1,6
bisphosphate
D-Ery th
rose 4
-phosphate
6-Phospho-2
-dehy dro-D
-gluconate
Deoxy
ribose 2.7.1.15
D-Gly
cerate
D-Glucono-1,
-lactone 6
-phosphate
2-Deoxy
-D-ribose
1-phosphate
beta-D
-Fructose 6
-phosphate
D-Gly cer
aldehy de
D-Xy lulose 5
-phosphate
D-Xy lulose 5
-phosphate
2.7.4.23
5.3.1.9
4.1.2.-
Succiny l
-CoA
Thiamin
diphosphate
S-succiny l
dihy droli
poy lly sine
2-(a-Hy dr
oxy ethy l)
thiamine
S-acety ld
ihy drolip
oy lly sine
N6-(lipo
y l)ly sine
Enzy me N6
-(lipoy
l)ly sine
1.3.99.1
-(dihy drolipo
ly sine
Acety lenedi
carboxy late
Lactal
dehy de
y drolipoy lly s
-residue
ty ltransf eras
D-Man
nonate
2-Hy droxy e
thy lenedic
arboxy late
Phosphoen
olpy ruv ate
Acety l
adeny late
(3S)-Citr
amaly l-CoA
2-Propy
lmalate
(2S)-2-Isop
ropy lmalate
(R)-2
-Ethy lmalate
(R)-Lact
aldehy de
(S)-Lact
aldehy de
Methy l
gly oxal
2.3.1.9
4.1.1.32
1.2.4.14.1.1.78
D-Fruct
ose 6
-phosphate
D-arabino-3
-Hexulose 6
-phosphate
D-Galac
turonate
D-Gluc
uronate
CDP-ribitol
D-Xy lonate
(4S)-4,6
-Dihy droxy
-2,5-diox
Xy litol
D-Xy lulose
D-Ly xose
L-Ly xose
L-Ara
binose
Pectin
Mesaco
ny l-CoA
Pectate
Gly cola
ldehy de
3.2.1.15
5.1.3.1
4.1.1.34
2.4.1.17
(3S)-Citr
amaly l-CoA
Itacony l
-CoA
Acetate
Propan
oy l-CoA
D-threo-3
Methy lmalate
Methy loxa
loacetate
2-Oxogl
utarate
2-Oxobu
tanoate
D-ery thro
-3-Meth
y lmalate
4-Methy
lene-L
-glutamate
2-Methy
lmaleate
4-Methy
lene-L
-glutamine
my o-Inositol
4-Methy l
-L-glutamate
2.8.3.11
4.1.3.22
2.3.3.112.8.3.7
alpha
-D-Glucose
D-Gal
actose
UDP-g
lucose
2-Dehy dro-3
-deoxy -D
-galactonate
Gly cerol
D-Galac
tose 6
-phosphate
alpha-D
Galactosy l-(
->3)-1D-my o
D-Tagat
ose 6
-phosphate
D-Glucose
Gly cerone
phosphate
D-Tagatose
1,6-bis
phosphate
D-Galac
tosamine
N-Acety l-D
-galacto
samine 6
Galactitol
3.2.1.22
1.1.1.21
4.1.2.403.2.1.22
UDP-glu
curonate
1.1.3.9
D-Glu
carate
L-Gal
actose
L-Gulose
GDP-L
-gulose
L-Ribulose 5
-phosphate
L-Xy lulose 5
-phosphate
L-Gulonate
D-Gluc
uronate
3-Dehy dro-L
-gulonate 6
-phosphate
D-Galac
turonate
D-Gala
ctarate
4R,5S)-4,5,6
-Trihy droxy
-2,3
2-Oxogl
utarate
Monodehy dr
oascorbate
5-Hy droxy
-2,4-diox
opentanoate
L-Xy lonate
L-Arab
inonate
Gly cosami
nogly can
degradation
L-Arabi
nono-1,4
-lactone
L-Galac
tono-1,4
-lactone
5.1.3.-
5.1.3.18
2.7.1.69 PTS
-Ula-EIIC,
3-Dehy dro-L
-gulonate
5.1.3.-
5.1.3.18
1.13.99.1
Peptidogly can
biosy nthesis
D-Glutamine
and D-glutamate
metabolism
1.1.1.2
map01051
1.2.1.5
Purine
metabolism
Nitrogen
metabolism
Vitamin
B6 metabolism
Glutamate
metabolism
Gly cine, serine
and threonine
metabolism
Histidine
metabolism
Urea cy cle and
metabolism
of amino groups
Alanine and
aspartate
metabolism
Py rimidine
metabolism
Fatty acid
biosy nthesis
Fatty acid
elongation
n mitochondria
Nicotinate and
nicotinamide
metabolism
1.2.4.1
4.1.1.1
Acetate
-Phosphatidy
