Unraveling the segmentation clock with
bioinformatics tools and techniques.
Sourav Roy
School of Informatics
Indiana University
Advisors: Dr. Predrag Radivojac & Dr. Santiago Schnell
Capstone Instructor: Dr. Mehmet Dalkilic.
11/24/2011 1
Capstone Presentation 04/21/2006
Overview
Introduction
• Background and Motivation
• Notch Signaling Pathway (NSP)
• Structure of NSP proteins (NSPPs)
• Potential Binding Sites in NSPPs
• Probable Protein-Protein Interactions
Sourav Roy
• Interaction Map & Boolean Model for
School of Informatics
the Pathway
Indiana University
11/24/2011 2
Capstone Presentation 04/21/2006 1
Beginning of Life
Introduction
Background
zygote Cleavage
Sourav Roy
School of Informatics
Indiana University
Blastula Start of Gastrulation
11/24/2011 3
Capstone Presentation 04/21/2006 2
Vertebrate segmentation
Introduction
Background
• Segmentation was nature’s answer to the
development of complex organisms
• Presomitic mesoderm (PSM) or paraxial
mesoderm gives rise to vertebrate equal-
sized segments – somites
Sourav Roy
• Repeated formation of intersomitic
boundaries is driven by a molecular
School of Informatics
oscillator (segmentation clock)
Indiana University
11/24/2011 4
Source: Cancer Research UK Website
Capstone Presentation 04/21/2006 3
Periodicity
Introduction
Background
Motivation
Sourav Roy
School of Informatics
Indiana University
11/24/2011 5
Capstone Presentation 04/21/2006 4
Why are we interested?
Introduction
Background
Motivation
Sourav Roy
School of Informatics
Indiana University
To date models of the pathway have been built on the
basis of knockout experiments, but nobody knows
whether there is a single interaction or a cascade for
most of the cases
11/24/2011 6
Capstone Presentation 04/21/2006 5
Notch Signaling Pathway
Introduction
Background
Motivation
Notch Pathway
Sourav Roy
School of Informatics
Indiana University
11/24/2011 Source - 7
Biocarta
Capstone Presentation 04/21/2006 6
Proteins in Mouse Notch Signaling Pathway
Introduction
Background
Motivation
Notch Pathway
• Notch1 - Transmembrane receptor.
• Dll1 - Ligand.
Sourav Roy
• Dll3 - Ligand.
• Psen1 - Membrane bound protein with
School of Informatics γ- secretase activity.
• Lfng - Beta-1,3-N-
Indiana University acetylglucosaminyltransferase.
• RbpSuh - DNA binding transcription.
• Mesp2 - bHLH protein.
• Hes1 - bHLH transcription repressor.
• Hes7 - bHLH transcription repressor.
11/24/2011 8
Capstone Presentation 04/21/2006 7
Flexibility, Rigidity and binding
Introduction
Background
Motivation
Notch Pathway
Structure of NSPPs
Sourav Roy
School of Informatics
Indiana University
11/24/2011 9
Capstone Presentation 04/21/2006 8
Order and Disorder
Introduction
Background
Motivation
Ordered Regions
Notch Pathway
Structure of NSPPs
Sourav Roy
School of Informatics
Indiana University
Disordered Regions
11/24/2011 10
Source : Bachinsky and V.V. Solovyev
Capstone Presentation 04/21/2006 9
Disorder Prediction
Introduction
Background • Disorder has a role to play in signal transduction, cell
cycle regulation and transcriptional activity (Dunker et al.
Motivation
2002; Iakoucheva et al. 2002; Ward et al. 2004)
Notch Pathway
Structure of NSPPs • Swiss-Prot (June 2005, version) database was
downloaded and all mouse proteins were extracted with
the help of a Matlab code
• VL3 model (Obradovic et al., 2003) was used to
calculate the disorder for each protein
Sourav Roy
School of Informatics
Indiana University
11/24/2011 11
Capstone Presentation 04/21/2006 10
Notch Signaling Pathway Proteins
Introduction Intrinsically Disordered?
