of Regulatory Network Ping Ma Department of Statistics & - - PowerPoint PPT Presentation

of regulatory network
SMART_READER_LITE
LIVE PREVIEW

of Regulatory Network Ping Ma Department of Statistics & - - PowerPoint PPT Presentation

Nonparametric Modeling of Regulatory Network Ping Ma Department of Statistics & Institute for Genomic Biology University of Illinois Urbana-Champaign Central Dogma of Molecular Biology Transcription Regulation Gene A Gene B TF TF


slide-1
SLIDE 1

Nonparametric Modeling

  • f Regulatory Network

Ping Ma Department of Statistics & Institute for Genomic Biology University of Illinois Urbana-Champaign

slide-2
SLIDE 2

Central Dogma of Molecular Biology

slide-3
SLIDE 3

Transcription Regulation

Transcription factors (regulatory proteins) bind to genes, turning on or shutting off their expressions.

……TTCGA…….CCCGG……CCCGG…..CGCGGGCTTACGATATAACG

TF A TF B

Gene C

Gene A Gene B

slide-4
SLIDE 4

Transcription Factor Binding

 Transcription Factor Binding Motif (TFBM):

Common patterns in DNA sequences at transcription factor binding sites. RAP1 GCN4

slide-5
SLIDE 5

Transcription Factor Binding

slide-6
SLIDE 6

Transcription Factor Binding

Mardis Nat

  • Meth. 2007
slide-7
SLIDE 7

Gene Expression

 To quantify the abundance of each transcript  Two approaches:

Hybridization (Microarray) Sequence (RNA-Seq)

slide-8
SLIDE 8

Linking gene expression with TF binding

 Linear Regression

Motif Regressor (Conlon et al 2003 PNAS) Motif Express (Zamdborg and Ma 2009 NAR)

 Nonlinear Regression

RSIR (Zhong et al 2005, Bioinformatics) Correlation Pursuit (Zhong et al 2012, JRSSB)

slide-9
SLIDE 9

Converting Gene Expression to Clusters

 Gene expression is noisy  Clustering gene expression to get robust clusters  Linking gene clusters with TF binding data.

Bayesian Network (Beer and Tavazoie 2004 Cell) Proportional Odds Model (Yuan et al 2007 PLoS Comput. Biol.)

slide-10
SLIDE 10

Desirable Features

 Flexible function form to link gene expression

(clusters) with TF binding

 Integration of new expression data

slide-11
SLIDE 11

Our Method

 Gene expression clusters and TF binding

slide-12
SLIDE 12

Penalized Likelihood

slide-13
SLIDE 13

Functional ANOVA

slide-14
SLIDE 14

Penalized Likelihood

slide-15
SLIDE 15

Inference

slide-16
SLIDE 16

Bayesian Confidence Interval

slide-17
SLIDE 17

Mixed Effect Models

slide-18
SLIDE 18
slide-19
SLIDE 19
slide-20
SLIDE 20
slide-21
SLIDE 21
slide-22
SLIDE 22
slide-23
SLIDE 23

Software

 R package gss

http://cran.r-project.org/web/packages/gss/

slide-24
SLIDE 24

Joint work with Chong Gu