Microarray Cancer Data Visualization Analysis in Relation to Pharmacogenomics

1
Microarray Cancer Data Visualization Analysis in
Relation to Pharmacogenomics

• By Ngozi Nwana

2
Microarray Data Acquisition

• What is Microarray
• Microarray data (scanned image data of expressed
  genes) are obtained from microscope slides that
  contain an ordered series of samples (DNA, RNA,
  Protein, Tissue).
• The type of microarray depends on the material
  placed on it, for example DNA, DNA Microarray,
  RNA, RNA Microarray etc. The most commonly used
  microarray is the DNA microarray.
• DNA Microarrays are ordered sets of gene-specific
  probes fixed to a solid support to which
  fluorescently labeled samples (with reverse
  transcriptase enabling RNA to bind to spots of
  cDNA) are hybridized for use in massively
  parallel gene expression studies.

3
Background Definition of Keywords

• Genetics has been the primary discovery engine
  for modern biomedical science
• Genetics is the study of heredity and how traits
  are passed on through generations
• Genomics is the study of genes and their
  functions
• Every human cell (with some rare exceptions)
  contains 46 (organized as 23 pairs) linear
  chromosomes (pieces of DNA).
• The chromosomes contain genetic information,
  which is organized into thousands of different
  genes
• A gene is a stretch of DNA, which codes for a
  particular protein, whether it is a structural
  protein (a protein that makes up part of a
  structure of the cell, for example the cell wall)
  or an enzyme.

4
Microarray Technology and Pharmacogenomics

• Microarray technology has enabled many advances
  in gene study (genomics science).
• It provides a method of collecting thousands of
  individual qualitative (such as gene category )
  and/or quantitative (such as RNA level for an
  entire experiment), measurements/attributes
  simultaneously in a single sample.
• The oncology field has been especially active and
  to an extent successful in using microarrays to
  differentiate between cancer cell types and to
  obtain molecular signatures of the state of
  activity of diseased cells of patient samples.
• This approach of studying cancer provides a
  better understanding of the underlying mechanism
  for tumorigenesis, more accurate diagnosis, more
  comprehensive prognosis, and more effective
  therapeutic interventions

5
Microarray Data and Pharmacogenomics contd

• Pharmacogenomics
• - studies the way a person responds to a drug,
  by studying the inherited variations in genes
  that dictate drug response including negative,
  positive or no response)
• General Practice
• Current drug therapy is empirically prescribed
  to fit the needs of the average patient.
• Effect
• Empirical prescription leads to undue toxicity
  in cured patients and delays alternative active
  therapies while causing unnecessary toxicity in
  resistant ones.
• Goal
• To obtain new and widely applicable validated
  predictors of the likelihood of optimal drug
  therapy response that will enable individually
  tailored prescriptions.

6
Visualization of Microarray

• Visualization of microarray
• - Enable the simultaneous visualization of
  multiple expressed gene data attributes
• - Provides visualized summaries of gene
  expression data
• - Provides genome researchers with meaningful
  details (gene cluster summary, map position
  within the genome, gene /protein sequences for
  effective disease recognition
• Visualization attributes
• - Quantitative attributes
• - RNA level p-Value Size of expressed genes
• - Qualitative
• Color
• Size and color are two attributes that can be
  used to display quantitative differences in data
  using most visualization tools
• Visualization methods that enable the ability to
  simultaneously visualize multiple data attributes
  including the analysis of qualitative information
  about either gene families or biological function
  and quantitative information such as RNA level
  and p-value simultaneously are very important.

7
Source Microarray Visualization data

• BRAC 1 BRAC 2 (Onset) Microarray real-time data
• Control data from healthy cells
• Cells from patients undergoing treatment and have
  undertaken neoadjuvant chemotherapy (treatment of
  locally advanced and inoperable breast cancer-
  given before surgery)
• - aims at reducing tumor size and increasing
  rates of breast conserving treatment

8
GenePix Sample Data Format
Rank NAME Ch1 Net (Mean) Ch2 Normalized Net (Mean) Log(base2) of R/G Normalized Ratio (Mean) Regression Correlation Spot Flag
1 IMAGE199180 223 1 -7.986 0.799 0
2 IMAGE810625 119 1 -7.08 0.635 0
3 IMAGE52228 119 1 -7.08 0.611 0
4 IMAGE141726 631 10 -6.027 0.705 0
5 IMAGE74537 17093 330 -5.696 0.787 0
6 PEROU5D10 451 12 -5.195 0.946 0
7 IMAGE436741 11601 474 -4.613 0.686 0
8 IMAGE682522 711 30 -4.571 0.842 0
9 IMAGE782730 2852 126 -4.503 0.776 0
10 IMAGE41648 5618 269 -4.384 0.662 0
11 IMAGE46620 47 3 -4.155 0.782 0
12 IMAGE51865 4595 352 -3.707 0.615 0
13 IMAGE587847 4140 318 -3.701 0.71 0
14 IMAGE109440 10454 821 -3.671 0.679 0
15 IMAGE199367 21288 1749 -3.605 0.96 0
16 IMAGE247281 177 15 -3.565 0.674 0
17 IMAGE276688 110 10 -3.507 0.621 0
18 IMAGE80186 7123 636 -3.486 0.913 0
19 IMAGE214572 3764 339 -3.475 0.791 0
20 IMAGE810911 1497 136 -3.457 0.622 0
9
MATLAB Gene Spatial image Representations

• The command gprread reads the data from the file
  into a structure.
• pd.ColumnNames enabled the read to the Structure
  name/Fieldname fields with the following
  resulting spatial images of microarray data.
• Figuremaimage (pd,'F635 Median')
• Notice the very high background levels down the
  right side of the array. Areas of high color
  intensity signifies high level gene expression.

10
Visualization results contd

• Visualization results scanned at 532F for breast
  cancer cells
• The "F532 Median" field corresponds to the
  foreground of the green (Cy3) channel.
• Figure maimage(pd,'F532 Median')

11
Visualization results for the untreated Control
sample scanned at 532F
12
Clustering Commands

• The xlsread function can be used to read in the
  data from the XLS file and load the data into
  MATLAB
• numericData, textData xlsread(cancerdata.x
  ls)
• This reads the data in the spreadasheet in
  two variable, numericData (stores numeric values)
  and textData for text values
• giValues numericData (,2 end)
• drugMechanism textData(2 end,1)
• To perform the clustering, the command below is
  used
• clustergram(giValues, rowlabels, drug,
  columnlabels, tumorTypes)

13
Cluster figure
14
Visualization results

• A Subsection example of Unsupervised hierarchical
  clustering

15
Microarray Data Visualization Results

• Cross section of Hierarchical Clustering of
  expressed genes

16
Conclusion

• Significant differences in gene expression in
  cancer specimens before and after treatment were
  observed
• Differences in the microarray spatial images
  between the control and diseased cancer genes
  were observed.
• Further confirmation of whether the drug used is
  providing effective therapy is an oncologists
  call.

17
Q A

• Thank
  You