Integrative analysis: ChIP-seq - PowerPoint Presentation

Slide1

Integrative analysis: ChIP-seq and RNA-seq

Vladimir Teif

Intro to NGS analysis

Proficio

course 2020

Slide2

NGS data integration

http://determinedtosee.com/wp-content/uploads/2014/08/jigsaw-puzzle.jpg

Slide3

1. Signal + existing annotation

deepTools 2.0

https://github.com/fidelram/deepTools/wiki/Visualizations

Slide4

Comparing cluster

heatmaps

between two cell conditions

NucTools

Slide5

Histone modifications around TSS

http://www.ie-freiburg.mpg.de/bioinformaticsfac

Slide6

Different datasets in several tracks of a genome browser

Gifford et.al., Cell 2013

5mC

Slide7

Heat maps again: Signal from data 1 around regions in data 2Here:Nucleosome

occupancyaround bound CTCFin mouse stem cells

Vainshtein

et.al.,

BMC Genomics

2017

Slide8

http://homer.salk.edu/homer/ngs/quantification.html

Correlation analysis: any 2 datasets can be correlated

Slide9

RNA-seq: how many reads per gene

DESeq, edgeR,

Cuffdiff

Slide10

ChIP-seq: where binding is enrichedMACS, CISER, HOMER, PeakSeq,

edgeR, DESeq, CisGenome

Slide11

RNA-seq & ChIP-seq together:which protein regulates which gene

ChIP-seq

peak size

Gene expression

Log fold change

Slide12

Correlation of regulatory protein binding with gene expressionPavlaki et al., 2017

Slide13

Functional analysishttps://blog.arduino.cc/2018/08/17/build-a-4-button-arcade-game-out-of-lego/

Slide14

Gene Ontology (GO)Ontology: A set of concepts and categories in a subject area or domain that shows their properties and the relations between them

.Gene ontology (GO) is a major bioinformatics initiative to unify the representation of gene and gene product attributes across all species

.

http

://

www.geneontology.org

Slide15

Gene ontology typesCellular

component, the parts of a cell or its extracellular environment;

Molecular

function

, the elemental activities of a gene product at the molecular level, such as binding or catalysis

;

Biological

process

, operations or sets of molecular events with a defined beginning and end, pertinent to the functioning of integrated living units: cells, tissues, organs, and organisms.

Slide16

Cellular component

Slide17

Molecular function

drug transporter activity

Slide18

Biological process

Slide19

GO is manually curatedBased on experimental evidence

Based on computational predictionsBased on the claims reported in publications

For example,

the

Experimental Evidence codes are:

Inferred

from Experiment (EXP)

Inferred

from Direct Assay (IDA)

Inferred

from Physical Interaction (IPI)

Inferred

from Mutant Phenotype (IMP)

Inferred

from Genetic Interaction (IGI)

Inferred

from Expression Pattern (IEP

)

Slide20

GO is manually curated

Slide21

Characteristics of GO terms >40,000 terms and growing

GO is species independent, but some terms may be specific to a certain group (

e.g

.

photosynthesis)

GO is hierarchical (terms can have parents/

childs

)

GO terms are linked by relationships:

is-a



part-of



regulates

(and +/- regulates)

occurs-in

enables

involved-in

Slide22

GO hierarchy

24th Feb 2006 Jane Lomax EBI

Slide23

GO hierarchy

Slide24

GO hierarchy

http://geneontology.org/page/ontology-structure

Slide25

Anatomy of a GO term

Adapted from Melanie Courtot

, 2012

Slide26

How GO analysis tools workModified from Jane Lomax, 2006

input

a gene list and a subset of ‘interesting’

genes

tool shows which GO categories have most interesting genes associated with them i.e. which categories are ‘enriched’ for interesting genes

tool provides a statistical measure to determine whether enrichment is significant

Slide27

Whether enrichment is significant…

~60,000 genes in total in the mouse genome

RNA-seq

2,000

genes differentially

regulated

mitosis –

80

apoptosis –

40

positive control of cell proliferation

–

30

glucose

transport

–

20

Mitosis

Apoptosis

Glucose

transport

Proliferation

Slide28

Whether enrichment is significant…

Proliferation: 3-fold enriched

Glucose transport: 4-fold

Slide29

Whether enrichment is significant…

Slide30

Other enrichment tests

Slide31

GO tools online: GSEA

http://

software.broadinstitute.org/gsea

Slide32

GO tools online: DAVID

https://david.ncifcrf.gov

Slide33

GO tools online: GOrillahttp://cbl-gorilla.cs.technion.ac.il

Slide34

GO tools online: EnrichRhttp://amp.pharm.mssm.edu/Enrichr/

Slide35

https://www.dovepress.com/role-of-nsc319726-in-ovarian-cancer-based-on-the-bioinformatics-analys-peer-reviewed-fulltext-article-OTTGO analysis, first example

Unrealistically small P-values

Slide36

GO analysis, typical exampleDAVID, GOrilla, GREAT,

EnrichR

Calo et al. (2015) Nature 518, 249–253

Slide37

GO analysis, cool and easy to do

Massie et al., EMBO J. (2011) 30, 2719–2733

Slide38

Zao et al., Cell Death & Disease (2016), 7:e2053

GO analysis:cool,but

difficult

to do

Slide39

GO analysis“manual”

Red - up

Blue - down

Yellow -

unchanged

Massie

et al., EMBO J. (2011) 30, 2719–2733

Slide40

Problems of GO analysisSince we do many tests (one for each

term i) we encounter the multiple

testing problem: the

results significance is

not as

high

as individual P

i

suggest.

Both

the ontology and the annotations are updated

regularly. Results

can change

overnight. Make sure

you know

GO versions

you have used in each

analysis.

When

using a tool to study

GO enrichment

, make

sure you

know what the software is

doing.

Gene

lists produced in publications to tackle the same problem

using different software solutions can have

almost no overlap



Slide41

Adapted from http://cs273a.stanford.edu [Bejerano Fall09/10]

C

ombinatorial TF binding

Gene

Proteins

DNA

DNA

TF binding

site

Sequence

logo

R

egulatory region

The expression of a gene is determined by a combination of TFs simultaneously bound to its regulatory DNA region

Slide42

Blais

and

Dynlacht

(2005)

Genes Dev.,

19

, 1499-1511

Genes act together in gene networks

Slide43

Network visualisation

pathwaycommons.org