Bioinformatics manual for population epigenomics combining whole-genome and target genome sequencing

Odile Rogier, Isabelle Lesur Kupin, Mamadou Dia Sow, Christophe Boury, Alexandre Duplan, Abel Garnier, Abdeljalil Senhaji rachik, Peter Civan, Josquin Daron, Alain Delaunay, Ludovic Duvaux, Vanina Benoit, Erwan Guichoux, Gregoire Le Provost, Edmond Sanou, Christophe Ambroise, Christophe Plomion, Jérôme Salse, Vincent Segura, Jorg Tost, Stéphane Maury

Published: 2024-06-20 DOI: 10.17504/protocols.io.8epv5xw4ng1b/v1

DNA Methylation

Epigenetics

Epigenomics

Methylome

Natural population

Oak

Poplar

Transposon Insertion Polymorphism

SeqCapBis

WGS

WGBS

AI 解读

Abstract

We developed a strategy and a workflow for quantifying epigenetic diversity in natural populations combining whole genome and targeted capture sequencing for DNA methylation.

We first identified regions of highly variable DNA methylation in a representative subset of genotypes representative of the biological diversity in the population by WGBS. We then analysed the variations of DNA methylation in these targeted regions at the population level by Sequencing Capture Bisulphite (SeqCapBis).

Steps

Whole Genome Sequencing - Removing false C/T mutations

A preliminary Whole Genome Sequencing (WGS) step was considered for filtering purposes, to prevent C/T Single Nucleotide Polymorphisms (SNP) being interpreted as bisulfite conversions of unmethylated sites (i.e. false-positive calls). However, this C/T SNPs identification step is not required to study epigenetics levels along genomes.

Strategy for population epigenomics combining whole-genome and target genomesequencing.

Trimming

Software

Value	Label
Trimmomatic	NAME
http://www.usadellab.org/cms/?page=trimmomatic	REPOSITORY
http://doi.org/10.1093/bioinformatics/btu170	DEVELOPER
http://www.usadellab.org/cms/?page=trimmomatic	LINK
Version 0.39	VERSION

Publication: Bolger et al., 2014

Version: 0.38

Github: http://github.com/usadellab/Trimmomatic

Citation

Bolger AM, Lohse M, Usadel B 2014 Trimmomatic: a flexible trimmer for Illumina sequence data. http://doi.org/10.1093/bioinformatics/btu170

java -Xmx4G -jar trimmomatic.jar PE -threads 12 file_R1.fastq.gz file_R2.fastq.gz 
file_trimmed_1.fastq.gz  file_unpaired_1.fastq.gz file_trimmed_2.fastq.gz
file_unpaired_2.fastq.gz ILLUMINACLIP:TruSeq3-PE.fa:2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:35

Mapping

Software

Value	Label
BWA	NAME
Unix	OS_NAME
http://bio-bwa.sourceforge.net/	REPOSITORY
Li, H., Durbin, R.	DEVELOPER
http://arxiv.org/abs/1303.3997	LINK
0.7.12- r1039	VERSION

Publication: Li H, 2013

Version: 0.7.17

Citation

Heng Li Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM arXiv:1303.3997 [q-bio.GN] http://doi.org/10.48550/arXiv.1303.3997

Poplar genome: Populus trichocarpa v3.1

Publication: Tuskan GA et al., 2006.

Citation

Tuskan GA, Difazio S, Jansson S, Bohlmann J, Grigoriev I, Hellsten U, Putnam N, Ralph S, Rombauts S, Salamov A, Schein J, Sterck L, Aerts A, Bhalerao RR, Bhalerao RP, Blaudez D, Boerjan W, Brun A, Brunner A, Busov V, Campbell M, Carlson J, Chalot M, Chapman J, Chen GL, Cooper D, Coutinho PM, Couturier J, Covert S, Cronk Q, Cunningham R, Davis J, Degroeve S, Déjardin A, Depamphilis C, Detter J, Dirks B, Dubchak I, Duplessis S, Ehlting J, Ellis B, Gendler K, Goodstein D, Gribskov M, Grimwood J, Groover A, Gunter L, Hamberger B, Heinze B, Helariutta Y, Henrissat B, Holligan D, Holt R, Huang W, Islam-Faridi N, Jones S, Jones-Rhoades M, Jorgensen R, Joshi C, Kangasjärvi J, Karlsson J, Kelleher C, Kirkpatrick R, Kirst M, Kohler A, Kalluri U, Larimer F, Leebens-Mack J, Leplé JC, Locascio P, Lou Y, Lucas S, Martin F, Montanini B, Napoli C, Nelson DR, Nelson C, Nieminen K, Nilsson O, Pereda V, Peter G, Philippe R, Pilate G, Poliakov A, Razumovskaya J, Richardson P, Rinaldi C, Ritland K, Rouzé P, Ryaboy D, Schmutz J, Schrader J, Segerman B, Shin H, Siddiqui A, Sterky F, Terry A, Tsai CJ, Uberbacher E, Unneberg P, Vahala J, Wall K, Wessler S, Yang G, Yin T, Douglas C, Marra M, Sandberg G, Van de Peer Y, Rokhsar D 2006 The genome of black cottonwood, Populus trichocarpa (Torr. & Gray). http://doi.org/

bwa mem genome.fa file_trimmed_1.fastq.gz file_trimmed_2.fastq.gz -t 12 -M > file.sam

3.1.

