1. Data Mining Functionalities (2)
Classification and Prediction
Finding models (functions) that describe and distinguish classes or concepts for future prediction
E.g., classify countries based on climate, or classify cars based on gas mileage
Presentation: decision-tree, classification rule, neural network
Prediction: Predict some unknown or missing numerical values

2. Data Mining Functionalities (3)
Cluster analysis
Class label is unknown: Group data to form new classes, e.g., cluster houses to find distribution patterns
Clustering based on the principle: maximizing the intra-class similarity and minimizing the interclass similarity

3. Learned
Model
Use it to
predict
If acceptable,
Learning
Classification

4. Classification—A Two-Step Process
Model construction: describing a set of predetermined classes
Each tuple/sample is assumed to belong to a predefined class, as determined by the class label attribute
The set of tuples used for model construction: training set
The model is represented as classification rules, decision trees, or mathematical formulae
Model usage: for classifying future or unknown objects
Estimate accuracy of the model
The known label of test sample is compared with the classified result from the model
Accuracy rate is the percentage of test set samples that are correctly classified by the model
Test set is independent of training set, otherwise over-fitting will occur

5. Classification Process (1): Model Construction
Algorithms
Training
Data
Classifier
(Model)
IF rank = ‘professor’
OR years > 6
THEN tenured = ‘yes’
class label

6. Classification Process (2): Use the Model in Prediction
Classifier
Testing
Data
IF rank = ‘professor’
OR years > 6
THEN tenured = ‘yes’
Unseen Data
(Jeff, Professor, 4)
Tenured?

7. Supervised vs. Unsupervised Learning
Supervised learning (classification)
Supervision: The training data (observations, measurements, etc.) are accompanied by labels indicating the class of the observations
New data is classified based on the training set
Unsupervised learning (clustering)
The class labels of training data is unknown
Given a set of measurements, observations, etc. with the aim of establishing the existence of classes or clusters in the data

8. (1) Quality Control Issue:
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
Exp #j
xxxx
xxxx
(1) Quality Control Issue:
In each experiment, can you do some preliminary statistical
analysis that can assure at least the experiment has not
been contaminated? (e.g., add sufficient amount of controls)

9. (2) Generating Bias Free Data:
xxxx
Exp #1
Exp #2
Exp #3
Exp #4
Exp #5
Exp #6
(2) Generating Bias Free Data:
How to design the experiments so that you can offset variations
caused by experimental biases and errors?
e.g., redundant spot printing,
switch dyes,
change spot printing order, etc.

10. Microarray Expression
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
GeneID
(3) Getting the Representative Values:
How many experiments to produce statistically significant representative expression values? (average vs. mode)

11. How to partition expression data? Is one partitioning
Clustering 1
Clustering 2
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
(4) Clustering
How to partition expression data? Is one partitioning
better than the other – can you tell this without performing
association mining?

12. (5) Association mining - 1 Can you associate each fragment
Clustering i
xxxx A
xxxx
xxxx B
xxxx C
xxxx
xxxx D
xxxx
xxxx E
xxxx
xxxx F
xxxx
xxxx G
xxxx
xxxx H
xxxx
xxxx I
xxxx
(5) Association mining - 1
Can you associate each fragment
with some meaningful data?

13. (5) Association mining - 1 Can you associate each fragment
Clustering i
xxxx A
xxxx
xxxx B
xxxx C
xxxx
xxxx D
xxxx
xxxx E
xxxx
xxxx F
xxxx
xxxx G
xxxx
xxxx H
xxxx
xxxx I
xxxx
(5) Association mining - 1
Can you associate each fragment
with some meaningful data?

14. (5) Association mining - 1 Can you associate each fragment
Clustering i
xxxx A
xxxx
xxxx B
xxxx C
xxxx
xxxx D
xxxx
xxxx E
xxxx
xxxx F
xxxx
xxxx G
xxxx
xxxx H
xxxx
xxxx I
xxxx
(5) Association mining - 1
Can you associate each fragment
with some meaningful data?

