Genomics Flashcards
1
Q
Gene expression
A
- Exons code for proteins
- Not all genes are active at the same time → diversity in cells →ensured by cell type specific gene expression
- Gene expressed = the RNA transcribed from the gene is actually produced
2
Q
Measuring RNA production: qPCR
A
- rtPCR (retro-transcribed) –oldest method
- quantitative PCR allows gene of interest to understand how much cDNA is present in cells
- RNA cannot be directly measured by PCR → synthesise cDNA → complementary to RNA → reaction w/reverse transcriptase enzyme
- cDNA only includes exons because RNA is spliced
- cDNA made with specific fluorophore incorporated into RNA
- When Taq polymerase is completing second strand → fluorophore released→ qPCR quantifies how much fluorophore in reaction → more light = more RNA = more gene expression & earlier signal showing up
3
Q
qPCR Normalisation
A
- Results need to be normalised to measure actual change in transcription level → need to compensate for initial variations in mRNA and technical differences w/sequencing
- Housekeeping genes included in qPCR → have stable expression → important for cell components → always expressed above certain threshold
4
Q
qPCR Limitations
A
- Quick, relatively accurate, cheap but…
- Limited as to how many genes can be tested at any one time ~5-10 (not possible for 1000’s genes)
5
Q
Microarray
A
- Revolutionised gene expression analysis
- Allows for detection/comparison of thousands of genes simultaneously
- Relies on base-pairing hybridization with probes for each gene to be measured
- More expensive than PCR, but still relatively cheap
- Can measure:
o Differing expression of genes over time, between tissues and different states
o Co-expression of genes
o Identification of complex genetic diseases
6
Q
Affy Gene Chip & Array Experiment
A
- Each gene has 16-20 pairs of probes synthesized on the chip
1. RNA extraction
2. Make cDNA using biotin (important for binding to streptavidin which is probe-specific)
3. If binding between gene of interest and probe chip releases light quantify it
7
Q
Affy Expression Measurements
A
- A = absent
- M = marginal
- P = present (P-value gives confidence)
8
Q
Microarray Limitations
A
- Data is very noisy
- Probes not available for all genes; Affy probes only for ~75-80% of human genes
- Cannot detect genes w/very low expression levels
- Data requires lots of statistics and analysis
- Assay does not distinguish expression from different isoforms of the same gene
9
Q
Next generation sequencing
A
- Based on getting fragments out of a genome →adding sequences (adaptors) at the edges →adaptors always have same sequences →adaptors bind on flow cell and bend fragment then second adaptor binds → bridge-like formation
- Flow cell amplifies PCR → fragments w/adaptor still bind on sequencing machine to get sequenced →have millions of fragments ~75-150 base pairs → map onto reference genome
- “Seq” principle
10
Q
RNA-seq
A
- Uses next generation sequencing to measure gene expression
- Can assume that every mRNA present will be sequenced the same number of times
- If experiment shows 2x mRNA for particular gene as control, then gene expression is 2x greater
- Gives accurate measure of gene expression, even for genes w/v. low expression levels
- Can identify exact transcript being expressed
- Can potentially identify unknown transcripts with novel splice sites
- Method:
1. Extract all mRNA → convert to cDNA fragments
2. Add sequencing adaptors → obtain short sequence using high-throughput sequencing
3. Resulting sequence reads aligned w/reference genome or transcriptome
4. Base count profile for each gene is created - Same procedure for control and variant
- Read counts are proportional to gene expression level
11
Q
RNA-Seq Normalisation
A
- Important to normalise:
o Sequencing depth = how many reads are sequenced by the machine
o Length when dealing w/ different organisms (e.g. human vs mouse)
o Amount of fragments in each sample - 2 main methods to normalise data:
o Raw read count normalisation
o Reads/fragments per KiloBase per Million reads (RPKM -single end reads; FPKM – paired end reads)
RPKM = 109C / N L
C = raw count of reads in transcript
N = number of mappable reads in experiment
L = transcript length (bp)
Normalizes for gene length (C and L) and library size (N)