Mapping adjustments for Q. petraea Q. petraea

Oak genome: Quercus robur Haplome V2.3

Publication: Plomion C et al., 2018

Citation

Plomion C, Aury JM, Amselem J, Leroy T, Murat F, Duplessis S, Faye S, Francillonne N, Labadie K, Le Provost G, Lesur I, Bartholomé J, Faivre-Rampant P, Kohler A, Leplé JC, Chantret N, Chen J, Diévart A, Alaeitabar T, Barbe V, Belser C, Bergès H, Bodénès C, Bogeat-Triboulot MB, Bouffaud ML, Brachi B, Chancerel E, Cohen D, Couloux A, Da Silva C, Dossat C, Ehrenmann F, Gaspin C, Grima-Pettenati J, Guichoux E, Hecker A, Herrmann S, Hugueney P, Hummel I, Klopp C, Lalanne C, Lascoux M, Lasserre E, Lemainque A, Desprez-Loustau ML, Luyten I, Madoui MA, Mangenot S, Marchal C, Maumus F, Mercier J, Michotey C, Panaud O, Picault N, Rouhier N, Rué O, Rustenholz C, Salin F, Soler M, Tarkka M, Velt A, Zanne AE, Martin F, Wincker P, Quesneville H, Kremer A, Salse J 2018 Oak genome reveals facets of long lifespan. http://doi.org/10.1038/s41477-018-0172-3

3.2.

Mapping conversion, sorting & statistics

Software

Value	Label
SAMtools	NAME
http://github.com/samtools/	REPOSITORY
Li et al.	DEVELOPER
http://www.htslib.org/	LINK
1.3.1	VERSION

Publication: Danecek et al., 2021

Version: 1.8

Github: http://github.com/samtools/samtools

Citation

Danecek P, Bonfield JK, Liddle J, Marshall J, Ohan V, Pollard MO, Whitwham A, Keane T, McCarthy SA, Davies RM, Li H 2021 Twelve years of SAMtools and BCFtools. http://doi.org/10.1093/gigascience/giab008

samtools view -Sb file_trimmed.sam > file_trimmed.bam 
samtools sort file_trimmed.bam  -o file_trimmed_sorted.bam 
samtools flagstat file_trimmed_sorted.bam > file_flagstats.txt 
samtools stats file_trimmed_sorted.bam > file_stats.txt

Variant calling

4.1.

Adjustment for Q. petraea : Digital normalization Q. petraea : Digital normalization

Computational limitations associated with GATK and FreeBayes due to the very deep sequencing in oak (100X on average) necessitated a reduction of the complexity of each dataset. To reduce redundancy within the WGS dataset, we randomly downsampled sequencing reads over genome regions that are over-covered.

Software

Value	Label
KHMER	NAME
Linux	OS_NAME
http://khmer.readthedocs.io/en/latest/	REPOSITORY
Titus Brown	DEVELOPER

Publication: Crusoe et al., 2015

Version: 2.1.2

Github: http://github.com/dib-lab/khmer

Citation

Crusoe MR, Alameldin HF, Awad S, Boucher E, Caldwell A, Cartwright R, Charbonneau A, Constantinides B, Edvenson G, Fay S, Fenton J, Fenzl T, Fish J, Garcia-Gutierrez L, Garland P, Gluck J, González I, Guermond S, Guo J, Gupta A, Herr JR, Howe A, Hyer A, Härpfer A, Irber L, Kidd R, Lin D, Lippi J, Mansour T, McA'Nulty P, McDonald E, Mizzi J, Murray KD, Nahum JR, Nanlohy K, Nederbragt AJ, Ortiz-Zuazaga H, Ory J, Pell J, Pepe-Ranney C, Russ ZN, Schwarz E, Scott C, Seaman J, Sievert S, Simpson J, Skennerton CT, Spencer J, Srinivasan R, Standage D, Stapleton JA, Steinman SR, Stein J, Taylor B, Trimble W, Wiencko HL, Wright M, Wyss B, Zhang Q, Zyme E, Brown CT 2015 The khmer software package: enabling efficient nucleotide sequence analysis. http://doi.org/10.12688/f1000research.6924.1

