RNA-seq, ChIP-seq and 3C

Genomics tells us about the structure of the genome, transcriptomics tells us about gene expression. Microarrays and RNA-seq allow us to perform genome-wide experiments to measure the expression of every gene simultaneously. 

Answers questions like: Which genes are expressed?, When are they expressed?, How much are they expressed? 

Examples already looked at: 

Northern Blotting - Assess expression of a single gene at a time. Extract RNA and undergo electrophoresis which separates RNA by size. Transfer of RNA to membrane and add labeled probes of gene of interest. They hybridise to plot and stick to RNA of gene interested in giving qualitative idea of levels. Visualization of labeled RNA on x-ray film. 

RT-qPCR (reverse transcription - quantitative PCR) - Single gene analysis. Reverse transcription to form cDNA and then undergo PCR to amplify DNA. Quantitative as fluorescent primer is used so you can measure fluorescence levels at each cycle of PCR and compare to standard. 

Is a orderded arrangement of small spots of different DNA sequence probes on a glass slide. One spot contains lots of copies of the sequence. These probes of synthesized onto the surface of the slide. Labelled RNA samples then hybridise to probes promoting fluorescence. Intensity of spot is used to work out level of expression. A darker green dot shows a lower expression than a light green dot. A black dot shows no expression of that gene. 

Limitations: 

• Considered low- resolution sequencing technology
• If we get a signal for a probe we know that sequence or similar is present but don't know exact sequence
• Also don't know if there are any sequences present which are not covered by the microarray probes
• There is a limit to how uch RNA can hybridise to a particular spot which can limit ability to distinguish the expression levels of highly expressed genes that saturate the spot

Uses lots of sequence reads so only began with NGS. Fragment input RNA and convert to cDNA using reverse transcriptase and add sequencing adapters to create DNA library. Undergoes high throughout sequencing and creates RNA-seq reads. Reads are then mapped to a reference genome. Number of reads for a section of genome is used a proxy for expression. Lots of reads means a high level of RNA expression from that part of the genome. Few reads means low RNA expression. 

Split genes (exons and introns) in eukaryotes cause issues in RNA-seq. This is beause some reads are seperated if they run over two exons. When mapping, data programmes that are aware of this must be used. 

De-novo assembly of RNA seq reads can be done but is more difficult than DNA seq assembly. Once the sequence is in contigs it can be processed like the DNA. 

• Both well developed technologies with minimal technical variation 
• RNA-seq has a larger dynamic range than microarrays (greater ability to distinguish different levels of expression) 
• Microarrays only give info for pre-selected regions. RNA-seq is genome wide and can detect novel transcripts. 
• RNA-seq allows us to detect differences from the reference genome such as SNPs 
• RNA-seq can be done without a reference genome as de novo assembly is possible. 

Have a treated sample and a control sample eg compare drug vs no drug, mutant vs WT, ensuring the only difference in the thing you are interested in. 

Replicates need to be done so you can start to do reliable statistics. Biological replicates is using two different biological samples and tells us about variation in samples. Typically this variation is much larger than techincal replicates. Technical replicates is where you run many arrays/RNA-seq etc from the same biological sample. This tells us how reproducible the method is. 

Statistical output of an experiment: Use a log to base 2 fold chain relative to the control sample. This means if expression remains the same the value is 0 and if the expression doubles you get 1 and if it reduces you get -1. If it wasn't a log all reduced expression values would fall between 0 and 1, which is difficult for analysis. Something similar to a t-test is used to work out a p-value. 

Method used to isolate DNA bound by a specific protein. Useful to investigate promoters bound by transcription factors etc. A bit like microarrays but used less these days.  

• Proteins covalently crosslinked to DNA by treating with formaldehyde
• Chromatin sheared (segmentation) by sonication or endonuclease. Use of an exonuclease allows the bound DNA to be trimmed to binding site (trimming loose ends) 
• Immunoprecipitation and purification of bound DNA using an antibody specific to the protein of interest 

ChIP-seq 

Bind DNA to protein and digest 5' overhang using 5'-3' exonuclease (like before). Sequence (like illumina) DNA fragments and map reads to a reference genome. Binding sites is the overlap between the forward and reverse peaks. 

Looks for interacting parts of genome. 

• Cross-linked chromatin using formaldehyde as this links interacting parts of the genome
• Restriction digest 
• Ligate ends
• Remove cross links so now a single DNA molecule

3C: look for a one-to-one suspected link. PCR with primers specific to both interacting regions needed. 

4C: looking for all regions interacting with one remote region. Circularisation, via ligation, allows PCR using primers from just one region. 

5C: looking for many-to-many interactions (more likely use Hi-C). PCR with universal primers allows amplification of many interacting regions. 

Hi-C:looking for all-to-all interactions. Junctions are labelled with biotin during ligation step allowing purification and sequencing. Biotin can easily purify so you can easily find regions which are interacting.