Andreas Holzinger
Lecture 5 – Version 2013/14
Semi structured & weakly structured data
Structural homologies
VO 444.152 Medical Informatics
a.holzinger@tugraz.at
A. Holzinger 444.152
1/75
Med Informatics L5
Schedule
 1. Intro: Computer Science meets Life Sciences, challenges, future directions
 2. Back to the future: Fundamentals of Data, Information and Knowledge
 3. Structured Data: Coding, Classification (ICD, SNOMED, MeSH, UMLS)
 4. Biomedical Databases: Acquisition, Storage, Information Retrieval and Use
 5. Semi structured and weakly structured data (structural homologies)
 6. Multimedia Data Mining and Knowledge Discovery
 7. Knowledge and Decision: Cognitive Science & Human‐Computer Interaction
 8. Biomedical Decision Making: Reasoning and Decision Support
 9. Intelligent Information Visualization and Visual Analytics
 10. Biomedical Information Systems and Medical Knowledge Management
 11. Biomedical Data: Privacy, Safety and Security
 12. Methodology for Info Systems: System Design, Usability & Evaluation
A. Holzinger 444.152
2/75
Med Informatics L5
Learning Goals … at the end of the 5th lecture you …
 … have an overview on various dimensions of data in biomedical informatics;
 … are aware of the various contents of Electronic Patient Records;
 … have seen some application examples of topological structures from both macro‐cosmos and micro‐cosmos and are fascinated about it;
 … have a rough overview about some basics of computational topology;
 … have an understanding of the challenges of weakly structured data;
A. Holzinger 444.152
3/75
Med Informatics L5
Keywords of the 5th Lecture












Big data pools
Complex networks
Computational graph representation
Electronic patient record (EPR)
Homology modeling
Macroscopic structures
Medical documentation
Metabolic network
Microscopic structures
Network metrics
Structural data dimension
Topological structures
A. Holzinger 444.152
4/75
Med Informatics L5
Advance Organizer (1/3) A‐G
 Adjacency matrix = simplest form of computational graph representation, in which 0 or 1 denotes whether or not there is a directed edge from one node to another (in graph theory adjacent nodes in a graph are linked by an edge);
 Artifacts = not only a noise disturbance, which is contaminating and influencing the signal (surrogates) but also data which is wrong, however interpreted as to be reliable, consequently may lead to a wrong decision;
 Computational graph representation = e.g. by adjacency matrices  Data fusion = data integration techniques that analyze data from multiple sources in order to develop insights in ways that are more efficient and potentially more accurate than if they were developed by analyzing a single source of data. Signal processing techniques can be used to implement some types of data fusion (e.g. combined sensor data in Ambient Assisted Living);
 Global Distance Test (GDT) = a measure of similarity between two protein structures with identical amino acid sequences but different tertiary structures. It is most commonly used to compare the results of protein structure prediction to the experimentally determined structure as measured by X‐ray crystallography or protein NMRM;
 Graph theory = study of mathematical structures to model relations between objects from a certain collection;
 Graphs = a hypothetical structure consisting of a series of nodes connected by weighted edges (graphs can be directed/undirected and stoichometric/non‐
stoichometric regarding interaction classes);
A. Holzinger 444.152
5/75
Med Informatics L5
Advance Organizer (2/3) H‐P










Homology = in mathematics (especially algebraic topology and abstract algebra), it is (ὁμόιος homos = "identical") a certain general procedure to associate a sequence of Abelian groups (i.e. does not depend on their order) or modules with a given mathematical object such as a topological space or a group; Homology modeling = comparative modeling of protein, refers to constructing an atomic‐resolution model of the "target" protein from its amino acid sequence and an experimental three‐dimensional structure of a related homologous protein (the "template"); in Bioinformatics, homology modeling is a technique that can be used in molecular medicine. In silico = via computer simulation, in contrast to in vivo (within the living) or in vitro (within the glass);