Step1: Interleave reads

Parameters: Python-3.6.3

interleave-reads.py file_R1.fastq file_R2.fastq -o file_interleave_R1_R2.fastq
``` _Step2: Digital normalization_ 



Parameters: Python-3.6.3; -k 20 --> kmer size = 20bp; -C 30 --> maximal coverage; -N 4  -x 4e9 --> 16Gb

normalize-by-median.py -k 20 -C 30 -N 4 -x 4e9 file_interleave_R1_R2.fastq -o file_normalize_by_median_R1_R2.fastq




Parameters: Python-3.6.3

extract-paired-reads.py file_normalize_by_median_R1_R2.fastq -f --output-paired file_diginorm_paired --output-single file_diginorm_single

4.2.

Duplicates removing

Software

Value	Label
picardtools	NAME

Publication: “Picard Toolkit.” 2019. Broad Institute, GitHub Repository. http://broadinstitute.github.io/picard/; Broad Institute

Version: 2.18.2

Github: http://github.com/broadinstitute/picard

java -Xmx16g -jar picard.jar MarkDuplicates I=file_trimmed_sorted.bam O=file_trimmed_sorted_rmdup.bam CREATE_INDEX=true REMOVE_DUPLICATES=true M=file_output.metrics

4.3.

Variant Caller 1: GATK (Genome Analysis ToolKit)

Software

Value	Label
GATK	NAME
http://github.com/broadinstitute/gatk	LINK
4.1.7	VERSION

Publication: McKenna et al., 2010

Version: 4.0.11.1

Github: http://github.com/broadinstitute/gatk

Poplar genome: Populus trichocarpa v3.1

Citation

McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, Garimella K, Altshuler D, Gabriel S, Daly M, DePristo MA 2010 The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. http://doi.org/10.1101/gr.107524.110

## HaplotypeCaller
gatk --java-options "-Xmx16G" HaplotypeCaller -R genome.fa -I file_trimmed_sorted_rmdup.bam -ERC GVCF -O file_trimmed_sorted_rmdup.g.vcf
## GenomicsDBImport
gatk --java-options "-Xmx96G -Xms96G" GenomicsDBImport -V file1_trimmed_sorted_rmdup.g.vcf -V file2_trimmed_sorted_rmdup.g.vcf --genomicsdb-workspace-path my_database -L list_Chr+scaff.list --batch-size 50 -ip 500
## GenotypeGVCFs
gatk GenotypeGVCFs -R genome.fa -V gendb://my_database  -new-qual true -O all_trimmed_sorted_rmdup_gVCF_GATK.snps.indels.vcf

4.4.

GATK adjustments for Q. petraea Q. petraea

Version: GATK 3.8

Download: http://console.cloud.google.com/storage/browser/_details/gatk-software/package-archive/gatk/GenomeAnalysisTK-3.8-0-ge9d806836.tar.bz2;tab=live_object

Oak reference genome: Quercus robur Haplome V2.3

Parameters: java 1.8.0_72 ; HaplotypeCaller; GenotypeGVCFs

#HaplotypeCaller
GATK -R haplome_v2.3.fa -T HaplotypeCaller -nct 20 -I sample1_trimmed_vs_haploV23.bam -I sample2_trimmed_vs_haploV23.bam -I sample3_trimmed_vs_haploV23.bam -I sample4_trimmed_vs_haploV23.bam -I
 sample5_trimmed_vs_haploV23.bam -I sample6_trimmed_vs_haploV23.bam -I sample7_trimmed_vs_haploV23.bam -I sample8_trimmed_vs_haploV23.bam -I sample9_trimmed_vs_haploV23.bam -I sample9_trimmed_vs_haploV23.bam 
--emitRefConfidence GVCF -o gatk_nct20_slurm_1node-c20_snps.vcf

#GenotypeGVCFs
GATK -T GenotypeGVCFs -R haplome_v2.3.fa --variant sample1.vcf --variant sample2.vcf --variant sample3.vcf --variant sample4.
vcf --variant sample5.vcf --variant sample6.vcf --variant sample7.vcf --variant sample8.vcf --variant sample9.vcf --variant sample10.vcf -o gatk_all10samples_SNPs.vcf

4.5.

Variant Caller 2: samtools / bcftools

Software

Value	Label
SAMtools	NAME
Linux	OS_NAME
http://github.com/samtools/samtools	REPOSITORY
Wellcome Trust Sanger Institute	DEVELOPER
http://www.htslib.org/download/	LINK
1.8	VERSION

Publication: Danecek et al., 2021

Version: 1.8

Github: http://github.com/samtools/samtools

Poplar genome: Populus trichocarpa v3.1

Citation

Software

Value	Label
bcftools	NAME
http://github.com/samtools/bcftools	REPOSITORY
http://github.com/samtools/bcftools	LINK
1.12	VERSION

Publication: Li H, 2011

Version: 1.8

Github: http://github.com/samtools/bcftools

Citation

Li H 2011 A statistical framework for SNP calling, mutation discovery, association mapping and population genetical parameter estimation from sequencing data. http://doi.org/10.1093/bioinformatics/btr509

samtools mpileup -uf genome.fa mapping_file_sort_without_duplicate.bam | bcftools call -mv -Oz > file_bcftools_noduplicate.vcf.gz

4.6.

bcftools adjustments for Q. petraea Q. petraea

Oak genome : Q. robur haplome V2.3

bcftools version: 1.6

Download: http://sourceforge.net/projects/samtools/files/samtools/1.6/

4.7.

Variant Caller 3: FreeBayes

Software

Value	Label
freebayes	NAME
http://github.com/freebayes/freebayes	REPOSITORY
Garrison and Marth	DEVELOPER

Publication: Garrison and Marth, 2012

Version: 1.2.0-2

Github: http://github.com/freebayes/freebayes

Citation

Erik Garrison and Gabor Marth Haplotype-based variant detection from short-read sequencing arXiv preprint arXiv:1207.3907 [q-bio.GN] 2012 http://doi.org/10.48550/arXiv.1207.3907

Poplar genome: Populus trichocarpa v3.1

Oak genome : Q. robur haplome V2.3

freebayes -f genome.fa all_samples.bam > freebayes_all_samples.vcf

4.8.

SNP filtering

For poplar, we considered only biallelic intra-nigra SNPs with quality threshold ≥ 30.

Software

Value	Label
VCFtools	NAME
http://vcftools.github.io/man_latest.html	REPOSITORY
Adam Auton, Petr Danecek, Anthony Marcketta	DEVELOPER

Publication: Danecek et al., 2011

Version: 0.1.15

Github: http://vcftools.github.io/man_latest.html

Citation

Danecek P, Auton A, Abecasis G, Albers CA, Banks E, DePristo MA, Handsaker RE, Lunter G, Marth GT, Sherry ST, McVean G, Durbin R, 1000 Genomes Project Analysis Group 2011 The variant call format and VCFtools. http://doi.org/10.1093/bioinformatics/btr330

vcftools --vcf all_tool.snps.indels.vcf --out all_filtered_tool.vcf --remove-indels --max-alleles 2 --min-alleles 2 --minQ 30--recode --recode-INFO-all

```For oak, we considered bi-allelic SNPs, depth >= 20, maf >= 30% and <= 70%

4.9.

SNP identification

Only SNPs identified by at least 2 callers were selected to obtain the final set of SNPs.

Software

Value	Label
bcftools	NAME
http://github.com/samtools/bcftools	REPOSITORY
http://github.com/samtools/bcftools	LINK
1.12	VERSION

Publication: Danecek P, et al. 2021

Version: 1.8

Github: http://github.com/samtools/bcftools

Citation

Parameters: tabix-0.2.5, samtools-1.8, bcftools-1.8

bcftools index sample1_diginorm_gatk3.8_depth20_maf30.vcf.gz
bcftools index sample1_diginorm_FreeBayes_depth20_maf30.vcf.gz
bcftools index sample1_samtools_depth20_maf30.vcf.gz

bcftools isec -n +3 sample1_diginorm_gatk3.8_depth20_maf30.vcf.gz sample1_diginorm_FreeBayes_depth20_maf30.vcf.gz sample1_samtools_depth20_maf30.vcf.gz -O v -o common_SNPs_sample1_GATK_FreeBayes_samtools_depth20_maf30_bcftools.txt

Selection of C/T SNP

SMPs colocalizing with a C/T SNP (see the WGS and SNP detection section of the manuscript) will be removed at step #7 "SMPs filtering".

