Week 1
Lecture Epigenomic technologies: when what why when
Quantitative measurement methods of DNA and RNA in a genome-wide manner
qPCR arrays: can measure the presence of DNA or cDNA in a quantitative fashion. The
cycle number which is required to reach the logarithmic phase of amplification is highly
correlated to the amount of starting material. If you combine that with standards, you can
make a standard curve and determine the amount of DNA in the sample.
With the new array plate you can combine 96 samples with 96 primer pairs, so you can run
9216 reactions in one go. This is not genome wide, but you can study a set of markers for
example.
Array hybridisation: identifies losses or gains of DNA across the whole genome. Have
some dots on a platform, and the DNA
fragment can be hybridized with DNA from two
or multiple different samples. Since those are
fluorescently labeled, you can quantitatively
compare those two transcripts/samples.
Yellow = equal contribution
Red = one sample
Green = other sample
→ need to know your genome prior to the
experiment (relative measurement)
Tiling array: entering the true genomic area
(microarray). Instead of having one probe for each cDNA, from which you know if the gene
or cDNA is expressed or not, later on people used this microarray. The microarray has many
probes distributed in an evenly spaced way across an entire genome.
→ Can see alternative splicing for example, so if a gene is excluded from the transcript
Next generation sequencing: top technology nowadays, to define the abundance of DNA
fragments on a genome wide scale.
Two important parameters for these sequencing techniques:
Read length: how long the read can be. The mapping accuracy can be increased with this
parameter.
Throughput in gigabytes or number of reads: important to know the number of reads to
know how many times you cover the read (quantitative).
Longer fragments are better, otherwise if you fragmentise the DNA in small fragments it is
hard to find out the order. When you don’t know how the genome/DNA looks, longer
fragments make it easier = de novo assembly (don’t know what the genome looks like and
want to know it).
1
,Single molecule real-time sequencing: Oxford Nanopores
Nanopore is made of protein, which in nature forms holes in membranes created by
synthetic polymers. This membrane has high electronic resistance. When the single
molecules enter through the nanopore, the structure and current is disrupting, which helps
identifying the molecule by sequencing.
Nanopore strand sequencing: an intact polymer is sequenced when it enters the nanopore,
when a strand of DNA enters through the nanopore it can read base per base. The enzyme
in the nanopore makes the dsDNA to ssDNA so it can go through the nanopore. By
disrupting the current it can determine the order of bases on the DNA strand.
Big advantage of this technology compared to the other technologies: it can recognize
modified base pairs, for example when DNA methylation occurs, because this gives a
different signal. Have a hairpin in the end of the long read strand so then it will also read the
other strand. This is typically used for long read length sequences.
Practice questions
Which method can you use to measure the relative abundance of a few dozen RNA markers
in a cohort of 600 patients?
QPCR with the array plate of the 96 samples, with 6 of those plates you can check all of
these patients.
You would like to identify the exact position of the transcription start sites genome wide.
Which method is best suited for this purpose and why?
RNA sequencing and cap sequencing. You pull intact RNA sequences which contain the
cap, of this you remove the cap and sequence the 5’ end of this sequence.
Another option is tiling arrays, but not to base pairs, so RNA sequencing would be the best
option.
You are working with an intracellular parasite that resides in the liver and interested to see
which parasite genes are upregulated when the host has fever. Knowing that it is very
difficult to isolate RNA from these parasites without substantial contamination from the host,
which method is the most economical solution to your problem? Why?
?
Examples of epigenetic features in the genome: DNA modifications, histone modifications
and variants (acetylation, methylation, phosphorylation etc.), nucleosome occupancy,
chromatin interactions and chromatin domains, and RNA modifications and non-coding
RNAs.
If you have to choose 6-8 modifications to study, which ones are the most interesting?
- DNA methylation: H3K27 methylation is important because it leads to
heterochromatin formation and repression of transcription.
- H3K9 methylation
- K27 acetylation, which is good for enhancers.
2
,What are the most important epigenetic features to profile?
- H3K4me3 and H2AZ you can find
where promoters are.
- H3K36me3 and H3K79me2 you can
know where active coding sequences
are.
- H3K4me1 and H3K27ac you know
where enhancers are.
- CTCF is a structural protein.
- To know where the inactive genes are,
look for H3K9me2 and H3K9me3
methylation.
- H3K27me3 is involved in
heterochromatin silencing.
Measuring protein-DNA interactions to analyze the localisation of epigenetic features (e.g.
histone modifications, variants, reader proteins, chromatin remodelers) and transcription
factors
Chromatin immunoprecipitation sequencing (ChIP-seq): used to study all sorts of
protein-DNA interactions and where the protein of interest binds to the DNA.