12
Q
Raw Read Count Normalisation –DESeq2
A
- Aims to make normalized counts for non-differentially expressed genes similar between samples
- Does not aim to adjust count distributions between samples
- Assume that:
o Most genes are not differentially expressed
o Differentially expressed genes divided equally between up and down - Relies on housekeeping genes
- Normalisation looks for set of important, highly expressed genes → assume that expression is uniform across samples → same shift for housekeeping genes performed on all genes
13
Q
RNA-Seq Limitation
A
- Cofounded by heterogeneity of the sample:
o Different cell types
o Mutations
o Different cell cycle stage
o Epigenetic modifications
o Stochastic gene expression
14
Q
Single Cell RNA-Seq
A
- Allows analysis of single cells
o Enables improvement in resolution of gene expression within samples
o Enables identification of heterogeneity in cell populations i.e. different cell types
o Enables gene expression within single cells/cell types to be categorised - Tissue → dissociation of cells → isolation of cells → single cell → RNA extraction → cDNA synthesis → single-cell sequencing → expression profile → cell type identification
- Plots show differing gene expression in cell types clearly distinguished; even cells difficult to separated (e.g. podocytes) are effectively dissociated
- Results can be illustrated by heat maps or using dimension reduction analysis tools such as PCA or t-SNE
15
Q
Future
A
Profile gene expression in vivo w/o need of isolating cells
16
Q
DNA Methylation
A
- Reversible
- Symmetrical so maintained thorugh cell division
- Adding methyl group (CH3) to 5’C of cytosine by methyltransferases
- In mammals, mainly occurs at CpG sites – CpG islands
- CpG islans used for identification of potential promoter regions
- Methylation of CpG island = silencing of gene expression
- Represses gene expression by:
o Preventing binding of transcription factors
o Modifies chromatin structure to repress transcription - Methylation is major factor in epigenetic modifications
- Methyltransferases in mammals: DNMT3a and 3b
- During mitosis, hemi-methylated DNA is created → copied strand is unmethylated → recognised by DNMT1 →methylates new strand to maintain methylation state
- Methylation of histone → chromatin repressed → cannot be transcribed
17
Q
DNA methylation and disease
A
- Methylation patterns in disease tissue ≠ from normal tissue → aids in identification of disease-causing genes
o Specially in cancer and neurodegeneration → disease correlates with loss of methylation
o E.g Alzheimer’s disease (NEP gene); Colorectal cancer (MGMT gene); breast cancer (PRLR) - Abnormal methylation silences tumour suppressor genes
18
Q
Where does methylation occur
A
- Intergenic regions = usually methylated
o Maintains genomic integrity
o Methylated DNA forms compacted chromatin → less accessible for recombination and translocation
o DNMT1 deficient cells display genomic instability - Repetitive elements = usually methylated
o Transposable elements are highly mutagenic if they can transpose within genome → methylation protects genome from TEs
o Methylated C mutates to T over evolutionary time → prevent transposition
o Methylation prevents recombination - Gene upstream regions = usually unmethylated
- Promoter regions = usually unmethylated so create CpG islands
- Lack of methylation creates relatively higher density of CpG due to lower rate of mutation to T
19
Q
Avoiding methylation
A
- When region is methylated at all times, DNA tries to find evolutionary solutions for methylation to be avoided → not having cytosine anymore → tend to mutate to T overtime
- If transposons accumulates mutation it loses functionality; 50% of human genome is made of transposons → 99% of them have lost their ability to be a parasite → cannot move anymore
20
Q
Identifying DNA-Methylation
A
MeDIP-Seq
* Antibody recognizes methylated cytosine → binds meth DNA → immunoprecipitation → retain only antibody bound DNA →fragmented → next gen sequencing → sequences mapped back onto genome to identify methylated regions
Bisulphite sequencing
* Samples treated bisulphite → converts unmethylated C to U → sequence and compare samples to determine methylation e.g. cancer vs normal cells