Background
Motivation
Notch Pathway
Structure of NSPPs Proteins %Disorder
Disorder in the
Average of 9
Pathway pathway proteins 56.4
Average of all 9448 32.2
Mouse Proteins in the
Sourav Roy Swiss Prot database
on 05/10/2005
School of Informatics
Indiana •University was used with VL3 within it to
A Matlab code
calculate the average disorder
11/24/2011 12
Capstone Presentation 04/21/2006 11
Disorder Percentage in general
Introduction
Background
Motivation Sampling Method I:
Notch Pathway
Structure of NSPPs a) 10 / 30 / 60 / 90 / 100 / 300 / 600 / 900 /1000
Disorder in the random samples were taken from the set of
Pathway
9448 proteins.
b) Average Disorder and standard deviation
Sourav Roy with the help of a Matlab code.
calculated
School of Informatics
c) t-test done – Fischer’s t-test.
Indiana University
• Steps a) and b) were done with the help of a Matlab
code
• JAVA code was used for the t-test
11/24/2011 13
Capstone Presentation 04/21/2006 12
t-test results
Introduction
Background
# of Proteins Mean Sd P value Significance
Motivation
9 56.4 25.2 0.0062 Yes
Notch Pathway
10 26.6 23.0 0.5029 No
Structure of NSPPs
Disorder in the 60 31.6 26.5 0.8572 No
Pathway 100 29.8 23.1 0.3628 No
600 30.6 25.7 0.1556 No
1000 32.1 26.1 0.9283 No
Sourav Roy
Average Disorder in Mouse Proteins
School of Informatics
60
50
Disorder
%% Disorder
40
Indiana University
30 1
2
20
3
10
4
0
5
1
9 2
10 3
60 4 5 6
100 600 1000
6
## of Proteins
Sample size
of Proteins
11/24/2011 14
Capstone Presentation 04/21/2006 13
Sampling Method II
Introduction
Background
Motivation
Notch Pathway
a) Different sample sets 1,2,3,…,11 were taken.
Structure of NSPPs
Disorder in the b) Each of the sets had 9 random proteins from the
Pathway entire set of 9448.
c) The average of the 9 proteins in each set was
calculated.
Sourav Roy
The Informatics
School ofaverage of the averages and standard of the
d)
sample sets was calculated.
Indiana University
• A Matlab code was used for this method
11/24/2011 15
Capstone Presentation 04/21/2006 14
Sampling Method II results
Introduction
Introduction
Background
Background
Motivation
& Motivation Average Disorder in Mouse Proteins
Notch Pathway
Notch Pathway
Structure of NSPPs
Ordered & Dis- 60
Disorder in the
ordered Proteins 50
Pathway
Disorder in the
% Disorder 40
Pathway
30
Roy
Sourav20
10
School of Informatics
0
N
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 11 11
10 12
Indiana University # of Sample sets
11/24/2011 16
Capstone Presentation 04/21/2006 15
Potential binding sites
Introduction
Background
• MoREs - short, interaction-prone loosely-
Motivation
structured or semi-structured regions in
Notch Pathway
intrinsically disordered proteins
Structure of NSPPs
Disorder in the • α-MoRE a subclass of MoREs
Source: Oldfield et al Biochemistry 2005
Pathway
MoREs
Ordered Regions
Sourav Roy
School of Informatics
α-MoRE
Indiana University
Disordered Regions
11/24/2011 17
Source : Bachinsky and V.V. Solovyev
Capstone Presentation 04/21/2006 16
α-MoRE Prediction
Introduction
Background
Motivation
Notch Pathway
Structure of NSPPs
α-MoRE score
Disorder in the
Pathway
MoREs
Sourav Roy
Amino acid residue
School of Informatics
Indiana University
11/24/2011 18
Capstone Presentation 04/21/2006 17
Disordered Regions and α-MoREs
Introduction
Background
Motivation
Notch Pathway
Structure of NSPPs
Disorder in the
Pathway
MoREs
Sourav Roy
School of Informatics
Indiana University
Ordered regions, Disordered regions
α-MoREs at p >0.5 α-MoREs at p >0.7
11/24/2011 19
Capstone Presentation 04/21/2006 18
To check if the disordered regions
Introduction and the α-MoREs are conserved
Background
Motivation
Notch Pathway • PSI- BLAST of the pathway proteins