This method uses formaldehyde to crosslink the DNA and protein interactions. With
sonication you chop the DNA into smaller fragments and perform immunoprecipitation (uses
antibodies to purify) to study the individual fragments from which the transcription factor has
bound to the DNA. The binding caused by formaldehyde needs to be reversed so you only
keep the DNA to be isolated. This sequence will tell us which DNA sequences are bound by
TF in vivo for example.
Amplify the DNA to get enough for sequencing techniques and then you can study them by
microarray or sequencing. When you get the sequence read from this assay, you map
them to the reference genome, in other words: the number of reads mapped to a certain
loci is proportional to the abundance of the mark.
The peaks as a result from this assay in the genome browser show a binding site for the
transcription factor. Can analyze these results further with these examples:
3
, DNA adenine methyltransferase (DamID) as alternative to ChIP
ChIP relies on immunoprecipitation using an antibody to target the protein of interest and
sequencing of the co-precipitated DNA fragments. The primary advantage of DamID is that
there is no requirement for antibodies or other affinity reagents to purify the protein of
interest.
Drawback of CHIP is: you always lose a lot of material because you wash all the unbinding
material away and it is temporary. So if the binding is every 5 minutes it might be hard to
capture this precisely with ChiP-seq.
DamID is more sensitive and it involves the identification of DNA sequences carrying the
methylation mark that is deposited by the bacterial methylase dam, which is expressed as
a fusion to the protein of interest.
DNA gets isolated and the PCR adaptors get ligated so
that the methylated fragments can be amplified and
sequenced. This shows just as ChiP-seq an binding
profile between DNA and proteins.
DamID identifies binding sites by expressing
DNA-binding protein as fusion protein with DNA
MTFase.
DamID is based on the expression of a fusion protein
consisting of a protein of interest and DNA adenine
methyltransferase. Binding of POI to DNA, localizes the
methyltransferase in the region of the binding site. This
leads to methylation of adenines near sites where the
protein of interest interacts with the DNA. Adenine
methylation doesn’t occur naturally in eukaryotes, so
therefore adenine methylation in any region concludes
that the adenine methylation had been caused by the
fusion protein. This implies that the region is located
near a binding site. These methylated sequences are
subsequently amplified by a methylation-specific PCR
protocol and identified by hybridization to microarrays.
4
Lecture Epigenomic technologies: when what why when
Quantitative measurement methods of DNA and RNA in a genome-wide manner
qPCR arrays: can measure the presence of DNA or cDNA in a quantitative fashion. The
cycle number which is required to reach the logarithmic phase of amplification is highly
correlated to the amount of starting material. If you combine that with standards, you can
make a standard curve and determine the amount of DNA in the sample.
With the new array plate you can combine 96 samples with 96 primer pairs, so you can run
9216 reactions in one go. This is not genome wide, but you can study a set of markers for
example.
Array hybridisation: identifies losses or gains of DNA across the whole genome. Have
some dots on a platform, and the DNA
fragment can be hybridized with DNA from two
or multiple different samples. Since those are
fluorescently labeled, you can quantitatively
compare those two transcripts/samples.
Yellow = equal contribution
Red = one sample
Green = other sample
→ need to know your genome prior to the
experiment (relative measurement)
Tiling array: entering the true genomic area
(microarray). Instead of having one probe for each cDNA, from which you know if the gene
or cDNA is expressed or not, later on people used this microarray. The microarray has many
probes distributed in an evenly spaced way across an entire genome.
→ Can see alternative splicing for example, so if a gene is excluded from the transcript
Next generation sequencing: top technology nowadays, to define the abundance of DNA
fragments on a genome wide scale.
Two important parameters for these sequencing techniques:
Read length: how long the read can be. The mapping accuracy can be increased with this
parameter.
Throughput in gigabytes or number of reads: important to know the number of reads to
know how many times you cover the read (quantitative).
Longer fragments are better, otherwise if you fragmentise the DNA in small fragments it is
hard to find out the order. When you don’t know how the genome/DNA looks, longer
fragments make it easier = de novo assembly (don’t know what the genome looks like and
want to know it).
1
,Single molecule real-time sequencing: Oxford Nanopores
Nanopore is made of protein, which in nature forms holes in membranes created by
synthetic polymers. This membrane has high electronic resistance. When the single
molecules enter through the nanopore, the structure and current is disrupting, which helps
identifying the molecule by sequencing.
Nanopore strand sequencing: an intact polymer is sequenced when it enters the nanopore,
when a strand of DNA enters through the nanopore it can read base per base. The enzyme
in the nanopore makes the dsDNA to ssDNA so it can go through the nanopore. By
disrupting the current it can determine the order of bases on the DNA strand.
Big advantage of this technology compared to the other technologies: it can recognize
modified base pairs, for example when DNA methylation occurs, because this gives a
different signal. Have a hairpin in the end of the long read strand so then it will also read the
other strand. This is typically used for long read length sequences.
Practice questions
Which method can you use to measure the relative abundance of a few dozen RNA markers
in a cohort of 600 patients?
QPCR with the array plate of the 96 samples, with 6 of those plates you can check all of
these patients.
You would like to identify the exact position of the transcription start sites genome wide.
Which method is best suited for this purpose and why?
RNA sequencing and cap sequencing. You pull intact RNA sequences which contain the
cap, of this you remove the cap and sequence the 5’ end of this sequence.
Another option is tiling arrays, but not to base pairs, so RNA sequencing would be the best
option.
You are working with an intracellular parasite that resides in the liver and interested to see
which parasite genes are upregulated when the host has fever. Knowing that it is very
difficult to isolate RNA from these parasites without substantial contamination from the host,
which method is the most economical solution to your problem? Why?
?
Examples of epigenetic features in the genome: DNA modifications, histone modifications
and variants (acetylation, methylation, phosphorylation etc.), nucleosome occupancy,
chromatin interactions and chromatin domains, and RNA modifications and non-coding
RNAs.
If you have to choose 6-8 modifications to study, which ones are the most interesting?
- DNA methylation: H3K27 methylation is important because it leads to
heterochromatin formation and repression of transcription.
- H3K9 methylation
- K27 acetylation, which is good for enhancers.
2
,What are the most important epigenetic features to profile?
- H3K4me3 and H2AZ you can find
where promoters are.
- H3K36me3 and H3K79me2 you can
know where active coding sequences
are.
- H3K4me1 and H3K27ac you know
where enhancers are.
- CTCF is a structural protein.
- To know where the inactive genes are,
look for H3K9me2 and H3K9me3
methylation.
- H3K27me3 is involved in
heterochromatin silencing.
Measuring protein-DNA interactions to analyze the localisation of epigenetic features (e.g.
histone modifications, variants, reader proteins, chromatin remodelers) and transcription
factors
Chromatin immunoprecipitation sequencing (ChIP-seq): used to study all sorts of
protein-DNA interactions and where the protein of interest binds to the DNA.
This method uses formaldehyde to crosslink the DNA and protein interactions. With
sonication you chop the DNA into smaller fragments and perform immunoprecipitation (uses
antibodies to purify) to study the individual fragments from which the transcription factor has
bound to the DNA. The binding caused by formaldehyde needs to be reversed so you only
keep the DNA to be isolated. This sequence will tell us which DNA sequences are bound by
TF in vivo for example.
Amplify the DNA to get enough for sequencing techniques and then you can study them by
microarray or sequencing. When you get the sequence read from this assay, you map
them to the reference genome, in other words: the number of reads mapped to a certain
loci is proportional to the abundance of the mark.
The peaks as a result from this assay in the genome browser show a binding site for the
transcription factor. Can analyze these results further with these examples:
3
, DNA adenine methyltransferase (DamID) as alternative to ChIP
ChIP relies on immunoprecipitation using an antibody to target the protein of interest and
sequencing of the co-precipitated DNA fragments. The primary advantage of DamID is that
there is no requirement for antibodies or other affinity reagents to purify the protein of
interest.
Drawback of CHIP is: you always lose a lot of material because you wash all the unbinding
material away and it is temporary. So if the binding is every 5 minutes it might be hard to
capture this precisely with ChiP-seq.
DamID is more sensitive and it involves the identification of DNA sequences carrying the
methylation mark that is deposited by the bacterial methylase dam, which is expressed as
a fusion to the protein of interest.
DNA gets isolated and the PCR adaptors get ligated so
that the methylated fragments can be amplified and
sequenced. This shows just as ChiP-seq an binding
profile between DNA and proteins.
DamID identifies binding sites by expressing
DNA-binding protein as fusion protein with DNA
MTFase.
DamID is based on the expression of a fusion protein
consisting of a protein of interest and DNA adenine
methyltransferase. Binding of POI to DNA, localizes the
methyltransferase in the region of the binding site. This
leads to methylation of adenines near sites where the
protein of interest interacts with the DNA. Adenine
methylation doesn’t occur naturally in eukaryotes, so
therefore adenine methylation in any region concludes
that the adenine methylation had been caused by the
fusion protein. This implies that the region is located
near a binding site. These methylated sequences are
subsequently amplified by a methylation-specific PCR
protocol and identified by hybridization to microarrays.
4