-D-my o
-inositol
UDP-D-gal
acturonate
Undecapreny
phosp
hate alpha
3.1.3.11
3.1.3.37
4.2.1.11 2.7.1.403.1.3.9
5.3.1.9
2.7.1.1
2.7.1.2
5.1.3.15
5.3.1.9
3.1.3.10
2.7.1.412.7.1.63
1.3.13 5.4.2
5.4.2.4
3.5.99.6
2.3.1.157
2.7.7.23
5.1.3.7
1.14.18.2
5.4.2.10
2.3.1.3 3.2.1.-
3.5.1.41
2.7.1.64
3.1.3.26
2.7.1.69
PTS-Mur
2.7.1.134
2.2.1.5
(R,R)-Tar
taric acid
4.2.1.44
3.5.1.49
trans-2,3
-Epoxy s
uccinate
3.5.4.9
6.3.4.3
3.5.1.31
3.5.1.8
3.5.1.15
2.7.2.6
1.1.99.14
3.1.2.10
1.12.1.2
1.2.1.2
1.2.2.1
1.2.1.21
4.1.1.40
3.2.1.21
4.1.1.54
4.1.3.24
6.2.1.9
1.1.1.37 4.2.1.32 1.1.1.93
2.7.1.31
5.1.3.10
4.1.1.39
3.2.1.39
2.4.1.12 2.4.1.35
2.7.7.34 3.2.1.31
2.4.1.172.4.1.15
3.2.1.58
3.1.3.9
2.7.1.1
2.7.1.2
5.3.1.9
2.4.1.5
3.2.1.54
3.2.1.28
2.4.1.64
3.2.1.122
3.2.1.2
3.2.1.3
3.2.1.15
3.2.1.67
2.4.1.25
2.7.1.69
PTS-Tre 2.4.1.21
3.6.1.21
2.4.1.20
4.2.1.55
Thiamin
diphosphate
1.1.1.30
2.8.3.12 1.2.7.1
2.7.2.7
2.3.1.54
2.8.3.8
5.2.1.1
1.1.1.83
1.2.4.1
2.2.1.6 1.3.99.1
2.7.1.69 PTS
-Man-EIIC,
4.1.1.5
1.1.1.4
1.1.1.14
2.7.1.3
3.2.1.80
1.1.1.172.4.1.-
1.1.1.135
4.2.1.47
3.2.1.77
3.2.1.137
2.7.1.90
2.7.1.5
5.3.1.1
5.3.1.25
4.1.2.-
1.1.1.173
4.1.2.18
3.6.1.21
6.2.1.4
6.2.1.5
2.3.3.5
1.3.99.3 4.2.1.54
3.1.2.4
6.2.1.-
2.3.1.9
4.3.1.6
1.1.1.-
1.1.1.
- 2.1.2.-
1.3.1.-
2.7.7.12
5.4.2.7
4.1.2.4
1.1.1.44
3.1.3.11
3.1.3.37
2.2.1.1
5.3.1.6
4.1.2.13
3.1.1.17
1.2.7.5
1.1.5.2 1.1.99.3
Ribose 1
-phosp
1.2.7.3
1.8.1.4
1.2.1.22
4.2.3.3
1.3.5.1
4.2.1.3
4.1.3.6
4.2.1.3
2.3.3.1
1.1.1.376.4.1.1
1.2.4.1
4.1.1.49
3.1.2.1
2.7.1.40
1.1.1.38
1.1.39 1.1.1.
1.1.1.82
1.1.1.21
1.1.1.77
1.2.1.49
4.1.1.78
1.1.2.4
1.1.1.37
1.1.99.7
1.1.2.3 1.2.1.3
2.8.3.1
6.4.1.2
ACCB, bccP
2.3.3.14
1.1.1.13
1.2.4.1
5.-.-.-
4.1.1.85
4.1.2.36
3.6.1.7
1.13.12.4
3.2.1.315.3.1.4
1.1.1.56
5.3.1.17
2.7.1.16
2.7.1.47
5.3.1.15
2.7.1.53
5.3.1.5
5.3.1.-
1.1.1.12
3.1.1.681.1.1.21 1.1.1.121
1.1.1.175
1.1.1.22
1.1.1.57
4.3.1.2
4.2.1.84.2.1.7
5.3.1.12
1.1.1.58
1.1.2.2
4.2.2.2
4.2.2.6
4.1.2.19
4.1.1.5
2.7.1.11
2.7.1.144
3.2.1.26
3.2.1.108
3.2.1.22
2.4.1.67
3.2.1.26 3.2.1.22
2.7.1.69
PTS-Lac
5.4.2.2
3.2.1.223.2.1.22
3.2.1.22
2.7.1.1
2.7.1.2
2.7.1.58
2.7.7.9
3.1.1.25
5.1.3.2
5.3.1.26
1.1.1.16
3.1.3.9
1.1.1.122
2.7.1.433.1.1.17
1.1.1.130 1.6.5.4
4.1.2.20
1.2.1.3
4.2.1.41
1.10.3.3
CO2
3.2.1.33
(R)-3-Hy dro
xy butanoate
1.2.- 4.1.2.1
4.1.3.16
C00687
Carbon f ixation
n photosy ntheti
organisms
Lipopoly sa
ccharide
biosy nthesis
Gly ceropho
spholipid
metabolism
Retinol
metabolism
Sy nthesis and
degradation
f ketone bodie
Valine, leucine
and isoleucine
degradation
map01056
3-Chloroac
ry lic acid
degradation
Pantothenate
and CoA
biosy nthesis
map00523
beta-Alanine
metabolism
Methane
metabolism
Fatty acid
metabolism
Ty rosine
metabolism
Valine, leucine
and isoleucine
biosy nthesis
Ribof lav in
metabolism
Ly sine
biosy nthesis
D-Glucose
D-Glucose 1