Multi‐scale representation = in a graph, nodes do not have to represent biological objects on the same scale, one node (e.g. a molecule) may have an edge connecting it to a node representing a cell or tissue (the edge indicates that the molecule exerts an effect on the cell/tissue);
Network = graphs containing cycles or alternative paths;
Network analysis = a set of techniques used to characterize relationships among discrete nodes in a graph or a network;
Network topology = the shape or structure of a network;
Petri‐Net = a special class of graph, consisting of two general classes or node: place and transition nodes;
Predictive modeling = a set of techniques in which a mathematical model is created or chosen to best predict the probability of an outcome (e.g. regression);
P‐System = addresses the slowness of Petri‐nets
A. Holzinger 444.152
6/75
Med Informatics L5
Advance Organizer (3/3) R‐V










Radius of a graph = average minimum path length (biological networks are not arranged in a regular or symmetrical pattern);
Scale‐free Topology = ensures that there are very short paths between any given pair of nodes, allowing rapid communication between otherwise distant parts of the network (e.g. the Web has such a topology);
Semi‐structured data = does not conform with the formal structure of tables/data models assoc. with relational databases, but at least contains tags/markers to separate semantic elements and enforce hierarchies of records and fields within the data; aka schemaless or self‐describing structure; the entities belonging to the same class may have different attributes even though they are grouped together;
Spatial analysis = a set of techniques, applied from statistics, which analyze the topological, geometric, or geographic properties encoded in a data set;
Structural homology = similar structure but different function; Supervised learning = machine learning techniques that infer a function or relationship from a set of training data (e.g. classification and support vector machines);
Time series analysis = set of techniques from both statistics and signal processing for analyzing sequences of data points, representing values at successive times, to extract meaningful characteristics from the data;
Time series forecasting = use of a model to predict future values of a time series based on known past values of the same or other series (e.g. structural modeling); decomposition of a series into trend, seasonal, and residual components, which can be useful for identifying cyclical patterns in the data;
Unstructured data = complete randomness, noise; (wrongly, text is called unstructured, but there is some structure, too, so text data is a kind of weakly structured data);
Vertex degree = within a topology, the numbers of edges connecting to a node;
A. Holzinger 444.152
7/75
Med Informatics L5
Glossary





















ANSI = American National Standards Institute
CD = cardiac development
CDA = Clinical Document Architecture
CHD = congenital heart disease
CMM = Correlated motif mining
DPI = Dossier Patient Integre´ = integrated patient record
E = Edge
EPR = Electronic Patient Record
G(V,E) = Graph
GI = gastrointestinal
HER = Electronic Health Record
HL7 = Health Level 7
KEGG = Kyoto Encyclopedia of Genes and Genomes
NP = nondeterministic polynomial time
OWL = Web Ontology Language
PPI = Protein‐Protein Interaction
SGML =
Standard Generalized Markup Language
TF= Transcription factor
TG = Target Gene
V = Vertex
XML = Extensible Markup Language
A. Holzinger 444.152
8/75
Med Informatics L5
Slide 5‐1: Mathematically seen our world is …
Complex and
High dimensional
Geschwind, D. H. & Konopka, G. 2009. Neuroscience in the era of functional genomics and systems biology. Nature, 461, (7266), 908‐915.
A. Holzinger 444.152
9/75
Med Informatics L5
Weakly‐Structured
Slide 5‐2: Remember: Standardization/Structurization
Omics Data
Natural
Language
Text
Well‐Structured
XML
Databases
Libraries
RDF, OWL
Standardized
A. Holzinger 444.152
Non‐Standardized
10/75
Med Informatics L5
Slide 5‐3: Example: Well‐Structured Data
http://care2x.org
A. Holzinger 444.152
11/75
Med Informatics L5
Slide 5‐4: Example: Semi‐structured Data: XML
<?xml version="1.0"?>
<patient>
<patient-id>11111</patient-id>
<Name>Chen</Name>
<Date of Birth>1.1.1900</Date of Birth>
<diagnosis>
<code>123</code>
<diagnosistext>Myocardinfarct</diagnosistext>
</diagnosis>
</patient>
Holzinger, A. (2003) Basiswissen IT/Informatik. Band 2: Informatik. Das Basiswissen für die Informationsgesellschaft des 21. Jahrhunderts. Wuerzburg, Vogel Buchverlag.