15. (5) Association mining - 1 Can you associate each fragment
Clustering i
xxxx A
xxxx
xxxx B
xxxx C
xxxx
xxxx D
xxxx
xxxx E
xxxx
xxxx F
xxxx
xxxx G
xxxx
xxxx H
xxxx
xxxx I
xxxx
(5) Association mining - 1
Can you associate each fragment
with some meaningful data?

16. (5) Association mining - 2
Clustering i
xxxx A
xxxx
xxxx B
xxxx C
xxxx
xxxx D
xxxx
xxxx E
xxxx
xxxx F
xxxx
xxxx G
xxxx
xxxx H
xxxx
xxxx I
xxxx
(5) Association mining - 2
Can you associate each fragment with transcription
binding site data (promoter analysis) ?

17. Signal Transduction Pathway

18. (Gene_id, Family, Domain)
 Protein id
 Items
MGR4_HUMAN
 G-protein Receptor Family3,
Extracellular Ligand-binding Receptor
MGR4_RAT
G-protein Receptor Family3,
COPP_HELFE
 Copper Ion Binding Family,
Heavy Metal Transport/Detoxification Protein Domain
GLHA_RABIT
 Glycoprotein Hormones Alpha Chain Family,
Growth Factor Domain
Table 6: BioDataset1 Transactional Database

19. (Gene_id, Domain, GO Cellular Component, GO Biological Process )
 Protein name
 Items
 Homeobox Protein Antennapedia Type Domain,
Nucleus,
Regulation of Transcription DNA-Dependent
HXA7_HUMAN
 Homeobox Protein Antennapedia Type Domain,
Nucleus,
Regulation of Transcription DNA-Dependent
 O42504
 3’5’Cyclic Nucleotide Phosphodiesterose Domain, ?,
Signal Transduction
 CN4B_HUMAN
3’5’Cyclic Nucleotide Phosphodiesterose Domain, ?,
Signal Transduction
CN4D_RAT
Table 7: BioDataset2 Transactional Database

20. (Gene_id, Family, fingerprint, GO Cellular Component, GO Biological Process )
 Protein name
 Items
ORA1_LOCMI
GPCR Rhodopsin-like Superfamily,
GPCRRHODPSN,
Integral to Membrane,
G-Protein Coupled Receptor Protein Signaling Pathway
ORA2_LOCMI
Octapamine Receptor Family,
OCTOPAMINER,
A1AB_MESAU
ML1A_HUMAN
Melatonin1A Receptor Family,
MELATONIN1AR,
Table 8: BioDataset3 Transactional Database

21. What are the genotype features that could be used in clustering microarray expression data?
Codon Usage
Degree of Phylogenetic conservation 2.1 Strong purifying selection 2.2 Positive selection 2.3 Neutral selection
Domain architecture – Single or Multiple domain
Promoter Data 4.1 Boxes (Transcription factor binding site) 4.2 TATA regions 4.3 Regulatory Protein (Transcription factor)
# of introns, exons

22. Dynamic class labeling?
Phenotype
Genotype
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
GeneID
GeneID
Dynamic class labeling?

23. Joint DMS/NIGMS Initiative to Support Research
in the Area of Mathematical Biology - NSF
Examples of areas of research
1. Evolutionary theory and practice arising from genomics advances
2. Statistical and other approaches to the discovery of genes
contributing to complex behavior, and their environmental interactions
3. Explanatory and predictive models of the cellular state
4. Growth, motility, cell division, membrane trafficking, and other cellular behavior
5. Metabolic circuitry and dynamics
6. Signal transduction
7. Informational molecule dynamics
8. Design principles and dynamics of pattern formation in development
and differentiation
9. New approaches to the prediction of molecular structure
10. Improved algorithms for structure determination by x-ray crystallography,
Nuclear Magnetic Resonance (NMR), and electron microscopy
11. Simulations of the human systemic responses to burn, trauma, and other injury
12. New approaches to understanding system-wide effects of pharmacological
agents and anesthetics, and their genetic and environmental modifiers

24. Finding Pathway partners
You find Ecsit
Now study Ecsit!

25. Discussion
Use genotype and phenotype (expression intensity) together to do analysis
The efforts required to build the genotype list
How to utilize the promoter sequences in microarray experiments