Whole Genome Bisulfite Sequencing - Detecting mC whole genome and Identifying outlier DMCs

Galaxy pipeline

SMPs were identified with the GALAXY (The Galaxy Community, 2022) pipeline (Dugé de Bernonville et al., 2022; Sow et al., 2023).

Citation

Dugé de Bernonville T, Daviaud C, Chaparro C, Tost J, Maury S 2022 From Methylome to Integrative Analysis of Tissue Specificity. http://doi.org/10.1007/978-1-0716-2349-7_16

Citation

Sow MD, Rogier O, Lesur I, Daviaud C, Mardoc E, Sanou E, Duvaux L, Civan P, Delaunay A, Lesage-Descauses MC, Benoit V, Le-Jan I, Buret C, Besse C, Durufle H, Fichot R, Le-Provost G, Guichoux E, Boury C, Garnier A, Senhaji-Rachik A, Jorge V, Ambroise C, Tost J, Plomion C, Segura V, Maury S, Salse J Epigenetic Variation in Tree Evolution: a case study in black poplar (Populus nigra) bioRxiv 2023.07.16.549253 http://doi.org/10.1101/2023.07.16.549253

Following Sow et al., 2023:

6.1.

Trimming

Software

Value	Label
TrimGalore	NAME
http://github.com/FelixKrueger/TrimGalore	REPOSITORY
Felix Krueger	DEVELOPER
http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/	LINK

Publication: Krueger F et al., 2023. FelixKrueger/TrimGalore: v0.4.3.1

Version: v0.4.3.1

Github: http://github.com/FelixKrueger/TrimGalore

Parameters: --paired read1.fastq read2.fastq --clip_R1 10 --clip_R2 30

Citation

Felix Krueger; Frankie James; Phil Ewels; Ebrahim Afyounian; Michael Weinstein; Benjamin Schuster-Boeckler;Gert Hulselmans; sclamons FelixKrueger/TrimGalore: v0.6.10 Zenodo http://doi.org/10.5281/zenodo.5127898

6.2.

Mapping

Software

Value	Label
BSMAP	NAME
http://github.com/genome-vendor/bsmap/	REPOSITORY

Publication: Xi Y and Li W, 2009

Version: v1.0.0Github: http://github.com/genome-vendor/bsmap/

Parameters: default options

Citation

Xi Y, Li W 2009 BSMAP: whole genome bisulfite sequence MAPping program. http://doi.org/10.1186/1471-2105-10-232

Poplar genome: Populus trichocarpa v3.1

Mapping adjustments for Q. petraea Q. petraea

Oak genome: Quercus robur Haplome V2.3

6.3.

Methylation calling (SMP)

Software

Value	Label
BSMAP methylation caller	NAME
Greg Zynda	DEVELOPER
http://github.com/zyndagj/BSMAPz	LINK
1.0.0	VERSION

Publication: Xi Y and Li W, 2009

Version: v1.0.0Github: http://github.com/genome-vendor/bsmap/

Citation

Xi Y, Li W 2009 BSMAP: whole genome bisulfite sequence MAPping program. http://doi.org/10.1186/1471-2105-10-232

Poplar genome: Populus trichocarpa v3.1

methratio.py   --ref ref_genome.fa --zero-meth TRUE --trim-fillin 2 --combine-CpG --min-depth 8 --context all  bsmap_sample*.sam

``` **Mapping adjustments for Q. petraea**  _Q. petraea_ 



Oak genome:  _Quercus robur_  Haplome V2.3

SMP filtering

Each methylation context (CpG, CHG, CHH) was considered separately.

Software

Value	Label
methylKit	NAME
http://github.com/al2na/methylKit/releases	REPOSITORY
Alexander Blume	DEVELOPER
http://github.com/al2na/methylKit/releases	LINK
0.99.2	VERSION

Publication: Akalin et al., 2012

Version: Methylkit R package v0.99.2

Github: http://github.com/al2na/methylKit/releases

Site: http://bioconductor.org/packages/release/bioc/html/methylKit.html

Parameters: R (v3.5.1), library(methylKit)

Citation

Akalin A, Kormaksson M, Li S, Garrett-Bakelman FE, Figueroa ME, Melnick A, Mason CE 2012 methylKit: a comprehensive R package for the analysis of genome-wide DNA methylation profiles. http://doi.org/10.1186/gb-2012-13-10-r87

Step1: Forward and reverse strands were merged for the CG context only and 30% missing data were tolerated for each context.

meth.CpG <- unite(CpG, destrand = TRUE, min.per.group = 7L)
meth.CHG <- unite(CHG, destrand = FALSE, min.per.group = 7L)
meth.CHH <- unite(CHH, destrand = FALSE, min.per.group = 7L)
```      **Step2:**  Positions corresponding to C/T SNPs were removed.

SNPdat <- read.delim("SNP_file.txt", header = F)

#with SNP_file.txt:

ScaffoldID position allele1 allele2

SNPdatScaff, SNPdatSNP <- paste(SNPdatAlt, sep ="/") MethPos2 <- paste(meth.CpG2start, sep = "_") MethPosMatchSNP2 <-which(MethPos2 %in% SNPdatScaff_Pos))

for (i in 1:19) { cov <- getData(meth.CHG.filtind.filtSNP.filtCov)[,colnames(meth.CHG.filtind.filtSNP.filtCov) == paste0("coverage", i)] cov_filt <- sort(c(which(cov < 7), which(is.na(cov)))) meth.CHG.filtind.filtSNP.filtCov[cov_filt, colnames(meth.CHG.filtind.filtSNP.filtCov) == paste0("numCs", i)] <- NA meth.CHG.filtind.filtSNP.filtCov[cov_filt, colnames(meth.CHG.filtind.filtSNP.filtCov) == paste0("numTs", i)] <- NA rm(cov, cov_filt) }

Identification of target regions for the SeqCapBis design

We first grouped SMPs into 1kb sliding windows of 250bp for each methylation context. Following the calculation of the methylation levels in each window, the outlier DMRs were identified using two strategies (see 8.2 and 8.3) with homemade scripts (given as examples). Finally, target sequences correspond to outlier DMRs identified by the two strategies.

