GTRD Workflow

ChIP-seq experiment information were collected in semi-automated way from literature, GEO and ENCODE.

Raw ChIP-seq data in the form of fastq and SRA files were fetched from ENCODE and SRA databases.

Sequenced reads were aligned using Bowtie2 aligner.

ChIP-seq peaks were called using 4 different methods: MACS SISSRS GEM and PICS.

Bowtie2

We use bowtie2 version 2.2.3 for ChIP-seq read alignment to the reference genomes of human (GRCh38) and mouse (GRCm38).

Bowtie2 was run with following parameters:

bowtie2 -x $genome -U $fastq_files -p 8 --mm --seed 0

The resulting alignments were converted to bam files, then sorted and indexed using samtools version 1.0

MACS

MACS version 1.4.2 was used for peak calling with following parameters:

macs14 f BAM -g $species -n $peaks -t $alignment_bam

or if control experiment was available:

macs14 f BAM -g $species -n $peaks -t $alignment_bam -c $control_bam

SISSRS

SISSRS requires alignments in bed format, bam files were converted to bed files using bedtools version 2 by:

bamToBed -i $input_bam > $output_bed

Version 1.4 of SISSRS were used for peaks calling with following parameters:

sissrs.pl -i $alignment_bed -s 3000000000 -o $peaks.sissrs

or if control experiment was available:

sissrs.pl -i $alignment_bed -s 3000000000 -o $peaks.sissrs -b $control_bed

GEM

GEM version 2.5 was used with following parameters:

java -Xmx4G -XX:+UseSerialGC -jar /srv/local-main/tools/gem/gem.jar --d /srv/local-main/tools/gem/Read_Distribution_default.txt
--g /srv/local-main/tools/gem/$species.chrom.sizes --s 2000000000 --f SAM --t 1 --out $peaks --expt $bam

or if control experiment was available:

java -Xmx4G -XX:+UseSerialGC -jar /srv/local-main/tools/gem/gem.jar --d /srv/local-main/tools/gem/Read_Distribution_default.txt
--g /srv/local-main/tools/gem/$species.chrom.sizes --s 2000000000 --f SAM --t 1 --out $peaks --expt $bam --ctrl $control

For the large datasets -Xmx24G parameter was set.

PICS

We use following R script to call peaks using PICS: