Meth10X

• Meth10X (old name: WGBSX10) is a pipeline of parallel alignment of whole genome bisulfite sequencing (WGBS) or nucleosome occupancy and DNA methylation (NOMEseq) from inputs as fastq files of multi samples (approx. 8-16 samples with more than 20X coverage for each) and generating:

  • An HTML report of quality control detail metrics: percent mapped reads, methylation average of CpG and non-CpG, bisulfite conversion rate, CpG islands/shores bias, coverage bias between all sites and CpG sites, percent SNP overlapping with CpG sites, Top 5 SNPs in each sample.
  • MDS plots to compare among samples
  • DNA methylation distribution plots of all CpG sites, CpG island sites, CpG shore sites...
  • A tsv big table of CG/CHH/CHG with filtered/unfiltered SNP
  • SNP calling
  • PMD, LMR, UMR, HMR calling
  • Bigwig files of whole genome coverage, only CpG coverage, CpG methylation of each sample for IGV visualization of these samples

• Meth10X is designed for running on cluster such as SGE/PBS. It requires 4GB of memory to maintain the running of the whole workflow.

• Meth10X helps to speed up the whole workflow of high coverage WGBS as well as enhance the accurate of methylation and SNP calling.

• Meth10X is built based on in-house bash/python/perl/R script, Bpipe and a collection of software packages:

  • [bwa, 0.7.9a] https://github.com/lh3/bwa
  • [bwa-meth, 0.10] https://github.com/brentp/bwa-meth
  • [samtools, 1.4] https://github.com/samtools/
  • [tabix, 1.4] https://github.com/samtools/tabix
  • [qualimap, 2.2.1] http://qualimap.bioinfo.cipf.es/
  • [SAMStat, 1.08] http://samstat.sourceforge.net/
  • [picardtools, 2.3.0] https://github.com/broadinstitute/picard
  • [bedtools, 2.22.0] https://github.com/arq5x/bedtools2
  • [Bpipe, 0.9.9.2] https://github.com/ssadedin/bpipe
  • [bamUtil, 1.0.7] https://github.com/statgen/bamUtil
  • [MethylDackel, 0.2.0] https://github.com/dpryan79/MethylDackel
  • [Biscuit, 0.2.0] https://github.com/zwdzwd/biscuit
  • [MethylSeekR, 1.0] https://github.com/Bioconductor-mirror/MethylSeekR
  • [Vcf-tools, 0.1.15] https://github.com/vcftools
  • [fastqc, 0.11.5] http://www.bioinformatics.babraham.ac.uk/projects/fastqc/
  • [trimCEGX, 1.0] https://github.com/luuloi/trim.paired.read
  • [makeFullDataFrame, 1.0] https://github.com/luuloi/make.full.data.frame
  • [merge_columns_multi_tsv, 1.0] https://github.com/luuloi/merge_columns_multi_tsv
  • [UCSC-format-file-converter, 1.0] http://hgdownload.soe.ucsc.edu/admin/exe/linux.x86_64/
  • [ggplot2] https://github.com/tidyverse/ggplot2
  • [BSgenome.Hsapiens.UCSC.version] https://github.com/Przemol/BSgenome.Hsapiens.UCSC.version
  • [GenomicRanges] https://github.com/Bioconductor-mirror/GenomicRanges
  • [rtracklayer] https://github.com/Bioconductor-mirror/rtracklayer
  • [python modules] configobj, argparse

• Dependencies

  • Java version >= 1.8
  • All the above listed software packages, here are some commands to install python and R packages
    • [python version >= 2.7.6]
      • pip install configobj argparse
    • [R version >= 3.1.2]
      • source("https://bioconductor.org/biocLite.R")
      • biocLite(c("BSgenome.Hsapiens.UCSC.hg19", "rtracklayer", "GenomicRanges", "MethylSeekR"))
      • install.packages(c("ggplot2", "data.table", "ggfortify", "knitr"))

• Installation

  • git clone git@github.com:luuloi/Meth10X.git
  • or
  • git clone https://github.com/luuloi/Meth10X.git

• Inputs require

  • Paired fastq files in .gz format for each sample
  • Filled sample config file, example https://github.com/luuloi/WGBSX10/blob/master/config/sample.Test_ProstateC.config
  • Filled system config file, example https://github.com/luuloi/WGBSX10/blob/master/config/system.Test_ProstateC.config
  • Note:
    • The fastq files are located in project/raw/LNCaP/lane1/lane1_R1.fastq.gz, project/raw/sample/lane2/lane1_R2.fastq.gz of each sample.
    • Lane and sample name should not have any dots '.' in it.

• First Run [TODO]

  • Indexing the human genome/or download the indexed genome before alignment
  • Run the pipeline: https://github.com/luuloi/WGBSX10/blob/master/example/example.sh

    • module load phuluu/python/2.7.8

    • module load gi/java/jdk1.8.0_25

    • module load phuluu/bpipe/0.9.9.2

    • mkdir -p Test_ProstateC/raw

    • copy fastq files to Test_ProstateC/raw as following

      • cp sample1/lane1/*.fastq Test_ProstateC/raw/sample1/lane1/
      • cp sample1/lane2/*.fastq Test_ProstateC/raw/sample1/lane2/
      • cp sample2/lane1/*.fastq Test_ProstateC/raw/sample2/lane1/
      • ....

    • python "WGBS10X/pipe/run_Bpipe.py" "WGBS10X/config/sample.Test_ProstateC.config" "WGBS10X/config/system.Test_ProstateC.config"

• Second Run

  • module load phuluu/python/2.7.8

  • module load gi/java/jdk1.8.0_25

  • module load phuluu/bpipe/0.9.9.2

  • python "WGBS10X/pipe/run_Bpipe.py" "WGBS10X/config/sample.Prostate.config" "WGBS10X/config/system.Prostate.config"

• Outputs

  • HTML report

    • Table metrics of quality control of WGBS
    • Depth vs Breadth of CpG sites
    • Overlapping of DNA methylation and SNP calling
    • PMD, UMR, LMR and HMR calling
    • Distribution of methylation
    • MDS clustering of samples

  • CpG big table contain the count of C and coverage of all CpG sites in the reference genome similar to CHG and CHH big table

    • #chr	position	LNCaP.C	LNCaP.cov	PrEC_1.C	PrEC_1.cov	PrEC.C	PrEC.cov
      chr1	10469	10	11	2	12	3	14
      chr1	10471	15	15	1	14	2	12
      ...	...	...	...	...	...	...	...

  • CpG big table [filtered out CpG overlapping with SNP] contain the count of C and coverage of all CpG sites in the reference genome [LNCaP having SNP at postion chr1:10471 therefore this CpG methylation value is ommitted ]

    • #chr	position	LNCaP.C	LNCaP.cov	PrEC_1.C	PrEC_1.cov	PrEC.C	PrEC.cov
      chr1	10469	10	11	2	12	3	14
      chr1	10471	0	0	1	14	2	12
      ...	...	...	...	...	...	...	...

  • Output folder structure

    •