* PCR only able to amplify U-containing DNA (non-methylated); with other primers can amplify all fragments that contain methylated DNA
Both
* Expensive
* Great resolution
* MeDIP-Seq requires antibody
21
Q
X inactivation
A
- The silencing of one of the X chromosomes in all female mammals
- Required for dosage compensation to avoid over expression of genes on X chromosome
- Inactivated X chromosome packaged as compacted heterochromatin
o Compaction by chromosome wide histone methylation –H3K27M3 - Inactivation by Xist gene (long non-coding RNA)
22
Q
Long non-coding RNA sequences
A
- Longer than 200 nucleotides
- Thousands identified but function largely unknown
o Target different aspects of gene transcription mechanism
o Can function as co-regulators or transcription factors - Act in ‘cis’ (same chromosome they are transcribed from) or ‘trans’ (different chromosome)
- ncRNA Evf-2 = a co-activator for homeobox transcription factor Dlx2, involved in forebrain development and neurogenesis
23
Q
Xist
A
- 17kb long; acts in cis
- Expressed from only one of 2 X chromosomes first detectable event in X inactivation
- Xist contains many repeats → 6 identified so far
- Repeat A (RepA) silences function of Xist → binds to PRC2 (Polycomb repressive complex –a histone methyltransferase complex) → lays down histone methylation along chromosome at Lys27
24
Q
HOTAIR
A
- Long ncRNA expressed from HOXC locus on chromosome 12 → represses HOXD on chrom.2
- ‘HOX’ = important developmental genes
- Acts in trans
- Binds to PRC2 and LSD1 → PCR2 adds repressive H3K27me → LSD1 removes active H3K4me → combined function produce repressive chromatin structure
- In cancer, HOTAIR acts on regions other than HOXD
25
Transcription factors
* Up- or down-regulate a gene
* Recognise specific DNA motifs ~5,6,7 sequences
* Use several mechanisms to regulate gene expression:
o Stabilising or blocking the binding of RNA polymerase to DNA
o Recruit coactivator or corepressor proteins to the transcription factor DNA complex
o Catalyse acetylation/deacetylation of histone proteins:
i) Histone acyltransferase (HAT) activity –weakens association of DNA w/histones making DNA more accessible to transcription (up-reg)
ii) Histone deacetylase (HDAC) activity –strengthens association of DNA w/histones making DNA less accessible to transcription (down-reg)
* ~95% of Tf can only bind motifs when chromatin is in active methylated state (non-condensed)
26
Variability of TF
* Actual sequence of particular Tf binding site →variable → harder to identify → so described by general motif not fixed sequence
* E.g. Neurod family recognise small sequences; HOX recognise dimers
* P53 recognises dimers → most important tumour suppressor gene in mammalian cells → if mutation → cancer
* Very specific motifs; even single mismatch would affect function
27
Pioneer transcription factors
* Able to go to region of inactive repressed chromatin → bind it → signal for active machinery to arrive
28
Experimental Method (Tf) –ChIP-Seq
* Chromatin immunoprecipitation followed by sequencing
* Used to identify binding regions if binding protein is known
* Based on antibodies
* ChIP-Seq directly sequences the bound DNA → can then be mapped back onto genome for precise localization
* Consensus/variability of binding sites can be determined from sequence:
o Map reads back to reference genome
o Most frequently sequenced fragments form coverage peaks at specific locations
* Same approach as used for DNA methylation
29
ChIP-Seq Method
1. Chromatin in nucleus is cross-linked (Tf cannot detach) → fragment it
2. DNA fragments include those w/target protein bound → incubate w/antibody specific for Tf of interest
3. Immune precipitation → everything bound to Tf will precipitate (normally use magnetic beads that recognise the antibody to keep fragments bound to Tf stuck in tube)
4. Reverse cross-linking
5. Use specific proteases to degrade Tf → have DNA only
6. Sequence using next-gen sequencing
7. Mapped back onto genome → reads only map in regions where Tf was binding
30
Identifying histone modifications
* Also use ChIP-Seq
* Fractionate DNA → use antibody that binds to modified histone being studied → separated by immunoprecipitation →sequence isolated fraction using next gen seq → map sequence back onto genome to identify regions w/modified histones ie. genes under regulation