-phosphate
Gly cerone
phosphate
beta-D
-Glucose
Enzy me N6
-(lipoy
l)ly sine
2-(alpha
Hy droxy ethy
l)thiamine
y drolipoy lly s
-residue
ty ltransf eras
Thiamin
diphosphate
Acetal
dehy de
Enzy me N6
-(dihy droli
poy l)ly sine
Arbutin
6-phosphate
Arbutin
2.7.2.3
Oxaloa
cetate
2,3-Bisp
hospho
-D-gly cerate
4.1.1.1
1.8.1.4
6.2.1.1
1.1.1.1
EUTG
1.2.4.1
4.1.2.132.7.1.11
5.3.1.9 2.3.1.121.1.1.27
1.3.13 5.4.2
5.4.2.4
5.4.2.2
2.7.1.1
2.7.1.2
1D-my o
nositol 1,3,4
-tetrak
2.7.1.63
2.7.1.69 PTS
-Arb-EIIC,
2.7.1.69
PTS-Asc
4.1.1.32
3.2.1.86
1.1.99.8
4.1.1.49
1.2.7.5
1.2.7.6
1.2.1.9
1.2.7.1
3.1.3.57
Ferricy to
chrome b5
N-Acety l
-D-glu
cosamine
Chitin
D-Gluc
osamine
Chitobiose
D-Glucos
amine 6
-phosphate
UDP-N
-acety l-D
-glucosamine
N-Acety lne
uraminate 9
-phosphate
UDP-N
-acety l-D
mannosamin
Mucopoly s
accharide
2-Amino
-2-deoxy -D
-gluconate
D-Gluco
saminide
Gly colipid
UDP-N-acety
-2-amino-2
-deoxy
ominic acid(
-reducing N
- or O-acy lne
Gly cop
rotein
3.2.1.52
D-Fruct
ose 6
-phosphate
1.6.2.2
N-Acety l
muramate
1.1.1.158
2.5.1.71.1.1.136
2.4.1.165.1.3.14
1.1.1.-
3.1.4.-
5.1.3.144.1.3.32.7.7.43
1.14.18.2
3.1.3.29
2.7.7.43
2.5.1.56
2.5.1.57
3.1.3.-
2.7.1.60
5.1.3.9
3.5.1.25
2.3.1.4
5.1.3.8
2.7.1.59
3.2.1.14
2.7.1.69
PTS-Dgl
3.2.1.14
2.3.1.96
3.2.1.132
2.7.1.69
PTS-Nag
2.3.1.157
2.7.7.23
3.5.1.33
1.1.3.-
1D-my o
nositol 1,4,5
-tetrak
4-Hy droxy
-2-oxog
lutarate
D-my o-Inosit
1,3,4
trisphosphate
my o-Inositol
hexak
isphosphate
1D-my o
nositol 3,4,5
-tetrak
D-my o-Inosito
1,2,4,5,6
-penta
1D-my o
-Inositol 1,3,
4,5,6-pent
1D-my o
nositol 1,3,4
-tetrak
Inositol
,2,3,5,6-pent
kisphosphate
D-my o-Inosito
3,4
bisphosphate
D-Glucose 6
-phosphate
1-Phospha
tidy l-D-my o
-inositol
Phosphatid
y linositol
-3,4,5
Inositol
1-phosphate
D-Gluc
uronate
-Phosphatidy
-1D-my o
-inositol 4
1-O-Met
hy l-my o
-inositol
3.1.3.36
2.7.1.149
2.7.1.153
3D-(3,5/4)
-Trihy drox
cy clohexane
2.7.1.151
3.1.3.67
2.7.1.151
3.1.3.56
3.1.3.62
2.4.1.11
2.7.1.127
2.7.1.134
3.1.3.8
2.7.1.140
2.7.1.134
3.1.3.56
3.1.3.57
3.1.3.64
3.1.3.66
3.1.3.25
5.5.1.4
1.13.99.1
3.1.3.25
3.1.4.3
4.6.1.13
2.7.1.67
2.7.1.137
3.1.4.11
PLCE PLCZ
3.1.3.64
3.1.3.25
2.1.1.39
2.7.1.158
1.1.1.18
D-2,3-Diketo
4-deoxy -epi
-inositol
3-Oxopr
opanoate
D-Gly cera
ldehy de
3-phosphate
2-Deoxy -5
-keto-D
gluconic acid
Gly cerone
phosphate
5-Deoxy
glucur
onic acid
2,4,6/3,5
-Pentah
y droxy cy
2.7.1.92
1.2.1.18
1.2.1.27
4.1.2.29
1.-.-.-
Citrate
(3S)-3
-Carboxy -3
-hy droxy pro
iolE1.1.1.18
(S)-Malate
Acety l-CoA
3-Phospho-D
-gly cerate Gly colate
Gly cola
ldehy de
Ethy lene
gly col
Gly oxy late
Formate
Oxalate
H+
Hy drogen
Formy l
phosphateOxaly l-CoA
cis-Ac
onitate
Isocitrate
Succinate
3-Oxal
omalate
3-Ethy
lmalate
2-Hy droxy
3-oxoadipate
Butanoy l
-CoA
Pentan
oy l-CoA
3-Propy
lmalate
4.2.1.3
Oxaloa
cetate
5.3.1.22
2.3.3.12
2.3.3.7
3.5.1.56
4.1.3.134.1.3.1
4.1.3.16
3.5.1.68
3.5.1.27
3.5.1.9
3.5.1.10
4.2.1.3
1.2.1.17
4.1.1.8
1.1.3.15
glcE glcF
2.8.3.2
1.1.1.26
1.1.1.29
1.2.3.5
4.1.1.2
1.12.7.2
1.2.3.4
1.1.1.77
2.3.3.1
1.1.1.26
1.1.1.29
1.1.1.60
4.1.1.47
2.3.3.9
5.1.2.5 3.3.2.41.3.1.7
1.1.1.92
3.1.3.18
1.1.1.93
D-Glucose 6
-phosphate Pectin
D-Xy lose
1,4-beta-D
-Xy lan
Pectate
D-Galac
turonate
Sucrose
6-phosphate
Cellulose UDP-glu
curonate
beta-D
Glucuronosid
beta-D
-Glucose
D-Glucose
UDP-g
lucose
alpha
-D-Glucose
CDP-3,6
-dideoxy -D
-glucose
alpha-D
-Glucose 1,6
bisphosphate
D-Glucose 1
-phosphate
CDP-g
lucose
1,3-beta-D
-Glucan
Maltose
Maltod
extrin
Dextrin
Cy clomal
todextrin
alpha,alpha
-Trehalose
beta-D
-Glucose
1-phosphate
Maltose
6'-phosphate
CDP-4
-dehy dro-3,6
-dideoxy
DP-4-dehy d
-6-deoxy
-D-glucose
SucroseCellobiose
Starch
alpha
-D-Glucose
D-Fructose
D-Fructose
Cellobiose
Cellulose
(2,6-beta
D-Fructosy l)
Cellotriose
Cellop
entaose
Celloh
exaose
Celloh
eptaose
Sucrose
-6-phosphate
Maltose
Isomaltose
.7.1.1 2.7.1.
2.7.1.4
3.2.1.26
1.1.99.13
3.2.1.20
3.2.1.26
3.2.1.4
3.2.1.91
2.4.2.24
3.1.3.24
2.4.1.29
4.1.1.35
2.7.1.69
PTS-Scr
4.2.1.45
2.7.1.41
3.6.1.9
2.4.1.342.4.1.14
1.1.1.22
5.1.3.6
2.7.7.33
2.7.1.106
2.7.1.10
2.4.1.432.7.7.9
3.1.1.11Thiamin
diphosphate
2.7.7.27
5.4.2.2
2.3.1.61
2.4.1.18
3.1.3.12
5.4.2.6
3.2.1.1
3.2.1.20 2.4.1.8
2.4.1.4 2.4.1.10
2.4.1.10
3.2.1.74
3.2.1.65
3.2.1.74
3.2.1.74
3.2.1.74
3.2.1.74
3.2.1.74
2.4.1.7
3.6.1.21
3.2.1.10
3.2.1.3
Maleic acid
(S,S)-Butane
-2,3-diol
Succinate
(R)-Acetoin
Fumarate
(R,R)-Butane
-2,3-diol
(R)-Malate
Butanoy l
phosphate
Poly -beta
-hy drox
y buty rate
Butanal
Butanoy l
-CoA
Acetoac
ety l-CoA
Croton
oy l-CoA
Glutacony l
-1-CoA
Acety l-CoA
(S)-3
-Hy droxy -3
-methy lgl
Acetoa
cetate
(R)-3
-Hy droxy but
anoy l-CoA
1-Butanol
3-Buty noate
3-Buty n
-1-al
Diacety l
3-Buty n
-1-ol
1.2.99.3
1.1.99.8
1.2.1.3
4.2.1.27
3.1.1.22
2.8.3.5
6.2.1.16
1.1.1.157
1.1.1.61
1.1.1.36 4.1.3.4
4.2.1.17
2.3.3.10
1.1.1.35
2.3.1.9
5.1.2.3
3.1.1.75
2.3.1.-
5.3.3.3
2.6.1.194.1.1.15
1.3.1.44
1.3.99.2
1.2.1.10 1.2.1.16
1.2.1.24
1.1.1.-
4.2.1.31
2.2.1.6
5.1.2.4
1.1.99.2
4.1.1.70
3.1.2.11
1.1.1.76
4.2.1.-
1.1.1.5
D-Mannose
6-ph
osphate
L-Sorbose
alpha
-D-Glucose
D-Fruct
ose 1
-phosphate
D-Manni
tol 1
-phosphate
1,4-beta-D
-Mannan
Manno
gly can
D-Sorbitol
D-Gly cer
aldehy de
(Alginate)n
L-Rham
nulose
L-Rhamnu
lose 1
-phosphate
2-Dehy dro-3
-deoxy -L
-rhamnonate
(S)-Lact
aldehy de
beta-D
-Fructose 6
-phosphate
GDP-6-deoxy
-D-talose
D-Mannose
1-ph
osphate
L-Rhamn
ono-1,4
-lactone
6-Deoxy
-L-galactose
2-Dehy dro-3
-deoxy -D
-f uconate
(R)-Lact
aldehy de
Sorbose
1-phosphate
1.1.1.21
2.7.1.69
PTS-Fru
1.1.1.187
5.3.1.5
1.1.1.11
1.1.1.67 3.1.3.22
3.1.3.-
5.3.1.7
2.7.1.69
PTS-Mtl
3.2.1.78
2.7.1.1
2.7.1.2
Propy noate
.7.1.1 2.7.1.
