CHIP-Seq GENOME ANALYSIS

TFBS/Chromatin landscape important

Euka: Pomoteur distal, proximal, enhancer, silencer …

● Traditional approaches: Footprint Assay, EMSA = determine factor binding site

CHIP: IN VIVO
detect specifid DNA region associated with protein of interest in VIVO

Chip-​SEQ​​ : DNA size selection for sequecing library

Etapes:
- Cross Linking FA (De-Cross Linking => Optimiszation Crucial)
- Cell Lysis
- Imunoprecipitation (antibody bead = SPECIFICITY NEEDED)
- Decrosslinking & purification. (​Reverse FA: SDS 65°C + NACL)
- Analysis: rtPCR, sequencing

Sequencing
improved sensitivity & less background
sequençage et alignement
POLONY MULTIPLEX Sequencing (=​immobilized DNA sequences in parallel) = Colony + Pol

Generation Polony Array: DNA Beads = 454, Solid

Chip-Seq > Chip-chip

but: sequencing error, bias toward high G-C, insufficient nb of read, technical issues &​ COST

ANTIBODY QUALITY:​​ validation : sensivitve & specific

/!\ Cross reactivity
CELL NUMBER​​: 10^6 ou 7 cells = 10-100 ng DNA
CHROMATIN FRAGMENTATION: ​manageable size (150 - 300pb) sonication ou RE ou Micrococcal
Nuclease. Different condition for cell type
CONTROL EXEPERIMENT:​​ /!\ repeat seq, shearing of DNA, open chromatine & stuff
- INPUT DNA (before IP) = control bias chromatine frag + greater coverage of genome
- DNA from nonspecific IP (random antibody IGG)
IP factor = several fold above control
- NO TAG Control: strain with protein that has no tag to be analysed

REPLICATE:​​ duplicate at least of bio expe + different AB

Final Analysis of DNA: 5np, 100-400 pb with gel

Library Construction:​​ standard protocol specific to seq platform

No PCR overamplification
Sequencing:​​ read of 30 to 35 pb. 25 min. Number of read depends of AB and target
Paired-End: ​more coverage, improved effeciency for repeat, detect frag size
Dont over sequence: not statisticaly significant results for peak
Multiplexing
DATA ANALYSIS:
Small number of mismatch & SNP indel. Multi-reads discarded (multiple sides alignement)
2 peak: for the two short reads from both side of fragment => moyenne = un seul pic résultant

Challenges
- Genome Alignement
- Identification of enriched regions = la moyenne des deux pics
- DATA management (storing)

ATF7: bind près de gènes, juste avant le +1

enrichissement + methylation tag = transcriptional factor
BUT also inhibitor

Chip-Seq tools: ENCODE,

- MACS ​= Model-based Analysis for ChIP-seq = locate real biding site, model shift size of
chip-seq (moyenne des deux pics)
distance deux pics D: prot fixe a D/2 bravo macs (nul)
Remove duplicates tags too
+ donne un FDR = false discovery rate
MACS recommentadation: more chip tag than control: FDR overly optimistic
EPIGENOMICS

INTRO
Histone modification + DNA methylation
Epigenetic reprogramming: cancel epigen modification for multipotent cells (stem cells)
Remove epigenomic signature during dev
Genomic imprinting = need of male & female
Mamalian: two epigenetic reprogrammatin (between egg = embryo and between embryo =
gametes)
Transmission of epigenetic = controversial

DNA METH IN MAMALS

5mC principalement. Symmetrical CpG diNuc
catalysed with DNMTs
3A 3B: de novo meth
1: maintenance meth
Other form by TET prot of cytosines. (intermediaire of demeth)
CpG island = lot of CpG, 70% of prom, ​unmethylamted ​in normal cell
Methylated cytosines = SILENT chromatine

Some CpG island remains unmethylated:

Human distribution: far less CG than expected
Pattern of CpG: created by deletions of island everywhere in genome (constantly
lost) (and not positive selection of CpG island)
Methylation of CpG make you loose them: that’s why most CpG are not methylated
coz they would be removed

DISEASES
DNMT1 mut: neuropath, Immunodeficiency DNMT3B
PAttern of DNA met in cancer.
Leukemia = DNMT et TET mut
The CpG island methylator phenotype (CIMP) in cancer
Drug: degrade DNMT1 for cancer = very toxic

ANALYSE EPIGENOMIC
Illmunia + Chip-Seq. (Histone modif)
AB @5mC. Or 5mC RE
- Sodium Bisulfite conversion
Unmethylated C -> U -> T in sequencing
Count how many read have a T and how many have a C
7T 3C in reads = 30% methylated
Single nucleotide resolution

RRBS: reduced representation bisulfite sequencing​​: only seq interesting part of the genome
because high sequencing cost
Using MSPI (C^CGG) ​enrich for CpG island = 1/10 of the whole genome price
Methylation is quite similar between tissues and most of the genome is methylated