Deciphering regulatory networks by promoter sequence analysis

Transcription

1 Bioinformatics Workshop 2009 Interpreting Gene Lists from -omics Studies Deciphering regulatory networks by promoter sequence analysis Elodie Portales-Casamar University of British Columbia Bioinformatics Workshop - Interpreting Gene Lists from -omics Studies 1 Module #: Title of Module Bioinformatics Workshop - Interpreting Gene Lists from -omics Studies 2

2 Overview Part 1: Overview of transcription Lab 1: Promoters in Genome Browser (UCSC and PAZAR) Part 2: Prediction of transcription factor binding sites using binding profiles ( Discrimination ) Lab 2: TFBS scan (ORCAtk) Part 3: Interrogation of sets of co-expressed genes to identify mediating transcription factors Lab 3: TFBS Over-Representation (opossum) 3 Restrictions in Coverage Focus on Eukaryotic cells and PolII Promoters Principles apply to prokaryotes Will provide suggestions for similar tools for other species as requested Many of the examples drawn from the Wasserman lab s work there are equivalent tools 4

3 Part 1 Introduction to transcription in eukaryotic cells 5 Complexity in Transcription Chromatin Distal enhancer Proximal enhancer Core Promoter Distal enhancer 6

4 Studying gene expression at the bench EMSA DNase I footprinting ChIP- chip SELEX experiment Gene reporter assay Expensive and Time-Consuming!!! ww.chiponchip.org/ w w.abcam.com w w.hku.hk PAZAR and UCSC 8

5 Part 2 Prediction of TF Binding Sites Teaching a computer to find TFBS 9 TF Binding Profile Aligned binding sites TCACTATGATTCAGCAACAAA TCACAGTGAGTCGGCAAAATT TCATGCTGACTCAGCGGATCG CAACCATGACACAGCATAAAA CAGGCATGACATTGCATTTTT TAATGGTGACAAAGCAACTTT GGAGCATGACCCAGCAGAAGG CTGGGATGACATAGCATTCAT TCAGAATGACAAAGCAGAAAT TCACCGTTACTCAGCACTTTG AGGTGGTGATGTTGCATCACA CCAGGATGACTTAGCAAAAAC AGCCTGTGACTGGGCCGGGGC AGACAATGACTAAGCAGAAAT TCCCCGTGACTCAGCGCTTTG TCAGCATGACTCAGCAGTCGC CCTCCATGACAAAGCACTTTT AGCGGGTGACCAAGCCCTCAA TCAGGGTGACTCAGCAGCTTG TCTGTGTGACTCAGCTTTGGA Position Frequency Matrix (PFM) A C G T Position Specific Scoring Matrix (PSSM) A C G T A T G A T T C A G C A Score = 13.6 Binding Profile Logo 10

6 JASPAR: AN OPEN-ACCESS DATABASE OF TF BINDING PROFILES ( jaspar.genereg.net ) 11 Analysis of TFBS with Phylogenetic Footprinting Scanning a single sequence Scanning a pair orf orthologous sequences for conserved patterns in conserved sequence regions A dramatic improvement in the percentage of biologically significant detections Low specificity of profiles: too many hits great majority not biologically significant 12

7 Phylogenetic Footprinting Dramatically Reduces Spurious Hits Human Mouse Actin, alpha cardiac 13 Choosing the right species for pairwise comparison... CHICKEN MOUSE HUMAN COW HUMAN HUMAN 14

8 ORCAtk 15 TFBS Discrimination Tools Phylogenetic Footprinting Servers FOOTER CONSITE rvista ORCAtk SNPs in TFBS Analysis RAVEN Prokaryotes or Yeast PRODORIC YEASTRACT Software Packages TOUCAN Programming Tools TFBS ORCAtk 16

9 Part 3: Inferring Regulating TFs for Sets of Co-Expressed Genes 17 Two Examples of TFBS Over-Representation Foreground Foreground Background More Genes with TFBS Background More Total TFBS 18

