A Typical Eukaryotic PremRNA Structure:

1
(No Transcript)
2
(No Transcript)
3
A Typical Eukaryotic Pre-mRNA Structure
4
(No Transcript)
5
What is alternative splicing?
6
pre-mRNA of fibronectin
Exon 2
Exon 3
Exon 4
Exon 5
Intron 1
Intron 2
Exon 1
Alternative splicing
Exon 3
Exon 5
Exon 1
Mature mRNA in Firoblast cell
Mature mRNA in Hepatocyte cell
7

• A few facts about alternative splicing in human
  genome
• Many alternatively spliced mRNAs may be expressed
  
• simultaneously in the same tissue.
• 30-60 of genes undergo alternative splicing.
• But the presence of multiple alternatively
  spliced mRNA
• forms is not addressed in microarray design and
  analysis.

8
Known Gene Structure information 21 well
characterized genes
Design a special chip of splicing variants
Statistical Modeling for assessing the amounts
of Splicing variants
9
Idea
Concentration of features containing the probes
Intensities of Specially designed probes
Amounts of the Individual splicing variants
Features are exons or exon-exon junctions Gene
Structure specifies the features of each
alternative Splicing variant (or form).
10
Design of probes
11
Statistical Modeling
Splicing variants
Features
Mapped to
Probe
Mapped to
12
Relationship between features and splicing
variants (or forms, transcripts)
G is q x t matrix rows q features
columns t splicing variants
entry glk is 1 or 0 1 means presence of l
feature in k variant (or transcript)
0 means absence of l feature in k
variant (or transcript)
Relationship between splicing variants (or forms,
transcripts) and samples (or experiments)
T is t x x matrix rows t splicing variants
columns x samples entry tkj is
the concentration of variant k in sample j
variant
samples
features
variant
13
C is a q x x matrix with the entry clj
representing the concentration of feature l in
sample j
samples
features
Next, Capture the relationship between the probes
and features with Matrix F
F (fil) is a p x q matrix with values 0 or
1. fil 1 if probe pair i belongs to feature
l 0 otherwise
features
Probe pairs
14
In matrix X, the entry xij represents the actual
concentration of all the target variants (or
transcripts) in sample j interrogated by probe
pair i
samples
Probe pairs
15
Question how to relate X to the matrix Y of
observed probe pair intensity differences
(PM-MM)?
samples
Probe pairs
16
Affinity matrix A for probe pairs, like that in
Li and Wongs model
A (aii) is a p x p diagonal matrix where aii
represents the affinity term of probe pair i.
17
With random error
Obtain the model estimates for A and T by least
square A affinity terms for probe pairs T
The amounts of splicing variants in samples
18

• Model validation with experiment
• Dilution experiments to test the accuracy and
• sensitivity.
• Two-variant spike experiments
• genes like CD44 have two splice variants
• Three sets
• Set 1 first variant ranges from 0 to 64 pM
• second variant ranges from 64 to 0 pM
• the two variants mixed and the total
  Concentration held at 64 pM.
• Set 2 set 1 diluted 4 times, and the total
  concentration reached 16 pM
• Set 3 Set 1 diluted 16 times, and the total
  concentration reached 16 pM

19
(No Transcript)
20
Three-variant spike experiments A third CD44
variant added Designed to test all possible
combinations of clones at 0 and 4 pM
21
Tissue experiments with TPM2 gene Two variants
TPM2-A, containing exon A, mainly present in
skeletal muscle TPM2-B, containing exon B,
in esophagus, stomach, uterus, etc TagMan
experiment is done for validation.
Microarray experiment
TaqMan Experiment
22
(No Transcript)
23
(No Transcript)