8.1.

Grouping SMPs in windows and DMRs identification

Software

Value	Label
methylKit	NAME
http://github.com/al2na/methylKit/releases	REPOSITORY
Alexander Blume	DEVELOPER
http://github.com/al2na/methylKit/releases	LINK
0.99.2	VERSION

Publication: Akalin et al., 2012

Version: 1.18.0

Github: http://github.com/al2na/methylKit/releases

Site: http://bioconductor.org/packages/release/bioc/html/methylKit.html

Parameters: MethylKit package

Citation

Input files: pre-filtered SMPs in each context.

meth.CpG.window <-tileMethylCounts(meth.CpG.filtind.filtSNP.filtTE.filtCov.filtNA,win.size = 1000, step.size = 250)
meth.CHG.window <-tileMethylCounts(meth.CHG.filtind.filtSNP.filtTE.filtCov.filtNA,win.size = 1000, step.size = 250)
meth.CHH.window <-tileMethylCounts(meth.CHH.filtind.filtSNP.filtTE.filtCov.filtNA,win.size = 1000, step.size = 250)
```<Note title="Citation" type="info" >Akalin A, Kormaksson M, Li S, Garrett-Bakelman FE, Figueroa ME, Melnick A, Mason CE 2012 methylKit: a comprehensive R package for the analysis of genome-wide DNA methylation profiles. <a href="http://doi.org/10.1186/gb-2012-13-10-r87">http://doi.org/10.1186/gb-2012-13-10-r87</a></Note>

8.2.

Strategy I: STANDARD DEVIATION OF THE MEANS

Calculate average C-methylation by averaging the methylation level across all (pre-filtered) cytosines in each window for each individual. Then calculate standard deviation of this average across individuals.

#Identification of windows to remove
percmeth.CpG.window.sd <- rowSds(percmeth.CpG.window, na.rm = TRUE)
sum(percmeth.CpG.window.sd == 0)

# Removal of windows showing the less variable levels of methylation
percmeth.CpG.window <-percmeth.CpG.window[which(percmeth.CpG.window.sd != 0), ] dim(percmeth.CpG.window)

#Identification of the windows associated with the most variable methylation levels
percmeth.CpG.window.sd <- rowSds(percmeth.CpG.window, na.rm = TRUE)
layout(matrix(c(rep(1, 2), 2), nrow = 1))
hist(percmeth.CpG.window.sd, col = "grey", main = "")
bp <- boxplot(percmeth.CpG.window.sd, col = "grey")
length(bp$out)
bp$stats

8.3.

Strategy II: MEAN OF THE STANDARD DEVIATIONS

For each (pre-filtered) cytosine, calculate the standard deviation of methylation across individuals. Then calculate the mean standard deviation from all cytosines in a window.

dag_window_size=1000
dag_step=250

load("meth.CHG.filtind.filtSNP.filtTE.filtCov.filtNA.Rdata")
y<-x[,c("chr","start","end","strand")]

for (i in 1:length(colnames(x)[colnames(x) %like% "coverage"])){  # To recover the C/coverage values
  j=5+3*(i-1)
  print(paste0(j," ",j+1))    
  y[,paste0("in",i)]<-x[,j+1]/x[,j]
}
yy<-x[,c("chr","start","end","strand")]
rm(x)

z<-rowSds(as.matrix(y[,5:ncol(y)]),na.rm=TRUE) # Calculate row standard deviations
yy$STDEV<-z
rm(z)
y<-yy
rm(yy)


# Do last adaptations and launch
dag_window=dag_window_size/dag_step
colnames(y)<-c("CHR","START","END","STRAND","STDEV")
y$MEAN<-(y$START+y$END)/2
y$CHR<-gsub("Chr0","Chr",y$CHR,perl=TRUE)
y$WINDOW<-floor(y$MEAN/dag_step)+1


stdev_counts = data.table(
   CHR = character(),
   WIN = numeric(),
   POS = numeric(),
   STDEV = numeric()
)

count=0
for (i in unique(y[y$CHR %like% "Chr" | y$CHR %like% "scaffold",]$CHR)){
  window_size=dag_window_size
  step=dag_step
  #i<-paste0("Chr",i)
  z<-y[y$CHR==i,]
  min=0
  max=max(z$WINDOW)
  #print(paste(i,min,max,min(z$MEAN),max(z$MEAN)))
  count=count+1
  print(paste(i,min,max,min(z$MEAN),max(z$MEAN),count,length(unique(y[y$CHR %like% "Chr" | y$CHR %like% "scaffold",]$CHR))))
  zz<-data.frame(matrix(ncol=2,nrow=max*step))
  colnames(zz)<-c("MEAN","STDEV")
  zz$MEAN<-rownames(zz)
  
  zz[zz$MEAN %in% z$MEAN,]$STDEV<-z[z$MEAN %in% zz$MEAN,]$STDEV

 # Sliding window
  total <- nrow(zz)
  if (max(z$MEAN)<window_size){ # Adapted to avoid problems with scaffolds smaller than window_size
    spots <- 1
  }
  else {
    spots <- seq(from=1, to=(total-window_size), by=step)
  }
  
  if (spots[length(spots)]<=total-window_size){spots<-c(spots,(spots[length(spots)]+step))} # Adapted to recover the last bits inside smaller window
  result <- vector(length = length(spots))
  for(j in 1:length(spots)){
    if (j%%50000==0){print(paste(j,length(spots)))}
    if ((spots[j]+window_size)>=total){window_size=(total-spots[j])} # Adapted to recover the last bits inside last smaller window 
    result[j] <- mean(zz[spots[j]:(spots[j]+window_size-1),"STDEV"],na.rm=TRUE)
  }
  
  stdev_counts<-rbind(stdev_counts,data.frame(CHR=i,WIN=1:length(spots),POS=spots,STDEV=result))
}

x<-stdev_counts
write.table(x,file=paste0(save_file_name))

8.4.

Outlier threshold

The threshold for DMRs is defined as (Q3+1.5*(Q3-Q1)) where Q1 and Q3 are the first and third quartiles (i.e. the threshold is not defined by a percentile, but instead depends on the length of the boxplot box)

 *** Strategy I**

Parameters: Python 3.7

#$Id$

###run with python get_threshold_over_all_windows_calc1.py OUTPUT_FILE_from_calc1_get_mean_and_stdv_for_each_window.py > threshold_calc1.txt


import os
import re
import string
import sys
import glob
import numpy

file1 = sys.argv[1]
file1_stream = open(file1)
list_of_means = []

for line1 in file1_stream.readlines():
	if (line1.count('start') == 0):
		line1 = line1.replace('\n','')
		splitted_line1 = line1.split('\t')
		scaffold = splitted_line1[0]
		start = splitted_line1[1]
		end = splitted_line1[2]
		
		mean = splitted_line1[13]
		mean = float(mean)
		list_of_means.append(mean)

list_of_means.sort()
nbre_de_means = len(list_of_means)
##XXX corresponds to the first half of the dataset 
##YYY corresponds to the second half of the dataset
Q1 = numpy.median(list_of_means[:XXX])
Q3 = numpy.median(list_of_means[YYY:])

##for CHH context, hreshold = (Q3 + 3*(Q3- Q1))
threshold = (Q3 + 1.5*(Q3- Q1))
threshold = round(threshold,5)

print 'threshold = ',threshold