2.7.1.4
2.7.1.1052.7.7.22
1.1.1.271
5.4.2.8
2.7.7.132.7.7.30
1.1.1.140
2.7.1.69
PTS-Gut
4.2.2.3
2.4.1.33
2.7.1.56
5.3.1.14
4.1.2.13 2.7.1.28
4.1.2.19
4.1.2.13
3.1.1.654.2.1.90
3.1.3.11
3.1.3.37
2.4.1.-
4.1.2.17
2.7.1.51 4.2.1.67
3.6.1.21
1.1.1.- 2.7.1.69 PTS
-Sor-EIIC,2.7.1.52
(S)-2-Met
hy lmalate
(Z)-But-2-ene
-1,2,3
tricarboxy lat
Propanoy l
phosphate
1-Aminocy cl
opropane-1
-carboxy late
Propan
oy l-CoA
Propen
oy l-CoA
(S)-Lactate
Lacto
y l-CoA
(S)-Methy l
malony l-CoA
Methy lm
alonate
Succinate
Succiny l
-CoA
(2S,3R)-3
y droxy butan
-1,2,3
Propan-2-ol
beta-Al
any l-CoA
2-Methy
lcitrate
3-Hy dro
xy propi
ony l-CoA
Acetone
Acetoa
cetate
3-Hy droxy
propanoate
Acetoac
ety l-CoA
Acety l-CoA
2-Hy droxy bu
tanoic acid
3-Oxopr
opanoate
4.1.1.41
2-Propy n
-1-ol
Propinol
adeny late
2.3.1.54
2.7.2.1
2.7.2.15
1.2.1.27
5.1.99.1
4.2.1.99
1.2.1.18
6.4.1.3
3.1.2.17
2.1.3.1
6.2.1.1
6.2.1.17
2.8.3.1
6.2.1.4
6.2.1.5
5.4.99.2
4.1.3.30
2.8.3.1
1.2.7.1
2.6.1.192.6.1.18
1.2.1.3 1.2.99.3
1.1.1.59
1.1.99.8 1.2.1.18
2.1.3.1
4.2.1.27
1.1.1.80
2.8.3.8
4.1.1.4
4.1.1.9 6.4.1.2
ACCB, bccP
3.5.99.7
1.1.1.27
1.2.1.18
4.2.1.17
UDP-L
-Ara4FN
UDP-L
-Ara4N
UDP-L
-Ara4O
UDP-D
-galactose
UDP-g
lucose
UDP-glu
curonate
UDP-D-gal
acturonate
UDP-6-sulf
oquinov ose
UDP-L
-iduronate
D-Glucose 1
-phosphate
alpha-D
-Galactose 1
-phosphate
TDP-6-deoxy
-L-talose
dTDP
-glucose
UDP-L
-rhamnose
alpha-D
-Xy lose
1-phosphate
dTDP-D
galacturonate
DP-4-dehy d
-6-deoxy
-D-glucose
L-Ara
binose
beta-L
-Arabinose
1-phosphate
alpha-L
-ArabinanPentosan
dTDP-D
-glucuronate
1.1.1.
- 2.1.2.-
up:Q7VUF1
2.6.1.-2.7.8.-
1.1.1.186
3.13.1.1
5.1.3.2 4.2.1.46
1.1.1.134
1.1.1.133
5.1.3.13
5.1.3.2
1.1.1.22
4.1.1.35
2.7.7.24
2.7.7.33
5.1.3.12 2.7.7.9
4.2.1.76
1.1.1.-
5.1.3.6
2.7.7.11
4.1.1.67
5.1.3.5
2.7.7.10
2.7.1.44
3.2.1.-
D-Ribose
1,5-bis
phosphate
2.7.7.373.2.1.37
2.4.1.43
3.2.1.-
1.1.1.1335.1.3.13
2.7.1.46
2.4.2.24
D-Glucono
-1,5-lactone
3.2.1.55
D-Gluco
nic acid
D-Ribose
5-phosphate
6-Phospho
-D-gl
uconate
beta-D
-Glucose
alpha-D
-Ribose
1-phosphate
D-Ribul
ose 5
-phosphate
2-Deoxy
-D-ribose
5-phosphate
beta-D
-Glucose
6-phosphate
2-Dehy dro-D
-gluconate
2-Phospho-D
-gly cerate5.3.1.9 1.1.1.49
1.1.1.435.1.3.1
2.7.6.15.4.2.2
5.4.2.7
2.7.1.11
2.7.1.15
2.2.1.1
2.7.1.134.3.1.9
2.7.1.45
4.2.1.12
3.1.1.31
2.7.1.12
4.2.1.39
1.1.3.4
1.1.3.5
1.1.99.10
2.2.1.2
1.1.1.472.7.1.-
1.1.1.215
4.1.2.9
Oxalosu
ccinate
Fumarate (S)-Malate
Oxaloa
cetate
Acety l-
CoA
Isocitrate
cis-Ac
onitate
Citrate
Succinate
N6-(lipo
y l)ly sine
4.1.1.-
N6-(dihy drol
ipoy l)ly sine
2-Oxogl
utarate
3-Carboxy -1
-hy droxy p
ropy l-ThPP
1.2.4.2
Phosphoen
olpy ruv ate
1.2.4.2
6.2.1.4
6.2.1.5
6.2.1.4
6.2.1.5
2.3.3.8
1.1.1.42
1.1.1.41
1.1.1.42
4.2.1.2
1.2.4.1
1.8.1.4
2.3.1.12
1.2.7.1
4.1.1.32
Enzy me N6
-(dihy droli
poy l)ly sine
2-(alpha
Hy droxy ethy
l)thiamine
Propane
-1,2-diol
Gly cerone
phosphate
Formate
Oxaloa
cetate (S)-Malate
Acety l
phosphate
Malon
y l-CoA
(R)-2-Hy dr
oxy butane
-1,2,4-tric
Acetoac
ety l-CoA
Acety l-CoA
Acetal
dehy deAcetate
(R)-S
-Lactoy lglu
tathione
(S)-Lactate
(R)-Lactate
1.2.7.1
Thiamin
diphosphate
4.1.1.31