31
ENCODE
* Encyclopaedia of DNA Elements
* 400 scientists involved
* 5-year project completed in 2012 → identify all functional elements in human genome sequence
o Goal: to identify and characterise everything in the genome that is non-coding →the DNA/RNA regions regulated and the factors regulating them
* 2003- human genome was sequenced → now have a reference genome allowing development of this project
32
What they did in ENCODE
* Looked at how many different genes are expressed in different cell types
* Profiled binding of Tf across ~56 different cell types using ChIP-Seq
* Profile DNA methylation across ~56 different cell types
* Looked at chromatin conformation
o Promoters & enhancers need to be in contact w/each other → contacts can be profiled by different techniques → look at 3D genome & chromatin
* Profile all accessible regions of chromatin (euchromatin state)
o Euchromatin = less condensed, gene-rich, more easily transcribed; nucleosomes are depleted; DNA is accessible for binding of Tf
o Heterochromatin = highly condensed, gene poor, transcriptionally silent
33
Complexity of gene expression
* Regulating gene expression → complex → involves many regulatory factors → Tf, enhancers, silencers, methylation patterns
* Transcripts regulated by splicing factors e.g. exon and intron splicing enhancers and silencers
* Functionality often cell/tissue and time specific
34
DNA Hypersensitive Sites (HS)
* Hypersensitive sites = regions of chromatin highly sensitive to DNase 1
o Nucleose→ less compact→ enables DNA to bind to proteins e.g. Tf
* DNase-Seq = technique where DNase 1 cuts DNA only when accessible → isolating fragments → amplify through next gen seq → add adaptors w/ligation → cluster into flow cell→ cluster → reads mapped onto genome → they pile up in regions where chromatin was open and accessible
* Mapping HS sites → identify location of genetic regulatory elements e.g. promoters, enhancers, silencers, locus control regions →(ChIP-Seq can only identify Tfs)
35
DNase-Seq Overview
* Open chromatin regions sequenced and mapped to reference genome
* Nuclear extraction → DNase 1 digestion → library preparation → PCR amplification → high-throughput sequencing
36
DNase-Seq Footprinting
* Number of fragments that map to a sequence is a measure of regulatory activity
* Sites bound by some Tfs show highly specific patterns of DNase I cleavage = 'DNase footprints’
* Footprints used to identify binding of specific Tfs; advantage over ChIP-seq
o w/ChIP_seq need to know the transcription factor for immunoprecipitation
o DNase-seq identifies the binding sites de novo
* DNAse-seq is a genome wide version of DNA footprinting method
* Footprint = prediction of what Tf could bind
37
FAIRE-Seq
* Similar method to DNase-Seq + addition of formaldehyde for cross-linking
o More efficient in nucleosome-bound DNA than in nucleosome-depleted regions of genome
* Phenol chloroform extraction of DNA → treatment to isolate nucleic acid from solutions
o Cross-linked chromatin will go to bottom of tube (organic phase)
o Condensed chromatin at bottom; active chromatin at top of tube
o These increase resolution and reduce noise
* DNA extracted and mapped to reference genome to identify open DNA regions
* FAIRE-Seq higher coverage at enhancer regions over promoter regions
* DNase-Seq higher sensitivity towards promoter regions
38
ATAC-Seq
* Uses mutated hyperactive transposase Tn5 instead of DNase1
* Tn5 enzyme derived from transposons → attacks and chews open active euchromatin efficienctly
* DNA fragments then isolated, sequenced, and mapped
* Advantages:
o Requires smaller sample than DNase-seq and FAIRE-seq (requires 1000x more cells)
o V. fast → completed in 3 hours
* Disadvantages:
o V. expensive because monopoly
39
MNase-Seq
* Uses micrococcal nuclease
* Cuts v. near to nucleosomes
40
ChIP-Seq vs all others
* ChIP-seq requires antibody
o Profiling something specific e.g. one Tf of interest
* All others don’t need antibody
o Less specific → profile all active chromatin regions
41
Chromatin Interaction
* Within each chromosome TAD (Transcriptional active domain)- portions of chromosome isolated from each other
o One TAD does not interact w/other TADs → enhancers/promoters can only regulate genes in same TAD (v. few expection)
