Functional enrichment for QTL variants

This script performs functional enrichment of QTL variants.

How to install

This will guide you through compiling fenrichcpp.

Compilation on Ubuntu 20.01.1 Desktop Distribution

1. Install all necessary packages.

sudo apt install g++ gcc gfortran make autoconf automake libtool zlib1g-dev liblzma-dev libbz2-dev lbzip2 libgsl-dev libblas-dev libx11-dev libboost1.71-all-dev git libreadline-dev libxt-dev libpcre2-dev libcurl4-openssl-dev

2. You will need to download the HTSlib library

mkdir ~/Tools
cd ~/Tools
wget https://github.com/samtools/htslib/releases/download/1.10.2/htslib-1.10.2.tar.bz2
tar jxvf htslib-1.10.2.tar.bz2

3. Compile HTSlib

cd ~/Tools/htslib-1.10.2/ 
./configure 
make

4. Edit fenrichcpp Makefile

cd ~/
git clone https://github.com/NLykoskoufis/fenrichcpp.git
cd ~/fenrichcpp/

##########################################
# SET THESE 6 PATHS CORRECTLY TO COMPILE #
##########################################
BOOST_INC=/usr/include
BOOST_LIB=/usr/lib/x86_64-linux-gnu
HTSLD_INC=$(HOME)/Tools/htslib-1.10.2
HTSLD_LIB=$(HOME)/Tools/htslib-1.10.2

5. Compile fenrichcpp

make

6. If you want to clean

make clean 

Usage example

The functional enrichment is performed in two steps. The first step creates a file containing for each variant inside a vcf, the MAF and the closest upstream and downstream feature (e.g gene). Second, it will read through all QTL variants and for each of them randomly pick variants (from the file created previously) that match MAF (±2%) and distance to TSS (±2.500) and that are not eQTLs. This allows for the generation of a null distribution. Then, it uses a Fisher exact test to perform enrichment over the null of the QTL variants.

Step by step example

Step1: Create null file

Before creating the null file you also need to provide a list of variants that are nominally significant eQTLs. These can be provided by using QTLtools cis --nominal 0.05. Then you need to only keep the var_id. It should be the 8th column.

fenrich null \
  --vcf genotypes.chr22.vcf.gz \
  --bed genes.50percent.chr22.bed.gz \
  --qtl nominal_only_significant_snps.txt.gz \
  --out null_variantFile.txt.gz

Step2: Run enrichment analysis

IMPORTANT: the eQTL results should be in QTLtools cis format. Please check QTLtools for the exact format to use.

fenrich enrich \
  --null null_variantFile.txt.gz \
  --phen TFs.encode.bed.gz \
  --qtl results.permute.genes.significant.txt.gz \
  --out qtlTFenrichment.txt \
  --random_var 100 \
  --window_size 2500 \
  --maf_window 0.02

The output of fenrich enrich mode looks like this:

#A       B       C       D
#59      22      214     531

A -> Number of eQTLs overlapping with the peaks.

B -> Number of null variants overlapping with peaks.

C -> Number of eQTls not overlapping with peaks.

D -> Number of null variants not overlapping with peaks.

Next, you can load the results file into R, create a matrix and run a fisher exact test as shown below.

D = read.table("qtlTFenrichment.txt",header=TRUE,sep="\t",stringsAsFactors=FALSE)
mat = matrix(D,ncol=2, byrow=TRUE)
fisher.test(mat)