A. Holzinger 444.152
12/75
Med Informatics L5
Slide 5‐5 Example: Generic XML template for a med. report
DPI = Dossier
Patient Integre´ = integrated patient record
Rassinoux, A.‐M., Lovis, C., Baud, R. & Geissbuhler, A. (2003) XML as standard for communicating in a document‐based electronic patient record: a 3 years experiment. International Journal of Medical Informatics, 70, 2‐3, 109‐115.
A. Holzinger 444.152
13/75
Med Informatics L5
Slide 5‐6 Comparison of XML ‐ RDF/OWL in Bioinformatics
Louie, B., Mork, P., Martin‐
Sanchez, F., Halevy, A. & Tarczy‐Hornoch, P. 2007. Data integration and genomic medicine. Journal of Biomedical Informatics, 40, (1), 5‐16.
A. Holzinger 444.152
14/75
Med Informatics L5
Slide 5‐6 Example: Weakly structured data set ‐ PPI
Kim, P. M., Korbel, J. O. & Gerstein, M. B. 2007. Positive selection at the protein network periphery: Evaluation in terms of structural constraints and cellular context. Proceedings of the National Academy of Sciences, 104, (51), 20274‐20279.
A. Holzinger 444.152
15/75
Med Informatics L5
Networks = Graphs + Data
A. Holzinger 444.152
16/75
Med Informatics L5
Slide 5‐7: Complex Biological Systems key concepts
 In order to understand complex biological systems, the three following key concepts need to be considered:
 (i) emergence, the discovery of links between elements of a system because the study of individual elements such as genes, proteins and metabolites is insufficient to explain the behavior of whole systems;  (ii) robustness, biological systems maintain their main functions even under perturbations imposed by the environment; and  (iii) modularity, vertices sharing similar functions are highly connected.  Network theory can largely be applied for biomedical informatics, because many tools are already available
A. Holzinger 444.152
17/75
Med Informatics L5
Slide 5‐8: Networks on the Example of Bioinformatics
, …
…
, ∈ ;
,
Hodgman, C. T., French, A. & Westhead, D. R. (2010) Bioinformatics. Second Edition. New York, Taylor & Francis.
A. Holzinger 444.152
18/75
Med Informatics L5
Slide 5‐9: Computational Graph Representation
,
,
Adjacency (ə‐ˈjā‐sən(t)‐sē) Matrix 2
1
5
3
4
6
1
3
1
2
3
,
2
3
2
5
1
∈
6
5
4
Simple graph, symmetric, binary
Directed and weighted
For more information: Diestel, R. (2010) Graph Theory, 4th Edition. Berlin, Heidelberg, Springer.
A. Holzinger 444.152
19/75
Med Informatics L5
Slide 5‐10: Example: Tool for Node‐Link Visualization Jean‐Daniel Fekete http://wiki.cytoscape.org/InfoVis_Toolkit
Fekete, J.‐D. The infovis toolkit. Information Visualization, INFOVIS 2004, 2004. IEEE, 167‐174.
A. Holzinger 444.152
20/75
Med Informatics L5
Slide 5‐11: Some Network Metrics (1/2)
Order = total number of nodes n; Size = total number of links (a):
Clustering Coefficient (b) = the degree of concentration of the connections of the node’s neighbors in a graph and gives a measure of local inhomogeneity of the link density:
2
1
1
Path length (c) = is the arithmetical mean of all the distances:
1
1
Costa, L. F., Rodrigues, F. A., Travieso, G. & Boas, P. R. V. (2007) Characterization of complex networks: A survey of measurements. Advances in Physics, 56, 1, 167‐242.
A. Holzinger 444.152
21/75
Med Informatics L5
Slide 5‐12: Some Network Metrics (2/2)
 Centrality (d) = the level of “betweenness‐ centrality” of a node I (“hub‐node in Slide 28);
 Nodal degree (e) = number of links connecting i to its neighbors: ∑
Modularity (f) = describes the possible formation of communities in the network, indicating how strong groups of nodes form relative isolated sub‐networks within the full network (refer also to Slide 5‐8).
A. Holzinger 444.152
22/75
Med Informatics L5
Slide 5‐13: Network Topologies
Das Bild k ann zurzeit nicht angezeigt werden.
