Genome Spot

When analyzing ChIP-seq data, we sometimes like to focus on promoters as a proxy of gene activation especially for histone marks that are normally associated with gene activation such as H3K9 acetylation or H3K4 tri-methylation.

FeatureCounts has emerged as a competitor to HTSeq and BedTools MultiCov for counting reads across features (ie, exons, genes, promoters). FeatureCounts is great for RNA-seq because it can natively read GTF annotation files, but can't read BED format (that we use a lot in ChIP-seq analysis).

In order to make featureCounts work, we need to extract the TSS coordinates and convert to a BED-like format that it can read (SAF format). In the below script, I extract the positions of the TSSs using a grep search followed by stripping only the neccessary information from the GTF, then using awk and BedTools to expand the region  around the TSS by 3kbp.

#!/bin/bash
#Generate an SAF file for TSS intervals from a GTF

#Specify some parameters
GTF=Homo_sapiens.GRCh38.78.gtf
…

Gene regulation is a really complicated thing. We have covalent marks to DNA, histones and transcription factors. Chromatin remodeling and long range enhancer interactions. Enhancer elements located in introns of genes hundreds of kilobases away from the gene they're controlling. Transcriptional control from microRNA networks and now there is an emerging model for the function of some of the thousands of long non-coding RNAs which are just now being uncovered with high resolution (directional) transcriptome analysis.

Many of you which studied molecular biology at Uni would (should) remember the model for how X chromosome inactivation is achieved. The mechanism centers around XIST, one of the first non-coding RNA genes identified. Expression of XIST from the inactive X chromosome essentially wraps it up at the same time that repressive epigenetic marks are established through its interaction with the Polycomb Repressive Complex 2 (PRC2). Sounds simple enough, but the model also inv…

There have been a number of high profile profile articles in recent times discussing the function of TET proteins, mostly in the conversion of methylated cytosine (5mC) into hydroxymethylated cytosine (5hmC), the 5th base. Hydroxymethylcytosine is much rarer than methylcytosine and is thought to be an intermediate towards demethylation of cytosine, a mechanism which remains incompletely resolved. A paper last year showed that TET proteins also convert 5hmc to 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC), termed the 6th and 7th bases.


OGT on the other hand is a fairly unique protein because it is the only known known O-GlcNAc transferase in mammals. What is GlcNAc you say? It stands for N-acetylglucosamine, a hexosamine. There has been a series of papers (here, here, here) discussing OGT as a nutrient sensor, transferring GlcNAc during period of surplus nutrient supply. GlcNAc can be transferred to the same amino acids as phosphorylation, so there is a suggested crosstalk betwee…