Structure of NSPPs • Extraction of the aligned regions with the help
Disorder in the of a PERL script
Pathway
• Clustering of the aligned regions with the help
MoREs
of PERL script
Evolution
• Multiple
Sourav Roy sequence alignment of the clustered
sequences by ClustalW
• Parsing the Clustal
School of Informatics report and calculating the
entropy of each column with the help of a
Indiana University
Matlab code
11/24/2011 20
Capstone Presentation 04/21/2006 19
Results for Hes7
Introduction
Background
Motivation
Notch Pathway
Structure of NSPPs
Disorder in the
Pathway
MoREs
Evolution
Sourav Roy
School of Informatics
Indiana University
11/24/2011 21
Capstone Presentation 04/21/2006 20
Prediction of possible
protein-protein interactions
Introduction ADVICE webtool was used for the prediction
Background ADVICE is a web server providing Automated Detection and Validation of
Motivation Interaction based on the Co-Evolutions between interacting proteins. It
automated the steps needed to compute the similarities between proteins'
Notch Pathway evolutionary histories to detect co-evolved interacting proteins.
Structure of NSPPs
Step 1. Homolog Search.
Disorder in the ADVICE detects orthologous sequences for pair(s) of protein sequences and
retrieve the orthologous sequences if both appear in the same species.
Pathway
MoREs Step 2. Distance Matrix Construction.
ADVICE then constructs the distance matrix for both orthologous group of
Evolution sequence. The distance matrices are derived from multiple sequences
Probable PPI alignments using
Sourav Roy ClustalW.
Step 3. Linear correlation coefficient computation.
School of Informatics
The Pearson's Correlation Coefficient formula is used to calculate the
similarities between the two distance matrices:
Indiana University
The result r will fall into -1 to 1. Previous studies have indicated that interacting
proteins share similarity in their evolutionary histories and have high r-value
(>=0.8)
11/24/2011 22
Source: ADVICE website
Capstone Presentation 04/21/2006 21
Predicted PPI
Introduction
Background
Motivation Chances of
Notch Pathway Protein Pairs interaction (r)
Structure of NSPPs Notch1 & Dll1 89.50%
Disorder in the Notch1 & Psen1 97.20%
Pathway
Notch1 & Lfng 98.90%
MoREs
Notch1 & RbpSuh 98.20%
Evolution
Probable PPI RbpSuh & Hes7 97.70%
Sourav Roy
Hes1 & Lfng 97.80%
School of Informatics
Notch1 & Hes7 97.70%
Lfng & Dll1 88.70%
Indiana University
11/24/2011 23
Capstone Presentation 04/21/2006 22
Interaction Map
Introduction
Background
Motivation
Notch Dll1
Notch Pathway
Structure of NSPPs Dll Notch NICD
Disorder in the
Psen1
Rbp/Suh
pathway
MoREs NICD
Rbp/Suh
Evolution
Nucleus
Probable PPI Sourav Roy dll
Interaction map
School of Informatics mesp2 Mesp2
Indiana University lfng Lfng
hes1/7 Hes1/7
Cell1 Cell2
11/24/2011 24
Capstone Presentation 04/21/2006 23
Boolean Logic
Introduction
Background
Motivation
Mesp2t ( Notch1t Psen1t RbpSuht )
Notch Pathway Dll1/3t+1
Structure of NSPPs
Disorder in the Dll1/3t Lfngt Notch1t+1
pathway
MoREs {( Notch1t Psen1t RbpSuht ) Mesp2t }
Evolution Hes1/7t Lfngt+1
Probable PPI Sourav Roy
Interaction map
Notch1t Psen1t RbpSuht Mesp2t+1
Boolean Model School of Informatics
(Notch1t Psen1t RbpSuht) Hes1/7t
Indiana University
Hes1/7t+1
Notch1t Psen1t RbpSuht+1
11/24/2011 25
Capstone Presentation 04/21/2006 24
Expression of hes/her regulates the
Introduction segmentation clock
Background
Motivation
Notch Pathway
Structure of NSPPs
Disorder in the
pathway
MoREs
Evolution
Probable PPI Sourav Roy
Interaction map
Boolean Model School of Informatics
The molecular network constituting the segmentation clock in two adjacent cells.