Scale‐free network
Modular network
Van Heuvel & Hulshoff (2010) A. Holzinger 444.152
23/75
Med Informatics L5
Slide 5‐14: Small‐World Networks
Increasing randomness
21.000 citations …
Watts, D. J. & Strogatz, S. (1998) Collective dynamics of small‐world networks. Nature, 393, 6684, 440‐442.
Milgram, S. 1967. The small world problem. Psychology today, 2, (1), 60‐67.
A. Holzinger 444.152
24/75
Med Informatics L5
Slide 5‐15 Graphs from Point Cloud Data Sets
Lézoray, O. & Grady, L. 2012. Graph theory concepts and definitions used in image processing and analysis. In: Lézoray, O. & Grady, L. (eds.) Image Processing and Analysing With Graphs: Theory and Practice. Boca Raton (FL): CRC Press, pp. 1‐24.
A. Holzinger 444.152
25/75
Med Informatics L5
Slide 5‐16 Graphs from Images
a) quadtree tessellation
b) RAG assoc. to the quadtree
c) Watershed Algorithm
d) SLIC superpixels
Lézoray, O. & Grady, L. 2012. Graph theory concepts and definitions used in image processing and analysis. In: Lézoray, O. & Grady, L. (eds.) Image Processing and Analysing With Graphs: Theory and Practice. Boca Raton (FL): CRC Press, pp. 1‐24.
A. Holzinger 444.152
26/75
Med Informatics L5
Slide 5‐18 Example Watershed Algorithm
Meijster, A. & Roerdink, J. B. A proposal for the implementation of a parallel watershed algorithm. Computer Analysis of Images and Patterns, 1995. Springer, 790‐795.
A. Holzinger 444.152
27/75
Med Informatics L5
Slide 5‐19 Graphs from Images: Watershed + Centroid
A. Holzinger 444.152
28/75
Med Informatics L5
Slide 5‐20 Graphs from Images: Voronoi <> Delauney
A. Holzinger 444.152
29/75
Med Informatics L5
Slide 5‐21 Points ‐> Voronoi ‐> Delaunay
Kropatsch, W., Burge, M. & Glantz, R. 2001. Graphs in Image Analysis. In: Kropatsch, W. & Bischof, H. (eds.) Digital Image Analysis. Springer New York, pp. 179‐197.
A. Holzinger 444.152
30/75
Med Informatics L5
Slide 5‐22 Example: Graph Entropy Measures
Holzinger et al. 2013. On Graph Entropy Measures for Knowledge Discovery from Publication Network Data. In: LNCS 8127, 354‐362.
A. Holzinger 444.152
31/75
Med Informatics L5
Slide 5‐23: Example for a Medical Knowledge Space
# Nodes: 641
# Edges: 1250
Agent
Condition
Pharmacological Group
Other Documents
A. Holzinger 444.152
Average Degree: 3,888
Average Path Length: 4.683
Network Diameter: 9
Holzinger, A., et al. 2013. Constraints of List‐based Knowledge Interaction. In: Medicine 2.0 London, in print.
32/75
Med Informatics L5
Slide 5‐24: Medical Details of the Graph
 Nodes






drugs
clinical guidelines
patient conditions (indication, contraindication)
pharmacological groups
tables and calculations of medical scores
algorithms and other medical documents
 Edges: 3 crucial types of relations inducing medical relevance between two active substances
 pharmacological groups
 indications
 contra‐indications
A. Holzinger 444.152
33/75
Med Informatics L5
Slide 5‐25: Example for the shortest path
Holzinger, A., et al. 2013. Constraints of List‐based Knowledge Interaction. In: Medicine 2.0 London
Henzinger, M. R., Klein, P., Rao, S. &
Subramanian, S. 1997. Faster shortest-path
algorithms for planar graphs. Journal of
Computer and System Sciences, 55, (1), 3-23.