* Lowest possible level = looping
* Knowing interactions is essential for understanding mechanisms of gene regulation in health and disease
* All possible interactions can be profiled by different techniques (e.g ChIA-PET, 3C, 4C, HiSEQ, etc)
o All based on same approach w/one variation
o Use restriction enzymes: fragmenting genome → religation→ profile what is ligated
* ChIA-PET → studies genome wide long range chromatin interactions involving protein factors
o Involves additional step: antibody precipitation → to identify chromatin interactions that are regulated by a specific transcription factor, between distal and proximal regulatory sites and their associated promoters
* Diff-linker, PETs, used to identify non-specific ligation noise; identify ligations between different ChIP complexes
42
RIP-seq and CLIP-seq
* V. accurate; can be reproduced
* Methods similar to DNA-protein interaction identification
* RIP-seq involves immunoprecipitation of RNA-binding protein (RBP) of interest → has to be done non-stringently
o Low stringency = low specificity
* Developed to CLIP-seq → includes cross-linking step using UV light (irreversible)
* Final steps:
o Digestion with proteinase K leaving peptide at binding site that modifies nucleotides to create cross-linked induced mutation sites (CIMS)
o Reverse transcription to make cDNA → identify RNA-binding sites
o Sequenced and mapped to transcript
43
CLIP-seq variants
* PAR-CLIP = improves crosslinking w/photoreactive RNA nucleotides
* iCLIP = uses reverse transcriptase stalling to map individual nucleotide-protein interactions
* miCLIP = modifies RNA methylase to map its binding sites
44
-seq key features
* All ‘seq’ methods use similar approaches and same final step to identify regions of interest (protein binding site, open chromatin, etc)
* Method:
o Isolate sequence e.g. fragmentation and immunoprecipitation, phenol/chloroform etc
o Sequence fragment and map back to genome
45
ENCODE controversy
* Most of the genome is "functional" → controversial statement from ENCODE
o ENCODE considers anything transcribed must be functional → but many transcripts are non-functional e.g. pseudogenes
* ENCODE emphasized sensitivity over specificity → lead to false positives
* Criticism: arbitrary choice of cell lines and transcription factors; lack of appropriate control experiments
46
ENCODE limitations
* Conducted in immortalised cell lines (derived from human cancers → v. easy to manipulate but v. unstable)
o Want something closer to real healthy cells
o E.g HeLa cells sometimes have 3/4 pairs of chromosomes → not diploid
* Roadmap consortium
o Profiled histone modifications across 25 human primary tissues (mark different chromatin states)
47
What is a SNP
* DNA sequence variations → occur when single nucleotide (A, T, C, or G) in genome sequence is altered; must occur in at least 1% of the population
* SNPs make ~90% of all human genetic variation; occur approx every ~1000 bases; ~4-5 million SNPs in an individual human genome
48
Why are SNPs important
* Can affect how humans develop diseases
* Can affect how an individual respond to pathogens
* Can affect how an individual respond to chemicals
* Can affect how an individual respond to drugs, etc
* Potentially their greatest importance in biomedical research is for comparing regions of the genome between cohorts
* Comparing cohorts with and without a disease – GWAS
49
SNP location
* Intergenic region → possibly transcription enhancer/regulatory region
* Within promoter or transcription factor binding region
* Within exon → Could affect protein coding
* Within intron → Possibly regulatory region e.g. affecting splicing
50
Disease SNPs
* Can be v. dangerous or neutral
* SNPs may be direct cause of disease or signal for increased likelihood of disease
* Disease associated SNPs:
o Monogenic → one nucleotide change leads to disease; relatively easy to detect/analyze; simple traits
o Polygenic → many nucleotide changes affect probability of disease; hard to detect/analyze; complex traits
51
Coding SNPs
* Coding SNPs = potentially disease causing as they can affect the protein
* Types: Synonymous (silent) & Non- synonymous
* Synonymous mutation: change base but AA is the same
o May still affect Exon Splicing Enhancers (ESE) or Exon Splicing Silencers (ESS) site so cannot always be ignored
