This is an workflow to call accessible chromatin regions (ACRs) in plants using differential MNase digestion and high-throughput sequencing (MNase-seq, left in the flow chart) and data analysis (right in the flow chart).
-
Running environment:
- The workflow was constructed based on the Linux system running the Oracle v1.6 to 1.8 java runtime environment (JREs).
-
Required software and versions:
The demo data used here is the paired-end fastq file generated by using Illumina platform.
- R1 FASTQ file:
input/demo_leaf_light.r1.fq - R2 FASTQ file:
input/demo_leaf_light.r2.fq
The fastq file includes data under light-digestion and heavy-gigestion.
- Light-digestion FASTQ file:
input/demo_leaf_light.r1.fq - Heavy-digestion FASTQ file:
input/demo_leaf_heavy.r1.fq
The data generated in two tissue types are provided.
- Leaf tissue FASTQ file:
input/demo_leaf_light.r1.fq - Root tissue FASTQ file:
input/demo_root_light.r1.fq
- Create a new working directory to store all the MNase-seq-related data and files.
cd data1/user/path # enter your working directory
mkdir MNase_seq
cd MNase_seq
mkdir raw_datathis directorydata # all your sequencing data here
- Use FastQC to evaluate the quality of your raw data.
mkdir ./fastqc_result
for i in ./raw_data/*.fq.gz
do
fastqc -o ./fastqc_result $i
done
- Use Trimmomatic software to obtain clean data. The main goal of this step is to remove the adapter sequence, leading and trailing low-quality or N bases and sequences too short in length.
name=leaf_light
mkdir clean_data
java -jar trimmomatic-0.36.jar PE -phred33 \
./raw_data/*.r1.fq.gz \
./raw_data/*.r2.fq.gz \
./clean_data/$name.1P.fq \
./clean_data/$name.1U.fq \
./clean_data/$name.2P.fq \
./clean_data/$name.2U.fq \
ILLUMINACLIP:data1/user/tools/Trimmomatic-0.36/adapters/\
TruSeq3-PE-2.fa:2:30:10:1:true \
LEADING:5 TRAILING:5 MINLEN:20 >trimmomatic.$name.log
- Use FastQC again to evaluate the quality of the clean data.
for i in ./clean_data/*.fq.gz
do
fastqc -o ./fastqc_result $i
done
- Align the filtered clean data to the reference genome using Bowtie2 software. Before that, library the genome with bowtie2-build.
bowtie2-build ref_genome.fa ref.genome
samtools faidx ref_genome.fa
bowtie2 --phred33 -p 10 --reorder -5 6 \
--no-mixed --no-discordant --no-unal --dovetail \
-x /index_path/ref.genome \
-1 ./$name.1P.fq \
-2 ./$name.2P.fq \
-S ./$name.sam >bowtie2.$name.log
- Screen the alignment according to the MAPQ value in the alignment results. After that, remove the repeated reads due to PCR library amplification with samtools markdup.
function samtools_scripts(){
samtools view -q 20 -bhS $name.sam -o $name.Q20.bam
samtools sort -n $name.Q20.bam -o $name.Q20.nsort.bam
samtools fixmate -m $name.Q20.nsort.bam $name.Q20.nsort.ms.bam
samtools sort $name.Q20.nsort.ms.bam -o $name.Q20.ms.psort.bam
samtools markdup -r $name.Q20.ms.psort.bam $name.Q20.psort.markdup.bam
samtools index -c $name.Q20.psort.markdup.bam
}
samtools_scripts $name >samtools.$name.log
- Subtract the mapped reads of heavy-digestion from light-digestion using bamCompare in the deepTools, and then normalize by the CPM method to output a DNS score file in bedgraph format.
bamCompare -b1 leaf_light.bam -b2 leaf_heavy.bam \
--scaleFactorsMethod None --normalizeUsing CPM --operation subtract \
--binSize 10 --smoothLength 50 --outFileFormat bedgraph \
-o leaf_diff.bdg
- Calculate the reads coverage from both light-digested and heavy-digested samples, and the CPM method is used for normalization. The Bayes factor is calculated for reads coverage in 10-bp intervals.
bamCoverage -b leaf_light.bam \
--binSize 10 --smoothLength 50 \
--normalizeUsing CPM --outFileFormat bedgraph \
-o leaf_light.bdg
bamCoverage -b leaf_heavy.bam \
--binSize 10 --smoothLength 50 \
--normalizeUsing CPM --outFileFormat bedgraph \
-o leaf_heavy.bdg
bedtools unionbedg \
-i leaf_light.bdg leaf_heavy.bdg \
-header -names light heavy \
>leaf.unionbdg
Rscript bayes_factor_caculator.R leaf.unionbdg
- DNS parts with values greater/less than 0 and Bayes values greater than 0.5 are defined as significant MSFs/MRFs. We referred to the significant MSFs as MNase-hypersensitive (MNase HS) regions or ACRs.
bedtools unionbedg \
-i leaf_diff.bdg leaf_unionbdg.bayes \
>leaf_diff.bayes
cat leaf_diff.bayes | awk '$4>0 && $5>0.5' | \
cut -f1-3 | bedtools merge -d 200 \
>leaf_MSF_bayes_0.5_merge_200.bed
cat leaf_diff.bayes | awk '$4<0 && $5>0.5' | \
cut -f1-3 | bedtools merge -d 200 \
>leaf_MRF_bayes_0.5_merge_200.bed
- Generate reads coverage depth in bigwig files with deepTools and then load them into the Integrative Genomics Viewer (IGV browser) for browsing.
bamCoverage -b leaf_light.bam \
--binSize 10 --smoothLength 50 \
--normalizeUsing CPM --outFileFormat bigwig \
-o leaf_light.bw
bamCoverage -b leaf_heavy.bam \
--binSize 10 --smoothLength 50 \
--normalizeUsing CPM --outFileFormat bigwig \
-o leaf_heavy.bw
bamCompare -b1 leaf_light.bam -b2 leaf_heavy.bam \
--scaleFactorsMethod None --normalizeUsing CPM --operation subtract \
--binSize 10 --smoothLength 50 --outFileFormat bigwig \
-o leaf_diff.bw
This is an example result generated from the demo data. The upper track illustrates an gene model, the middle and lower tracks show the signal of MSFs (in red) and MRFs (in blue) in leaf and root, and significant MSF region, or ACRs, are detected and shown in orange segments.
It is a free and open source software, licensed under (choose a license from the suggested list: GPLv3, MIT, or CC BY 4.0).