A. Holzinger 444.152
34/75
Med Informatics L5
Slide 5‐26: Example for finding related structures
Relationship between Adrenaline (center black node) and Dobutamine (top left black node)
Blue: Pharmacological Group
Dark red: Contraindication; Light red: Condition
Green nodes (from dark to light):
1. Application (one ore more indications + corresponding dosages)
2. Single indication with additional details (e. g. “VF after 3rd Shock”)
3. Condition (e.g. VF, Ventricular Fibrillation)
A. Holzinger 444.152
35/75
Med Informatics L5
Slide 5‐27: Example: The brain is a complex network
Van Den Heuvel, M. P. & Hulshoff Pol, H. E. (2010) Exploring the brain network: a review on resting‐state fMRI functional connectivity. European Neuropsycho‐
pharmacology, 20, 8, 519‐534.
A. Holzinger 444.152
36/75
Med Informatics L5
Slide 5‐28: Representative Examples of disease complexes
Examples of 4 functional networks driving the development of different anatomical structures in the human heart of a 37‐day old human embryo
Lage, K. et. al (2010) Dissecting spatio‐temporal protein networks driving human heart development and related disorders. Molecular systems biology, 6, 1, 1‐9.
A. Holzinger 444.152
37/75
Med Informatics L5
Slide 5‐29: Example: Cell‐based therapy Lage et. al (2010) A. Holzinger 444.152
38/75
Med Informatics L5
Slide 5‐30: Identifying Networks in Disease Research
Schadt, E. E. & Lum, P. Y. (2006) Reverse engineering gene networks to identify key drivers of complex disease phenotypes. Journal of lipid research, 47, 12, 2601‐2613.
A. Holzinger 444.152
39/75
Med Informatics L5
Slide 5‐31: Three main types of biomedical networks
Transcriptional regulatory network with two components:
TF = transcription factor
TG = target genes
(TF regulates the transcription of TG)
A. Holzinger 444.152
Protein‐Protein interaction network
Metabolic network
(constructed considering the reactants, chemical reactions and enzymes)
Costa, L. F., Rodrigues, F. A. & Cristino, A. S. (2008) Complex networks: the key to systems biology. Genetics and Molecular Biology, 31, 3, 591–601.
40/75
Med Informatics L5
Slide 5‐32: Example Transcriptional Regulatory Network
Salgado, H., Santos‐
Zavaleta, A., Gama‐
Castro, S., Peralta‐Gil, M., Peñaloza‐Spínola, M. I., Martínez‐
Antonio, A., Karp, P. D. & Collado‐Vides, J. 2006. The comprehensive updated regulatory network of Escherichia coli K‐12. BMC bioinformatics, 7, (1), 5.
A. Holzinger 444.152
41/75
Med Informatics L5
Slide 5‐33: Network Representations of Protein Complexes
Protein complex
True PPI topology
Spoke‐Model
Matrix‐Model
Wang, Z. & Zhang, J. Z. (2007) In search of the biological significance of modular structures in protein networks. PLoS Computational Biology, 3, 6, 1011‐1021.
A. Holzinger 444.152
42/75
Med Informatics L5
Slide 5‐34 Correlated Motif Mining (CMM)
Boyen, P., Van Dyck, D., Neven, F., van Ham, R. C. H. J. & van Dijk, A. (2011) SLIDER: A Generic Metaheuristic for the Discovery of Correlated Motifs in Protein‐Protein Interaction Networks. Computational Biology and Bioinformatics, IEEE/ACM Transactions on, 8, 5, 1344‐1357.
A. Holzinger 444.152
43/75
Med Informatics L5
Slide 5‐35 Steepest Ascent Algorithm applied to CMM
Boyen et al. (2011) A. Holzinger 444.152
44/75
Med Informatics L5
Slide 5‐36: Metabolic Network
E1
2 M5
M4
E2
M1
M2
E3
M3
M1
M2
M1
M4
M1
M5
M2
M1
M1
M2
M3
M4
M5
M2
M3
M1
0
1
0
1
1
M2
M4
M2
1
0
1
1
0
M4
M1
M3
0
0
0
0
0
M5
M1
M4
1
0
0
0
0
M5
1
0
0
0
0
Matrix contains many sparse elements ‐ In this case it is computationally more efficient to represent the graph as an adjacency list
A. Holzinger 444.152
45/75
Hodgman, C. T., French, A. & Westhead, D. R. (2010) Bioinformatics. Second Edition. New York, Taylor & Francis.