```     *** Strategy II** 



Parameters: Python 3.7

###run with python get_threshold_stdv_over_all_windows_calc2.py OUTPUT_FILE_from_get_stdv_between_individuals_for_each_window_calc2.py > threshold_calc2.txt

import os import re import string import sys import glob import numpy

file1 = sys.argv[1] file1_stream = open(file1) list_of_stdv = []

for line1 in file1_stream.readlines(): if (line1.count('start') == 0): line1 = line1.replace('\n','') splitted_line1 = line1.split('\t') scaffold = splitted_line1[0] start = splitted_line1[1] end = splitted_line1[2]

	stdv = splitted_line1[4]
	stdv = float(stdv)
	list_of_stdv.append(stdv)

list_of_stdv.sort() nbre_de_stdv = len(list_of_stdv) ##XXX corresponds to the first half of the dataset ##YYY corresponds to the second half of the dataset Q1 = numpy.median(list_of_stdv[:XXX]) Q3 = numpy.median(list_of_stdv[YYY:])

##for CHH context, hreshold = (Q3 + 3*(Q3- Q1)) threshold = (Q3 + 1.5*(Q3- Q1)) threshold = round(threshold,5)

print 'threshold = ',threshold

8.5.

Identification of capture targets

Target sequences correspond to outlier DMRs identified by the two strategies. This is a two-steps strategy where the 3 contexts are first merged and, then, sequence redundancy between the three methylation contexts is removed.

Software

Value	Label
bedtools	NAME
Linux	OS_NAME
http://ubuntu.pkgs.org/20.04/ubuntu-universe-amd64/bedtools_2.27.1+dfsg-4ubuntu1_amd64.deb.html	LINK
2.27.1	VERSION

Publication: Quinlan AR and Hall IM, 2010

Version: 2.27.1

Github: http://github.com/arq5x/bedtools2

Parameters: intersect, merge

Citation

Quinlan AR, Hall IM 2010 BEDTools: a flexible suite of utilities for comparing genomic features. http://doi.org/10.1093/bioinformatics/btq033

SeqCapBis - Detecting mC Target regions

Agilent Probes design and sequencing

A set of 120 bp probes was selected to capture 18 Mb of each genome (Agilent, http://earray.chem.agilent.com/suredesign/)..) The targeted regions corresponded to the regions identified as differentially methylated between populations. Custom targeted genome bisulfite sequencing was performed with SureSelect XT Methyl-Seq Target Enrichment (Agilent, Santa Clara, CA, USA) according to the manufacturer’s recommendations.

For poplar, in total, 17.84 Mb of sequence corresponding to the 25,434 DMRs was covered by 339,658 probes. Regarding oak, a set of 140,249 probes (120 bp) was designed by Agilent to cover 16.15 Mb DMRs.

10.

Trimming

Software

Value	Label
TrimGalore	NAME
Linux	OS_NAME
http://github.com/FelixKrueger/TrimGalore/releases	LINK
0.6.5	VERSION

Publication: Krueger F et al., 2023. FelixKrueger/TrimGalore: v0.6.5

Version: 0.6.5

Github: http://github.com/FelixKrueger/TrimGalore

Citation

trim_galore input_R1.fastq.gz input_R2.fastq.gz --paired ADAPTER1 -a2 ADAPTER2 -o output_directory --gzip -j {threads}

11.

Quality control

Software

Value	Label
FastQC	NAME
Linux	OS_NAME
Simon Andrews	DEVELOPER
http://www.bioinformatics.babraham.ac.uk/projects/fastqc/	LINK
0.11.9	VERSION

Publication: Andrews, S. (2010). FastQC: A Quality Control Tool for High Throughput Sequence Data [Online]. Available online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/

Version: 0.11.9

Github: http://github.com/s-andrews/FastQC

fastqc trimmed_reads.fq.gz -o fastQC_output_directory -t {threads}

12.