Indiana University
Notch pathway
HER1/HER7- based oscillator, excluding the Notch pathway components
Source: Pourquie & Goldbeter, Current Biology 2003
11/24/2011 26
Capstone Presentation 04/21/2006 25
Oscillations within the pathway
Introduction
Background time Dll1 Notch1 Psen1 Lfng Mesp2 RbpSuh Hes7
Motivation 0 1 0 1 0 0 0 0
Notch Pathway 1 1 1 1 0 0 0 0
Structure of NSPPs 2 1 1 1 1 1 1 0
Disorder in the 3 0 0 1 1 1 1 1
pathway 4 0 0 1 0 0 0 0
MoREs 5 1 0 1 0 0 0 0
Evolution 6 1 1 1 0 0 0 0
Probable PPI 7 1 1 1 1 1 1 0
Sourav Roy
8 0 0 1 1 1 1 1
Interaction map
9 0 0 1 0 0 0 0
Boolean Model School of Informatics
10 1 0 1 0 0 0 0
11 1 1 1 0 0 0 0
Indiana University
12 1 1 1 1 1 1 0
13 0 0 1 1 1 1 1
14 0 0 1 0 0 0 0
15 1 0 1 0 0 0 0
11/24/2011 27
Capstone Presentation 04/21/2006 26
Review
Introduction
Background What was known?
Motivation
Notch Pathway
• Somitogenesis - controlled by segmentation
Structure of NSPPs clock.
Disorder in the
pathway
• Segmentation clock in turn is controlled by the
MoREs
expression of the hairy and enhancer of split
Evolution family of genes.
Probable PPI Sourav Roy
Interaction map
• Expression of Hes family determined by the
School of Informatics
Boolean Model
Notch signaling pathway.
Indiana University
11/24/2011 28
Capstone Presentation 04/21/2006 27
Review
Introduction
Background Information from this study:
Motivation • The NSPPs have more disordered regions than
Notch Pathway usual
Structure of NSPPs
Disorder in the • The MoREs seem to be either in the disordered
pathway regions or near to those – probable binding sites
MoREs for binding partners
Evolution
Probable PPI • These
Sourav Roy regions have low entropy and therefore
Interaction map seem to be conserved
Boolean Model School of Informatics
• Interaction map on the basis of predicted ppi
Indiana University
and genetic interactions
• Got the required oscillations from the Boolean
11/24/2011
model 29
Capstone Presentation 04/21/2006 28
Review
Introduction
Background
Motivation
Notch Pathway
Structure of NSPPs
Disorder in the
pathway
MoREs
Evolution
Probable PPI Sourav Roy
Interaction map
Boolean Model School of Informatics
Indiana University
11/24/2011 30
Capstone Presentation 04/21/2006 29
Acknowledgements
Dr. S. Schnell.
Dr. P. Radivojac.
Dr. M. Dalkilic.
Dr. H. Tang.
Dr. S. Kim.
Dr. C. Raphael.
Junguk Hur.
Systems Biology Group. Thanks !!!
School of Informatics.
Sourav Roy
Everyone related to
Bioinformatics at IUB.
School of Informatics
My family and friends.
Indiana University
11/24/2011 31
Capstone Presentation 04/21/2006