Med Informatics L5
Slide 5‐37 Metabolic networks are usually big … big data … Schmid, A. K., Reiss, D. J., Pan, M., Koide, T. & Baliga, N. S. (2009) A single transcription factor regulates evolutionarily diverse but functionally linked metabolic pathways in response to nutrient availability. Molecular Systems Biology, 5, 1‐9.
http://www.nature.com/msb/journal/v5/n1/fig_tab/msb200940_F6.html
A. Holzinger 444.152
46/75
Med Informatics L5
Slide 5‐38 Using EPRs to Discover Disease Correlations
Electronic patient records remain a unexplored, but potentially rich data source for example to discover correlations between
diseases. Roque, F. S., Jensen, P. B., Schmock, H., Dalgaard, M., Andreatta, M., Hansen, T., Søeby, K., Bredkjær, S., Juul, A., Werge, T., Jensen, L. J. & Brunak, S. (2011) Using Electronic Patient Records to Discover Disease Correlations and Stratify Patient Cohorts. PLoS
Computational Biology, 7, 8, e1002141.
A. Holzinger 444.152
47/75
Med Informatics L5
Slide 5‐39: Heatmap of disease‐disease correlations (ICD)
Roque, F. S. et al (2011) Using Electronic Patient Records to Discover Disease Correlations and Stratify Patient Cohorts. PLoS
Comput Biol, 7, 8, e1002141.
A. Holzinger 444.152
48/75
Med Informatics L5
Slide 5‐40: Example: ὁμολογέω (homologeo)
He, Y., Chen, Y., Alexander, P., Bryan, P. N. & Orban, J. (2008) NMR structures of two designed proteins with high sequence identity but different fold and function. Proceedings of the National Academy of Sciences, 105, 38, 14412.
A. Holzinger 444.152
49/75
Med Informatics L5
Slide 5‐41 Conclusion
 Homology modeling is a knowledge‐based prediction of protein structures.  In homology modeling a protein sequence with an unknown structure (the target) is aligned with one or more protein sequences with known structures (the templates).  The method is based on the principle that homologue proteins have similar structures.
 Homology modeling will be extremely important to personalized and molecular medicine in the future. A. Holzinger 444.152
50/75
Med Informatics L5
Slide 5‐42: Future Outlook
Personalized
Medicine
EB
Proteomics
PB
TB
Genomics
2003
A. Holzinger 444.152
2013
51/75
2023
Med Informatics L5
Thank you! A. Holzinger 444.152
52/75
Med Informatics L5
Sample Questions
 Which are the four main “big data” pools in the health care domain and what problems involved?
 What is the main problem in medical documentation?  What is the advantage of an integrated Patient record?
 What are the advantages/disadvantages of XML/OWL for data in bioinformatics?
 What are the three key concepts in order to understand complex biological systems?
 What are the main symbols describing a network as used in Bioinformatics?
 How can networks represented computationally effectively?
 What are the main network metrics?
 What are the main network topologies used in Biomedical informatics?
 What is the Small‐World Theory?
 Why is the study of networks relevant for medical professionals?
 Which are the three main types of biomedical networks?
 What is a Motif?
 What benefits can we gain from Correlated Motif Mining (CMM)?
 What is more efficient if a matrix contains many sparse elements?
 Why are structural homologies interesting for biomedical informatics?
A. Holzinger 444.152
53/75
Med Informatics L5
Some Useful Links







http://www.cdisc.org
http://www.w3.org/Math/
http://www.sgpp.org/structures.shtml
http://salilab.org/modeller
http://swissmodel.expasy.org
http://www.expasy.org/tools
http://www.geneticseducation.nhs.uk
A. Holzinger 444.152
54/75
Med Informatics L5
Appendix: clustering network motifs in integrated networks
http://omics.frias.uni‐freiburg.de/
A. Holzinger 444.152
55/75
Med Informatics L5
Example from Immunology: Structural Homology
Calandra, T. & Roger, T. 2003. Macrophage migration inhibitory factor: a regulator of innate immunity. Nat Rev Immunol, 3, 791‐800.
