t ime warping of evolutionary distant temporal gene expression data based on noise suppression yury...
Post on 19-Dec-2015
217 views
TRANSCRIPT
1
TIME WARPING OF EVOLUTIONARY DISTANT TEMPORAL GENE EXPRESSION DATA BASED ON NOISE SUPPRESSIONYury Goltsev and Dmitri Papatsenko
*Department of Molecular and Cell biology, University of California, Berkeley, USA
* BMC Bioinformatics Oct. 2009
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwan
2
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
OVERVIEW
Introduction Dynamic Time Wraping
Kruskal's algorithm Orthologous Genes
Alignments of cell cycles Simulation data Real data
Conclusion Reference
3
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
INTRODUCTION
Dynamic Time Wraping(DTW)
0 10 20 30 40 50 60
-2-1
01
2
t
f
g
0 10 20 30 40 50
-2-1
01
23
4
t
f,g
*http://dblab.cs.nccu.edu.tw/presentation/DynamicTimeWarping.ppt
4
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
INTRODUCTION
Dynamic Time Wraping(DTW)
*http://www.evanlin.com/document/DataMiniing/Discovering similar time series patterns with fuzzy clustering and DTW methonds.ppt
5
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
DYNAMIC TIME WRAPING
Kruskal's algorithm
jiji
jiji
jiji
ji
badD
badD
badD
D
,2
,2
,2
min
1,
,1
1,1
,
*http://Fmccmb.belozersky.msu.ru/2009/Presentations/20090721MCCMB-Papatsenko.ppt
),...,,,(
,
321
1
2
aiM
ai
ai
ai
ai
M
k
bjk
aik
bj
aiE
xxxxX
xxXXd
6
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
INTRODUCTION
Orthologous Genes Homology
Budding yeast Fission yeast
Saccharomyces cerevisiae Schizosaccharomyces pombe
7
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
INTRODUCTION
Cell cycles
* Rustici G, et al., Nat Genet. 2004, 36:809-817
8
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
ALIGNMENTS OF CELL CYCLES
Periodic patterns (A,B: simulation data)
9
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
ALIGNMENTS OF CELL CYCLES
Periodic patterns (real data) A – Euclidean distance matrix B – Alignment path
No alignment
10
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
ALIGNMENTS OF CELL CYCLES
Periodic patterns (real data) A – Pearson distance matrix B – Alignment path
M
k
abbj
aiP nkjirXXd
1
2,,,1,
11
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
ALIGNMENTS OF CELL CYCLES
Use Pearson distance rather than Euclidean distance
12
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
ALIGNMENTS OF CELL CYCLES
Remove noise from the data(use Gaussian filter) A – Pearson distance matrix B – Alignment path
13
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
ALIGNMENTS OF CELL CYCLES
matching cell cycle markers to good valleys
14
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
ALIGNMENTS OF CELL CYCLES
matching cell cycle markers to good valleys G1 phase is longer in
S.cerevisiae G2 is longer in S.pombe.
15
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
CONCLUSION
Desynchronization of gene expression in evolution
Microarray data and low-level data processing
16
VC
Lab, D
ept. o
f Com
pute
r Scie
nce
, NTH
U, Ta
iwa
n
REFERENCE
http://dblab.cs.nccu.edu.tw/presentation/DynamicTimeWarping.ppt
http://www.evanlin.com/document/DataMiniing/Discovering similar time series patterns with fuzzy clustering and DTW methonds.ppt
http://Fmccmb.belozersky.msu.ru/2009/Presentations/20090721MCCMB-Papatsenko.ppt