Mapping

Software

Value	Label
BsmapZ	NAME
Linux	OS_NAME
http://github.com/zyndagj/BSMAPz	LINK
1.1.3	VERSION

Publications:

Xi Y, Li W, 2009
Citation
Xi Y, Li W 2009 BSMAP: whole genome bisulfite sequence MAPping program. http://doi.org/10.1186/1471-2105-10-232
* Zynda G. 2018. BSMAPz. http://github.com/zyndagj/BSMAPz Version: 1.1.3

Github: http://github.com/zyndagj/BSMAPz

Poplar genome: Populus trichocarpa v4.1

bsmapz -a fileR1.fq.gz -b fileR2.fq.gz -o {output.out} -d mapped_file.bam -d ref_genome.fa -p threads
 
``` **Mapping adjustments for Q. petraea**  _Q. petraea_ 



Oak genome:  _Quercus robur_  Haplome V2.3

12.1.

Duplicate Removing

Software

Value	Label
samtools	NAME
Linux	OS_NAME
http://sourceforge.net/projects/samtools/files/samtools/1.11/	LINK
1.11	VERSION

Publication: Danecek et al., 2021

Version: 1.11

Github: http://github.com/samtools/samtools

Parameters: stat, fixmate, sort, markdup

Poplar genome: Populus trichocarpa v4.1

Citation

samtools stats sample_bsmapz_sorted.bam -r ref_genome.fa -@ {threads} > sample.statics
samtools fixmate -@ {threads} -O BAM -m  sample_bsmapz_sorted.bam  sample_fixmate.bam 
samtools sort -@ {threads} -O BAM sample_fixmate.bam  -o sample_fixmate_sort.bam 
samtools markdup -r  ref_genome.fa -@ {threads} -s -f sample.statics sample_fixmate_sort.bam sample_fixmate_sort_temp.bam
``` **Mapping adjustments for Q. petraea**  _Q. petraea_ 



Oak genome:  _Quercus robur_  Haplome V2.3

13.

Detection of methylated cytosines (mC)

Software

Value	Label
BsmapZ	NAME
Linux	OS_NAME
http://github.com/zyndagj/BSMAPz	LINK
1.1.3	VERSION

Publications:

Xi Y and Li W, 2009.
Citation
Xi Y, Li W 2009 BSMAP: whole genome bisulfite sequence MAPping program. http://doi.org/10.1186/1471-2105-10-232
* Zynda G. 2018. BSMAPz. http://github.com/zyndagj/BSMAPz Version: 1.1.3