* Non-synonymous – change in base changes AA → mutation could be detrimental
52
Transition and transversion
* Transition (Ti) - most common substitution
o Replacing purine by purine i.e. A → G or pyrimidine by pyrimidine i.e. T → C
* Transversion (Tv) - less common
o Replacing purine by pyrimidine or vice versa i.e. A → C
* Ti/Tv ratio – varies within genome; used to assess GWAS data quality
o Across entire genome averages around 2
o In protein coding regions typically higher, often above 3 due to transversions in third base of codon being more likely to change the encoded amino acid
53
Sickle Cell Anaemia
* Inherited blood disorder due to mutations in beta globin HBB
* Found primarily in African and related populations
* Fragile, sickle-shaped cells deliver less oxygen to the body's tissues
* Get stuck more easily in small blood vessels; break into pieces that interrupt healthy blood flow
* Symptoms: shortness of breath; infections (bone, gall bladder etc); joint pain
* Causes:
o Mutation of β-globin gene at AA position 6 (HbS) - GAG → GTG: Glutamic acid → Valine
o Only individuals homozygous for allele (T:T genotype) have sickle cell anaemia
o Autosomal recessive mutation
54
Alzheimer’s Disease
* Early onset familial
o Hereditary; ~40yo
o V. rare ~5% of all cases
o Caused by mutation in amyloid precursor protein (APP) or presenilin-1 (PS1)
* Sporadic late onset
o ~70yo
o Associated w/many genes → e.g. Alipoprotein E (ApoE)
* ApoE contains 2 SNPs resulting in 3 possible alleles for the gene: E2, E3, E4
o Protein product of each gene differs by one amino acid
o E3 no effect regarding Alzheimer’s
o 1x E4 allele → greater chance of developing Alzheimer's
o 1x E2 allele → person is less likely to develop Alzheimer's
o 2x E4 alleles → may never develop Alzheimer's
o 2x E2 alleles → may develop Alzheimer's
55
Non-coding SNPs
* Studies report: disease associated SNPs are enriched in regulatory DNA regions
o Enhancers (~+1 million enhancers in human genome)
o Silencers
o Locus control regions
o Promoters
o Long non-coding RNAs maintaining higher order structure of 3D genome
* 98% of T2 diabetes associated SNPs were non-coding
* Changing 1 nucleotide in motif → Tf does not bind anymore → enhancer is not activated
56
SNPs disrupt splice sites
* Causes ~10% of all mutations causing human inherited disease
* Splice sites found in proximity to exon → provide signal for proteins to cut RNA
* If SNP in splice site → cannot cut anymore
* SNP most likely causes total loss of associated exon; or introduces cryptic splice site
* OAS1 gene –associated w/T1 diabetes
* Can have synonymous mutation in coding DNA but affects splicing machinery
57
Insertion or deletion
* Can cause:
o Disrupted start codon
o Disrupted stop codon
o Disrupted splice site
o Frame shift
* In a frameshift mutation, base is inserted/deleted, altering codon in which insertion or deletion took place, but also changing the reading frame so that all codons downstream are read out of frame → produces string of amino acid substitutions before a stop codon is reached (stop codons are frequent in coding sequences read out of frame)
58
Genome Wide Association Studies (GWAS)
* GWAS = If 500 people with the same disease all share a half dozen SNPs in common, but a group of 500 healthy people don’t share those SNPs, the mutations behind the disease is probably around those SNPs (now more like 10,000 people)
o GWAS has identified genetic variations that contribute to risk of: T2 diabetes, Parkinson's, Heart disorders, Obesity, Crohn's disease, Prostate cancer…
* GWAS is a very strong area of research → look at common SNPs → statistics
* The difficulty is accurately identifying the SNPs
59
100,000 Genomes Project
* Genomics England project in collaboration w/NHS
* Aim: to sequence the genomes from approximately 70,000 people
o Longer term aim → research on new & more effective treatments.
* Participants are NHS patients w/cancer or rare disease
* Genomes of families of patients w/rare diseases also sequenced → identify variants associated w/different conditions
* Objective: to create a new genomic medicine service for the NHS
* Patients may be offered a diagnosis where this wasn’t possible beforelonger term aim is research on new and more effective treatments.
