Chapter 9: Gene Regulation
Genes can be regulated in several different ways, and be regulated transcriptionally (at the gene-level), post-transcriptionally (at the mRNA level), translationally, and post-translationally. Splicing of mRNA can also be regulated and there are many other inputs to regulation still.
Clearly, transcriptional regulation and post-transcriptional regulation are best suited for study by nucleic acid sequence bioinformatics because techniques such as ChIP-seq and RNA-seq can be used to measure the inputs and outputs of these modes of regulation. Here we will focus on regulation by transcription factors using ChIP-seq and regulation by microRNAs using small RNA-seq and RNA-seq.
9.1 Transcription Factors and ChIP-seq
Transcription factors (TFs) are proteins that bind to DNA and are capable of regulating the transcription of genes. TFs can be activators, which activate transcription of the genes they regulate, or repressors, which block the transcription of the genes they regulate.
9.1.1 ChIP-seq Peak identification
ChIP-seq measures the genome-wide binding of a TF by sequencing DNA fragments that are bound to the TF in a controlled experiment [92 – 94]. Antibodies that recognize epitopes in the protein of interest are used to “pull-down” the protein. When cross-linking is used, such as by adding formaldehyde, everything sticks together, holding the protein to the DNA that it is bound to. After the antibodies and associated chromatin are precipitated out from the solution and isolated, they can be reverse-crosslinked, and then a DNA-extraction is performed. The fragments of DNA are then sequenced using deep sequencing. Despite best efforts to just isolate fragments directly bound the the TF of interest, a lot of spurious reads make it through the process as well. Therefore, we need to sequence a control sample without the antibody, either using the immunogloblulin (IgG) or just naked DNA (input), so that we can look for statistical enrichment of the reads that align to a particular genomic location from the ChIP-Sample compared to this control sample.
Short read alignment with [latex]\texttt{bowtie}[/latex]
Once we have the the reads from the high-throughput sequencer, typically in a FASTQ file, we need to align to the genome. This can be achieved with [latex]\texttt{bowtie}[/latex] and a bowtie index of the genome [95].
$ bowtie -m 1 -S -q hg38 TF_ChIPSeq.fastq TF_ChIPSeq_hg38_bowtie.sam
$ bowtie -m 1 -S -q hg38 DNA_input.fastq DNA_input_hg38_bowtie.sam
In this particular example, we are aligning the fastq files such that we are only retaining alignments that have one hit to the genome. This is done with the “[latex]\texttt{-m 1}[/latex]” option. If one wanted to retain only reads that map to at most [latex]\texttt{M}[/latex] hits, then we would use “[latex]\texttt{-m M}[/latex]” where [latex]\texttt{M}[/latex] is some integer. Recall that [latex]\texttt{hg38}[/latex] is the “base” of the bowtie index of the genome. It is actu