4.1.1.38
4.1.1.49
2.7.9.1
2.7.9.21.2.1.23
3.1.2.6
1.1.1.21
6.4.1.1
4.4.1.5
1.2.1.22
1.1.1.28
1.1.1.79
5.1.2.1
4.1.1.3
1.2.3.6
2.3.1.54
1.1.99.16
2.3.3.9
1.2.2.21.1.1.27
1.2.3.3
2.3.1.8
6.2.1.1
6.2.1.16.2.1.13 2.7.2.12
1.2.99.3
1.2.99.6
4.1.3.22
4.1.3.25
1.2.1.10
EUTE
2.3.3.6
2.3.3.13 4.1.3.-
2.3.1.12
1.8.1.4
3-Dehy dro-L
-gulonate 6
-phosphate
1.2.1.49
1.1.2.5
beta-D
Glucuronosid
1.1.3.3
2.7.2.1
2-Dehy dro-3
-deoxy -D
-gluconate
D-Fruct
uronate
D-Altronate
D-Glucose 1
-phosphate
D-Tagat
uronate
D-Glucur
onate 1
-phosphate
2-Dehy dro-3
-deoxy -6
-phospho-D
L-Gulonate
UDP-g
lucose
UDP-glu
curonate
L-Xy lulose
D-Xy lon
olactone
D-Gly cera
ldehy de
3-phosphate
D-Xy lose
2-Dehy dro-3
-deoxy -D
-xy lonate
D-Ribitol 5
-phosphate
2.7.1.16
D-Ribulose
Ribitol
D-Xy lulose 5
-phosphate
D-Ribul
ose 5
-phosphate
L-Xy lulose 5
-phosphate
L-Ribulose
L-Arabitol
D-Arabitol
3-Dehy dro-L
-gulonate
L-Ribulose 5
-phosphate
4-Deoxy -alp
-D-gluc-4
-enuronosy l)
Digalac
turonate
5-Dehy dro-4
-deoxy -D
-glucuronate
4R,5S)-4,5,6
-Trihy droxy
-2,3
L-Xy lonate
Gly cerone
phosphate
L-Ly xonate
L-Xy lulose 1
-phosphate
4.1.2.-
2.7.1.53
5.1.3.4
5.-.-.-
1.1.1.137
2.7.7.40
1.1.1.11
2.7.1.17
4.1.2.284.2.1.82
1.1.1.10 1.1.1.21
1.1.1.9
2.7.7.9
1.1.1.45
4.1.2.14
4.1.3.16
2.7.1.43
1.1.1.19
2.7.1.45
2.7.7.44
5.3.1.12
1.1.1.15
3.2.1.67
1.1.1.130
3.2.1.15
3.1.1.11
(R)-Acetoin
1.1.1.125
2.7.1.5
(S)-2-Ace
tolactate
Itaconate
cis-Ac
onitate
(S)-2-Met
hy lmalate
Gly oxy late
Mesac
onate
2-Hy droxy
glutarate
L-ery thro-3
-Methy l
maly l-CoA
L-threo-3
Methy lmalate
L-Glu
tamate
L-threo-3
-Methy l
aspartate
Parapy
ruv ate
(R)-2-Met
hy lmalate
4-Hy droxy -4
-methy l
glutamate
Propanoy l
phosphate
4-Methy lene
-2
-oxoglutarate
6.3.1.73.5.1.67
4.2.1.35
4.1.3.22
4.1.3.25
4.1.3.17
4.2.1.34
4.1.1.6
5.4.99.1
6.2.1.5 2.8.3.7
2.2.1.6
4.2.1.56
4.2.1.-
Galactitol 1
-phosphate
D-Galac
tono-1,4
-lactone
D-Fructose
D-Glucose 1
-phosphate
UDP-D
-galactose
2-Dehy dro-3
-deoxy -D
-galactonate
D-Gala
ctonate
alpha-D
-Galactose 1
-phosphate
Sucrose
Raf f inose
alpha-D
-Glucose 6
-phosphate
alpha-D
Galactosy l-(
->3)-1D-my o
Melibiitol
Epimel
ibiose
3-beta-D
-Galactosy l
-sn-gly cerol
D-Mannose
D-Sorbitol
Lactose
6-phosphate
D-Gly cera
ldehy de
3-phosphate
Melibiose
Galactan
D-Gal
actose
N-Acety l-D
-galac
tosamine
D-Gal alpha
1->6D-Gal
alpha 1
D-Tagatose
Stachy ose
Lactose
D-Galacto
samine 6
-phosphate
1.1.1.251
2.7.1.69
PTS-Gat
3.2.1.20
3.2.1.22
4.1.2.21
3.2.1.85
3.2.1.20
3.2.1.23
2.4.1.222.7.7.12
2.7.7.10
2.7.1.6 2.7.1.101
4.2.1.6
3.2.1.23 1.1.1.48
1.1.1.48
1.1.1.120
2.4.1.82
2.7.1.69 PTS
-Aga-EIIC,3.5.1.25
2.4.1.123
5.3.1.-
2.7.1.69 PTS
-Gam-EIIC,
my o-Inositol
L-Ascor
bate 6
-phosphate
UDP-g
lucose
D-Glucur
onate 1
-phosphate
L-Gala
ctonate
L-Galac
tose 1
-phosphate
GDP-L
-galactose
L-Gulose
1-phosphate
GDP-m
annose
D-Xy lulose 5
-phosphate
D-Glucuro
nolactone
2-Dehy dro-3
-deoxy -D
-glucarate
L-Gulono-1,4
-lactone
L-xy lo-Hexu
lonolactone
2-Hy droxy
-3-oxop
ropanoate
Ascorbate
5-Dehy dro-4
-deoxy -D
-glucarate
2,5-Dioxo
pentanoate
Dehy droa
scorbate
L-Ara
binose
2-Dehy dro-3
-deoxy -L
-arabinonate