60
Identifying Variants
* Sequence multiple genomes from a population at low coverage and pool the data
* Align to reference and identify variants
* Pooling works as most of the genome will be the same; some individuals will also share variants
* Variant prediction software identifies which variants are real and which sequencing errors
61
Limitations of GWAS
* Mostly white people genomes (UK or US) → no diversity → no representative pool
* Not easy to do functional validation
o SNP may have effect in one cell type not in the other
* Coding mutations alter AA sequence of protein → effect is clear
o Non-coding mutations disrupt regulatory elements are less clear
* Regulation of gene expression is dependent on multiple factors:
o Cell-type
o Temporal patterns such as circadian clock
o Cell-tissue development
* GWAS identifies candidate SNPs but confirmation requires additional work
* Biggest limitation: linkage disequilibrium and GWAS
o Linkage disequilibrium = the association of alleles at two or more loci within a population → haplotypes don’t occur at expected frequencies - not random
o Can be used to improve genetic association studies, such as cancer
o Enables identification of genetic markers for the associated disease
o E..g 6 SNPs found in people w/Alzheimer’s → 3 found in linkage disequilibrium → cannot figure out which SNP is causative of disease
62
Expression Quantitative Trait Loci (eQTL)
* First DNA sequencing → then RNA sequencing to quantify gene expression → involves mapping variants which alter gene expression
* eQTL = non-coding SNPs known to affect expression of specific gene ; variants associate w/RNA levels
* eQTL mapping enables identification of regulated genes → unlikely to be close to disease associated SNP
* Cis eQTL = affect expression of nearby gene
* Trans = does not map close to gene; could be other chromosome
63
GWAS and eQTL
* 1000 people → get DNA → find SNPs → get RNA → quantify gene expression → leads to identification of disease causing genes
64
SNP genotyping
* To only find out if pre-defined SNPs are present:
o Microarrays w/probes for specific SNP → DNA only binds to probe if there is SNP
65
Epigenetics
* Epigenetics = ‘above’ genetics → external modifications to chromatin that turn genes on/off
- Modifications do not change DNA sequence → they affect how cells read genes
* Epigenetic changes alter physical structure of DNA
- E.g. DNA and histone methylation
* Epigenetic modifications can be inherited → “An epigenetic system should be heritable, self-perpetuating, and reversible (Bonasio et al. - Science 29 October 2010: 612-616)”
* Depends on environment, diet, smoking
66
Nucleosomes
* Genome is condensed and compacted into nucleosomes
* 146bp of DNA wrapped around histone octamer (8 → 2x H2A, H2B, H3, H4)
* Space between nucleosomes = linker DNA
* Nucleosomes disassemble then reform during replication
* Nucleosomes = repeating units of chromatin
* Interaction DNA-histones = sequence independent (H-bonding & ionic interactions w/sugar-phosphate backbone)
67
H1
* Sits outside each nucleosome
* Structural function to keep nucleosome together
68
Histones
* Can be methylated and acetylated (the tail of the spheres) histone modifications
* Covalently modified
69
Euchromatin vs heterochromatin
70
2 main ways to repress chromatin
1. Constitutive heterochromatin
- H3K9me2/me3
- H3 = histone; K9 = lysine 9; di- or tri- methylation
- Older → weaker methylation of histones → heterochromatic regions get aberrantly activated e.g in Alzheimer’s
- Permament
2. Facultative heterochromatin
- Regions can turn on/off when necessary (e.g. gene promoters that don’t need to be active at all times i.e. developmental genes)
- H3K27me2/me3 → methylation of Lys27 by H3, tri- or di- methylation
- Non-permanent → deposited by PRC2
* Need heterochromatin → otherwise DNA too big
* Histone modifications repress transposons
71
PCR2
* EZH2 = catalytic subunit (enzyme)
* Other subunits in complex: EED + SUZ12
72
Typical signature of active enhancers
* Always acetylated in Lys27 of H3; 1x methylated in Lys4
* Sometimes methylation is a marker of activity
* Signature used by Tf to understand where to go; if chromatin is active or not
* Active promoters have trimethylation of H3 Lys4
73
Acetylation and methylation enzymes
* Histone acetylases and deacetylases
* Methyltransferases and demethylases