A. Holzinger 444.152
56/75
Med Informatics L5
Klein Bottle
http://www.maa.org/cvm/1998/01/tprppoh/article/Pictures/KleinBottle.gif
A. Holzinger 444.152
57/75
Med Informatics L5
Medical Documentation – Patient Record
A. Holzinger 444.152
58/75
Med Informatics L5
Medical Documentation ‐ Electronic Patient Record
http://care2x.org
A. Holzinger 444.152
59/75
Med Informatics L5
Challenge is in Genomic medicine …
 … to integrate and analyze these diverse and voluminous data sources to elucidate both normal and disease physiology.  XML is suited for describing semi‐structured data including a natural modeling of biological entities, because it allows features as e.g. nesting …
A. Holzinger 444.152
60/75
Med Informatics L5
Example: Comparison of XML and OWL data in bioinformatics
difficulty of modeling many‐to‐
many relationships, such as the relationship between genes and functions
Louie, B., Mork, P., Martin‐Sanchez, F., Halevy, A. & Tarczy‐
Hornoch, P. (2007) Data integration and genomic medicine. Journal of Biomedical Informatics, 40, 1, 5‐16.
A. Holzinger 444.152
61/75
Med Informatics L5
On time and space of data …
A. Holzinger 444.152
62/75
Med Informatics L5
… to microscopic atomistic structures
Wiltgen, M. & Holzinger, A. (2005) Visualization in Bioinformatics: Protein Structures with Physicochemical and Biological Annotations. In: Central European Multimedia and Virtual Reality Conference. Prague, Czech Technical University (CTU), 69‐74 A. Holzinger 444.152
63/75
Med Informatics L5
First yeast protein‐protein interaction network (2001)
Nodes = proteins
Links = physical interactions (bindings)
Red Nodes = lethal
Green Nodes = non‐lethal
Organge = slow growth
Yellow = not known Jeong, H., Mason, S. P., Barabasi, A. L. & Oltvai, Z. N. (2001) Lethality and centrality in protein networks. Nature, 411, 6833, 41‐42.
A. Holzinger 444.152
64/75
Med Informatics L5
The Nature of Space and Time
A. Holzinger 444.152
65/75
Med Informatics L5
Clouds of data – unordered sequence of points in n‐dim
Let us collect ‐dimensional observations:
Point cloud in topological space
metric space
Zomorodian, A. J. 2005. Topology for computing, Cambridge (MA), Cambridge University Press.
A. Holzinger 444.152
66/75
Med Informatics L5
Example: To predict the folding of a protein
Source: Theoretical and computational Biophysics Group: http://www.ks.uiuc.edu/
A. Holzinger 444.152
67/75
Med Informatics L5
Backup Slide: Overview Some Network Metrics
Van Heuvel & Hulshoff (2010) A. Holzinger 444.152
68/75
Med Informatics L5
Slide 5‐19: Watershed Principle
 Catchment basins:  treating an image as a height field or landscape, regions where the rain would flow into the same lake  Start flooding from local minima, and label ridges wherever differently evolving components meet
A. Holzinger 444.152
69/75
Med Informatics L5
A. Holzinger 444.152
70/75
Med Informatics L5
Slide 5‐15 Graphs from Images: Voronoi <> Delauney
A. Holzinger 444.152
71/75
Med Informatics L5
A. Holzinger 444.152
72/75
Med Informatics L5
Example: Cell based therapy (1) (Heart transplantation)
Chien, K. R., Domian, I. J. & Parker, K. K. (2008) Cardiogenesis and the complex biology of regenerative cardiovascular medicine. Science, 322, 5907, 1494.
A. Holzinger 444.152
73/75
Med Informatics L5
Example: Cell based therapy (2) (Heart transplantation)
Chien et al. (2008)
A. Holzinger 444.152
74/75
Med Informatics L5
Example: Network Generated by Gene Duplication
High Modularity (Modularity = 0.6717, Scaled
Modularity = 29);
Different colors represent different modules identified by Guimera and Amaral’s
algorithm [28].
Guimera R, Amaral LAN (2005) Functional cartography of complex
metabolic networks. Nature 433: 895–900.
Wang & Zhang (2007) A. Holzinger 444.152
75/75
Med Informatics L5