2-Dehy dro-3
-deoxy -D
-xy lonate
L-Ly xonate
Threonate
3-Dehy dro-L
-threonate
2.7.-.-
3.1.1.-
5.1.3.-
5.1.3.18
1.1.3.8
3.1.1.- 1.1.1.-
3.1.3.-
2.7.7.44
1.1.1.22
5.1.3.4
4.1.1.-
2.7.1.53
4.1.1.85
5.1.3.22
1.1.1.20
1.1.1.19
1.3.2.3 4.1.2.20
4.2.1.40 4.2.1.42
1.1.3.8
1.8.5.1
1.11.1.11
4.2.1.25
1.1.1.-3.7.1.-
4.2.1.40
3.1.1.19
1.2.1.26
4.2.1.43
4.1.1.- 1.14.-.-
1.13.11.13
1.1.1.129
1.1.1.46
1.3.3.12
3.1.1.15
1.3.2.3
3.1.1.17
Py ruv ate
Py ruv ate
Py ruv ate
Py ruv ate
Py ruv ate
Py ruv ate
Py ruv ate
Py ruv ate
Py ruv ate
Fig. 5. Bottom-level graph. Reactionnetwork of pathways associated with thecarbohydrate metabolism. The search re-sult for ”Pyruvate” is highlighted includ-ing its incident edges.
Fig. 6. GUI of the Visualization and Editing Tool. The top-level graph consisting of 154pathway vertices with 4 expanded pathways. The node ”Glycolysis/Gluconeogenesis”was selected in the Data Browser (right, top) resulting in highlighting all its compoundand reaction nodes including connections to adjacent pathways.
and chain-like structures because of the larger space available to unfold thosesubstructures, but resulted in increased runtime for the brute-force method.
7 Conclusion
The proposed software is able to layout and display complex graphs with a highnumber of elements. The development process was intensively accompanied bydomain experts from biology and biochemistry. For metabolic pathway networks,not only the graph topology is relevant, a high number of additional attributes—textual annotations in our case—need to be visualized. Semantic zooming andfocus&context methods are applied to accomplish this goal, instant highlightingof graph elements fitting the pattern of a string based search operation is anintuitive way to extract specific information on the dataset. The main benefit ofthe adapted Table Lens method is the preservation of the mental map. Many ofthe visualization tools lack this key feature. Even though node expansion andcollapse produce very discrete and rather abrupt changes in the graph appear-ance, only the row and the column of the grid position are affected while theremaining elements keep their relative position. In combination with continuouszooming, it is a straightforward task to explore even large graphs. Highlight-ing individual or groups of edges greatly facilitates the tracking of routes. Inthe presented grid layout algorithm, vertex placement and edge routing are per-formed in two separate steps. This offers the opportunity to develop alternativenode placement routines fitting the specific needs of pathway visualization in thefuture.
Visualizing Metabolic Networks using Advanced Focus&Context Techniques 11
Fig. 7. A more detailed view of the bottom level graph. This portion of the graph dis-plays the pathway Starch and Sucrose metabolism. The Algorithm Info Area (bottom,right) gives feedback on invoked algorithms and displays search results. In this scenario,a search for the term alpha-D-Glucose was performed and resulted in 13 matches beinghighlighted in the Graph View and in the hierarchical Data Browser view.