* If these are impaired → histone code is aberrant → have inappropriate labels → e.g. gene that should be repressed is labelled w/active markers → leads to disease e.g cancer
74
Epigenetics and Cancer
* Epigenetic change that silences tumour suppressor gene → lead to uncontrolled cellular growth
* Turn off genes that help repair damaged DNA → lead to increase in DNA damage → cancer risk
* Prostate cancer associated w/gene silencing by CpG island hypermethylation within promoter region of GSTP1 gene
* If issue upstream in epigenetics → leads to cascade of problems
75
X inactivation
* Example of epigenetics
* Marsupials: paternal X chromosome always silenced
* Tortoiseshell cat:
o All female
o Black/orange alleles of fur coloration are in X chromosome
o If heterozygous →resulting colour depends on which X is inactivated
o Tortoiseshell pattern (phenotype) determined by X inactivation
76
Agouti Mice
* Genetically identical mice
* Mother 1 →skinny-brown mouse→ methyl-rich diet → methylated agouti gene repressed
* Mother 2 →obese, yellow mouse prone to diabetes/cancer→ unmeth. agouti gene expressed
* Agouti gene common to all mammals
77
Epigenetics and Twins
* Identical twins = identical genes → differences due to epigenetic changes → lead to different disease susceptibility for example
* Label different histone modifications w/fluorescent probes → chromosome pairs in twins digitally superimposed →one tag red other green overlapping shown as yellow
78
Difficulty identifying inheritance
* Histone modifications can be inherited environmental factors impact it (e.g. smoking mother)
* Changes in epigenome inherited due to mechanisms (still under study) that allow cells of offspring to remember epigenome of parents
* Complicated to understand where inheritance is coming from → always present in family? → something environmentally driven?
* Need to show that epigenetic effect can pass through enough generations to rule out possibility of direct exposure → potentially 3 generations at once exposed to same environmental conditions → prove epigenetic inheritance requires epigenetic change in 4th generation
* WW2 in Netherlands → extreme lack of food diet poor in methyl groups → kids born now show these effects → e.g. hypomethylation of IGF2 (insulin-like growth factor) involved in diabetes and cardiovascular diseases
79
Erasing of methylation
* DNA of human sperm is highly methylated; eggs less
* Egg fertilized → methylation/acetylation in chromatin largely erased esp. from paternal genome
* As embryo develops, methylation marks continue to be lost from maternal genome up to blastocyst stage
* Not all methylation erased, hence inheritance possible
* After this → cells differentiate → DNA becomes methylated again for specialized cell types
80
How is methylation inherited
* Methylation patterns normally erased in primordial germ cells
* Methylation marks converted to hydroxymethylation then progressively diluted as cells divide
o V. efficient mechanism → resets methylation patterns of genes for each generation
* Research found: some rare methylation can escape this reprogramming process → can be passed on to offspring → enables inheritance of epigenetic traits
81
Imprinting
* Most genes inherit 2 working alleles from mother & father
* Imprinted genes inherit only one working copy → other is epigenetically silenced by addition of methyl groups → repressed forever but reset during egg/sperm formation
* Expression of gene comes from only 1 allele → less protein or less variation
* Almost all imprinted genes associated w/growth
82
Genetic conflict hypothesis
* Theory of why imprinting happens
* Males multiple offspring w/multiple partners simultaneously at low cost
* Females only one set of offspring at a time at high cost of personal resources
* Many imprinted genes involved in growth and metabolism
* Paternal imprinting favours production of larger offspring; maternal imprinting favours smaller offspring
83
Epigenetics in plants
* Plants depend on epigenetic processes for proper function
* Flowering controlled by set of genes affected by environmental conditions through alteration in expression pattern → ensures production of flowers even when plants are growing under adverse conditions
* Epigenetic modifications include DNA methylation, histone modifications, and production of micro RNAs (miRNAs)
