Get human transcription factor target genes

2015-03-05

I made a data package with human transcription factor target genes for use in R. It is a collection of data from three sources: TRED, ITFP, and ENCODE. I use them to test if the targets of a transcription factor are differentially expressed in my data. Also, I can test if a set of transcription factor target genes is enriched for some gene set of interest.

On Biostars, several people have asked how to find the targets of transcription factors:

To address these questions, I made commonly mentioned resources available for immediate use in R. This allows me to avoid navigating a website and typing in gene names.

Usage#

The github repository: https://github.com/slowkow/tftargets

Download the RData file:

# install.packages("RCurl")
library(RCurl)
download.file(
  url = "https://raw.githubusercontent.com/slowkow/tftargets/master/data/tftargets.RData",
  destfile = "tftargets.RData",
  method = "curl"
)
load("tftargets.RData")

List the Entrez Gene IDs for targets of a transcription factor:

> TRED[["STAT3"]]
 [1]      2    332    355    595    596    598    896    943    958   1026 1051
[12]   1401   1588   1962   2194   2209   2353   3082   3162   3320   3326 3479
[23]   3559   3572   3586   3659   3718   3725   3929   4170   4582   4585 4609
[34]   4843   5008   5021   5292   5551   5967   6095   6347   6654   7076 7078
[45]   7097   7124   7200   7422   7432   8651   8996   9021  11336  23514 26229
[56]  27151  55893 117153 201254

In my analyses, I am finding that TRED and ENCODE are both useful datasets. ITFP seems to be too noisy in my analyses, but you might have better luck.

Today, I found an additional dataset worth checking out by Neph et al. 2012 that comes with an interesting web-based visualization: http://www.regulatorynetworks.org/. I will eventually add it to my R package.

Note#

I have not performed careful filtering of the datasets, so each transcription factor might have low confidence targets, or targets found in different cell types. If you’re interested in a particular cell type, you might want to filter the data and prepare your own lists instead of using the lists that I prepared.

© 2024 Kamil Slowikowski