10 Statistical Methods for Identifying Over-represented TFBS Fisher exact probability scores Based on the number of genes containing the TFBS relative to background Hypergeometric probability distribution Binomial test (Z scores) Based on the number of occurrences of the TFBS relative to background Normalized for sequence length Simple binomial distribution model 19 opossum Procedure Set of coexpressed genes Automated sequence retrieval from EnsEMBL Phylogenetic Footprinting ORCA Putative mediating transcription factors Statistical significance of binding sites Detection of transcription factor binding sites 20

11 Validation using Reference Gene Sets A. Muscle-specific (23 input; 16 analyzed) B. Liver-specific (20 input; 12 analyzed) Rank Z-score Fisher Rank Z-score Fisher SRF e-02 HNF e-08 MEF e-04 HLF e-03 c-myb_ e-03 Sox e-01 Myf e-03 FREAC e-01 TEF e-03 HNF-3beta e-02 deltaef e-02 SOX e-01 S e-01 Yin-Yang e-01 Irf e-01 S e-02 Thing1-E e-02 Irf e-01 HNF e-01 COUP-TF e-01 TFs with experimentally-verified sites in the reference sets. 21 Empirical Selection of Parameters based on Reference Studies p65 NF- _B c-rel p50 HNF-1 SRF Z-score TEF-1 MEF2 FREAC-2 Myf cebp SP1 HNF-3 _ Muscle Liver NF-_B Z-score cutoff Fisher cutoff E E E E E-01 Fisher p-value 22

12 Structurally-related TFs with Indistinguishable TFBS Most structurally related TFs bind to highly similar patterns Zn-finger is a big exception 23 opossum Server 24

13 TFBS Over-representation Analysis Tools o P O S S U M : h t t p : / / w w w. c i s r e g. c a / o P O S S U M T F M - E x p l o r e r : h ttp :/ / b i o i n f o. lifl.fr/ T F M E / fo rm A s a p : h ttp :/ / a s a p. b i n f. k u. d k / A s a p / H o m e.h t m l 25 REFLECTIONS Part 2 Futility Theorem Essentially predictions of individual TFBS have no relationship to an in vivo function Successful bioinformatics methods for site discrimination incorporate additional information (clusters, conservation) Part 3 TFBS over-representation is a powerful new means to identify TFs likely to contribute to observed patterns of co-expression Generally best performance has been with data directly linked to a transcription factor Statistical significance is extremely sensitive to gene set size TFs in the same structural family tend to have similar binding preferences 26

14 The end More tomorrow in the lab 27 Part 4: de novo Discovery of TF Binding Sites (Gibbs sampling method) 28

15 Gibbs Sampling (grossly over-simplified) ttcgctcc cgatacgc tgctacct tgacttcc agacctca ctgtagtg acgcatct A C G T Pattern Discovery Gibbs sampling is guaranteed to return an optimal pattern if repeated sufficiently often Procedure is fast, so running many 1000s of times is feasible Unfortunately, we have a problem what if the mediating TFBS are not strongly overrepresented relative to other patterns 30

16 Applied Pattern Discovery is Acutely Sensitive to Noise PATTERN SIMILARITY vs. TRUE MEF2 PROFILE Pink line is negative control with no Mef2 sites included True Mef2 Binding Sites SEQUENCE LENGTH 31 Four Approaches to Improve Sensitivity Better background models -Higher-order properties of DNA Phylogenetic Footprinting Human:Mouse comparison eliminates ~75% of sequence Regulatory Modules Architectural rules Limit the types of binding profiles allowed TFBS patterns are NOT random 32

17 Pattern Discovery Summary Pattern discovery methods can recover overrepresented patterns in the promoters of coexpressed genes Methods are acutely sensitive to noise, indicating that the signal we seek is weak TFs tolerate great variability between binding sites As for pattern discrimination, supplementary information/approaches are required to overcome the noise 33