Acknowledgments. This work has been funded by the Volkswagen Stiftungunder grant I/82 719.
References
1. Misue, K., Eades, P., Lai, W., Sugiyama, K.: Layout Adjustment and the MentalMap. Journal of Visual Languages and Computing 6 (2), 183–210 (1995)
2. Rao, R., Card, S.K.: The table lens: merging graphical and symbolic representa-tions in an interactive focus + context visualization for tabular information. In:Proceedings of the SIGCHI conference on Human factors in computing systems,pp. 318–322. ACM, New York (1994)
4. Hu, Z., Mellor, J., Wu, J., DeLisi, C.: VisANT: an online visualization and analysistool for biological interaction data. BMC Bioinformatics 5 (1), e17 (2004)
5. Shannon, P., Markiel, A., Ozier, O., Baliga, N. S., Wang, J. T., Ramage, D., Amin,N., Schwikowski, B., Ideker, T.: Cytoscape: a software environment for integratedmodels of biomolecular interaction networks. Gen. Res. 13 (11), 2498–2504 (2003)
6. Schreiber, F.: High Quality Visualization of Biochemical Pathways in BioPath. InSilico Biology 2 (2), 59–73 (2002)
7. Karp, P.D., Ouzounis, C.A., Moore-Kochlacs, C., Goldovsky, L., Kaipa, P., Ahren,D., Tsoka, S., Darzentas, N., Kunin, V., Lopez-Bigas, N.: Expansion of the BioCyccollection of pathway/genome databases to 160 genomes. Nucleic Acids Research33 (19), 6083–6089 (2005)
12 M. Rohrschneider et al.
8. DiBattista, G., Eades, P., Tamassia, R., Tollis, I.G.: Graph Drawing: Algorithmsfor the Visualization of Graphs. Prentice Hall, New Jersey (1999)
9. Nicholson, D.E.: Metabolic Pathways Map (Poster). Sigma Chemical Co., St. Louis(1997)
10. Lungu, M., Xu, K.: Biomedical Information Visualization. In: Human-CenteredVisualization Environments, pp. 311–342., Springer (2006)
11. Kerren, A., Ebert, A., Meyer, J.: Human-Centered Visualization Environments,GI-Dagstuhl Research Seminar. In: LNCS, vol. 4417, Springer (2007)
12. Kerren, A.: Interactive Visualization and Automatic Analysis of Metabolic Net-works – A Project Idea. Technical Report, Institute of Computer Graphics andAlgorithms, Vienna University of Technology, Austria (2003)
13. Albrecht, M., Kerren, A., Klein, K., Kohlbacher, O., Mutzel, P., Paul, W.,Schreiber, F., Wybrow, M.: On Open Problems in Biological Network Visualiza-tion. In: Proc. of the 17th International Symposium on Graph Drawing (to appear),Chicago, USA, Springer (2009)
14. Albrecht, M., Kerren, A., Klein, K., Kohlbacher, O., Mutzel, P., Paul, W.,Schreiber, F., Wybrow, M.: A Graph-drawing Perspective to Some Open Problemsin Molecular Biology. Technical report TR08-01-003, Lehrstuhl XI fur AlgorithmEngineering, Fakultat fur Informatik, TU Dortmund, Germany (2008)
15. Brandes, U., Dwyer, T., Schreiber, F.: Visualizing Related Metabolic Pathways inTwo and a Half Dimensions. In: Proc. International Symposium on Graph Drawing(GD’03), pp. 111–122, LNCS, vol. 2912 (2003)
16. Kyoto Encyclopedia of Genes and Genomes, http://www.kegg.jp/kegg/17. Klukas, C., Schreiber, F.: Dynamic exploration and editing of KEGG pathway
analysis, systems evaluation and research agenda. Information Visualization 4 (3),191–205 (2005)
20. Shneiderman, B.: The Eyes Have It: A Task by Data Type Taxonomy for Infor-mation Visualizations. In: VL, pp. 336–343 (1996)
21. Streit, M., Kalkusch, M., Kashofer, K., Schmalstieg, D.: Navigation and Explo-ration of Interconnected Pathways. Eurographics / IEEE-VGTC Symposium onVisualization 27 (3) (2008)
22. Kamada, T., Kawai, S.: An algorithm for drawing general undirected graphs. Inf.Process. Lett. 31 (1), 7–15 (1989)
23. Kirkpatrick, S., Gelatt, C. D., Vecchi, M. P.: Optimization by simulated annealing.Science 220, 671–680 (1983)
24. Hart, P.E., Nilsson, N.J., Raphael, B.: A Formal Basis for the Heuristic Deter-mination of Minimum Cost Paths. IEEE Transactions on Systems Science andCybernetics 4 (2), 100–107 (1968)
25. Batini, c., Nardelli, E., Tamassia, R.: A Layout Algorithm for Data Flow Diagrams.IEEE Trans. Software Eng. 12 (4), 538–546 (1986)
26. Kaufmann, M., Wagner, D.: Drawing Graphs, Methods and Models. In: DrawingGraphs, LNCS, vol. 2025, Springer (2001)