Github: http://github.com/zyndagj/BSMAPz

Poplar genome: Populus trichocarpa v4.1

Parameters: methratio.py, python 2.7, samtools 1.11, pysam 0.16.0.1

python methratio.py sample.dedup.bam -o meth_sample.txt -d ref_genome.fa -N {threads} -I
``` **Mapping adjustments for Q. petraea**  _Q. petraea_ 



Oak genome:  _Quercus robur_  Haplome V2.3

14.

10X sequencing filtering

Software

Value	Label
methylKit	NAME
http://github.com/al2na/methylKit/releases	REPOSITORY
Alexander Blume	DEVELOPER
http://github.com/al2na/methylKit/releases	LINK
0.99.2	VERSION

Publication: Akalin A et al, 2012.

Version: 1.18.0

Parameters: MethylKit package

Github: http://github.com/al2na/methylKit/releases

Site: http://bioconductor.org/packages/release/bioc/html/methylKit.html

Citation

SeqCapBis_CHG = methRead( = path_to_the_files, sample.id = sample.ids, assembly = "quercus", mincov = 10, context = "CHG", treatment = rep(0,10))

15.

Splitting context

We set up a homemade bash script (splitting.sh) to obtain methylation files for each sample in the three contexts (CG, CHG and CHH).

#!/bin/bash 
# Splitting context:

usage()
{
cat << EOF
usage: $0 <options>
splitting context.
OPTION:
  -h      show this Help message.
  -o      Output.
  -i      Input.
EOF
}

# Get options
while getopts "ho:i:" OPTION
do
  case $OPTION in
    h)  usage; exit 1;;
    o)  output=$OPTARG;;
    i)  input=$OPTARG;;
    ?)  usage; exit;;
  esac
done

# Check that all options were passed
if [[ -z $output ]] || [[ -z $input ]] 
then
  printf "\n=========================\n ERROR: missing options\n=========================\n\n"
  usage
  exit 1
fi

#in_file = snakemake.input["isoforms"]
#out_file = snakemake.output["plot"]

# Fail on the first error
set -e

######################

file=$(echo $output|rev|cut -d "/" -f 1 |rev)
path=$(echo $output|rev|cut -d "/" -f 2- |rev)

for context in "CHH" "CG" "CHG"; do 
   
   awk "NR<=1 || \$4~/$context/" $input  > $path/$context-$file ;
done

16.

Methylation quantification

Software

Value	Label
methylKit	NAME
http://github.com/al2na/methylKit/releases	REPOSITORY
Alexander Blume	DEVELOPER
http://github.com/al2na/methylKit/releases	LINK
0.99.2	VERSION

Publication: Akalin A et al, 2012.

Version: 1.18.0

Parameters: MethylKit package

Github: http://github.com/al2na/methylKit/releases

Site: http://bioconductor.org/packages/release/bioc/html/methylKit.html

Citation

Functions: getMethylationStats(), getCoverageStats()

# Read methylation using methylkit function methRead
myobj <- methRead( = files, sample.id = sample_id, assembly = "populus tricharpa v3.1", mincov = 1, context = context,treatment = rep(0, length(files)), pipeline = list(fraction=TRUE, chr.col=1, start.col=2, end.col=2, coverage.col=6, strand.col=3, freqC.col=5 ))  

# Concatenate all samples tables into one unique table
finalFrame <- mergeMethylkitOutput(myobj)

#Write the final table as a csv2 file
write.csv2(finalFrame,file = table,)

# head(myobj)

# plots for statistcs and coverage simple :
pdf(file = XXX)
getMethylationStats(myobj[[1]],plot=TRUE,both.strands=FALSE)
getCoverageStats(myobj[[1]],plot=TRUE,both.strands=FALSE)
dev.off()

Transposon insertion polymorphisms (TIPs)

17.

Trimming

Eliminate unwanted or irrelevant parts of the read. Data trimming may include removing low quality bases or adapters used during sequencing.

Software

Value	Label
TrimGalore	NAME
Linux	OS_NAME
Felix Krueger	DEVELOPER
http://github.com/FelixKrueger/TrimGalore/releases/	LINK
0.6.5	VERSION

Citation

#Trim the paired sequences
trim_galore -q 30 --paired -o paired_1.fastq  paired_2.fastq

18.

Detection of TIPs on whole genome sequencing (WGS) data with TEFLoN

18.1.

Mapping

Alignment of DNA sequences to a reference genome.

Software

Value	Label
BWA	NAME
Linux	OS_NAME
Heng Li	DEVELOPER
http://github.com/lh3/bwa/releases/	LINK
0.7.17	VERSION

Citation

Heng Li; Richard Durbin Fast and accurate short read alignment with Burrows-Wheeler transform bioinformatics http://doi.org/10.1093/bioinformatics/btp324

#Index Genome
bwa index genome_ref.fa

#Align
bwa mem -Y genome_ref.fa  paired_trimmed_1.fastq  paired_trimmed_2.fastq > whole.sam

18.2.

Extracting unmapped reads

Search for TIPs from reads not aligning with the reference genome. It is interesting to choose non-mapped sequences, because we hypothesize that the insertion of a transposable element is one of the reasons which prevented the alignment of certain reads to their reference genome.

Software

Value	Label
samtools	NAME
http://github.com/samtools/samtools	REPOSITORY
http://github.com/samtools/samtools	LINK
1.12	VERSION

Citation

Petr Danecek, James K Bonfield, Jennifer Liddle, John Marshall, Valeriu Ohan, Martin O Pollard, Andrew Whitwham, Thomas Keane, Shane A McCarthy, Robert M Davies, Heng Li Twelve years of SAMtools and BCFtools GigaScience, Volume 10 http://doi.org/10.1093/gigascience/giab008

#From SAM2BAM
samtools view -S -b whole.sam -o whole.bam

#Extract Unmapped reads

#An unmapped read whose mate is mapped.
samtools view -u  -f 4 -F264 whole.bam  > tmps1.bam

#Both reads of the pair are unmapped
samtools view -u -f 12 -F 256 whole.bam > tmps2.bam

#merge
samtools merge unmapped.bam tmps1.bam tmps2.bam

Software

Value	Label
BamToFastq	NAME
Linux	OS_NAME
Maxime U Garcia	DEVELOPER
http://github.com/nf-core/bamtofastq/releases/	LINK
1.3.5	VERSION

Citation

Friederike Hanssen, SusiJo, Gisela Gabernet, Maxime U Garcia, Matilda Åslin, nf-core bot nf-core/bamtofastq: 2.1.0 Zenodo http://doi.org/10.5281/zenodo.4710628

#Extract the reads in FASTQ format (paired)
bamToFastq -bam unmapped.bam -fq1 unmapped_reads1.fastq -fq2 unmapped_reads2.fastq

18.3.

TIPs detection

Software

Value	Label
TEFLoN	NAME
Linux	OS_NAME
Jeffrey Adrion	DEVELOPER
http://github.com/jradrion/TEFLoN	LINK
0.4	VERSION

Citation

Adrion, J.R., M.J. Song, D.R. Schrider, M.W. Hahn, and S. Schaack Genome-wide estimates of transposable element insertion and deletion rates in Drosophila melanogaster Genome Biology and Evolution http://doi.org/10.1093/gbe/evx050

Software

Value	Label
RepeatMasker	NAME
Linux	OS_NAME
Robert Hubley	DEVELOPER
http://github.com/rmhubley/RepeatMasker	LINK
4.1.2	VERSION

WD="path/to/working/_directory"
PREFIX="prefix_you_want"

##For each samples
python teflon_prep_custom.py -wd ${WD}reference -g genome_ref -l path/to/TE_LIBRARY -p ${PREFIX}

bwa index ${WD}reference/${PREFIX}.prep_MP/${PREFIX}.mappingRef.fa

bwa mem -Y ${WD}reference/${PREFIX}.prep_MP/${PREFIX}.mappingRef.fa ${READS1} ${READS2} > ${WD}reference/${PREFIX}.sam

samtools view -Sb ${WD}reference/${PREFIX}.sam | samtools sort -o ${WD}reference/${PREFIX}.sorted.bam

samtools index ${WD}reference/${PREFIX}.sorted.bam

#Run Teflon
#For each samples
python teflon.v0.4.py -wd ${WD} -d ${WD}reference/${PREFIX}.prep_TF/ -s path/to/samples -i unique_ID -l1 family -l2 class

#Teflon collapse
##Only once
python teflon_collapse.py -wd ${WD} -d ${WD}reference/${PREFIX}.prep_TF/ -s path/to/samples -n1 minimum_reads_to_support_TE_in_one_sample -n2 minimum_reads_to_support_TE_in_all_samples

#Teflon Count
#For each samples
python teflon_count.py -wd ${WD} -d ${WD}reference/${PREFIX}.prep_TF/ -s path/to/samples -i unique_ID

#Teflon genotype
##Only once
python teflon_genotype.py -wd ${WD} -d ${WD}reference/${PREFIX}.prep_TF/ -s path/to/samples -dt pooled

19.

Detection of TIPs on whole genome bisulfite sequencing (WGBS) data with epiTEome

19.1.

Mapping and extracting unmapped reads

Alignment of DNA sequences to a reference genome. Search for TIPs from reads not aligning with the reference genome. We choose non-mapped sequences, because we hypothesize that the insertion of a transposable element is one of the reasons which prevented the alignment of certain reads to their reference genome.

Software

Value	Label
Bismark	NAME
Felix Krueger	DEVELOPER
http://github.com/FelixKrueger/Bismark/	LINK
0.19.0	VERSION

Citation

Felix Krueger, Simon R Andrews Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications Bioinformatics http://doi.org/10.1093/bioinformatics/btr167

bismark_genome_preparation --verbose genome_ref.fa

bismark --genome genome_ref.fa paired_trimmed_1.fastq  paired_trimmed_2.fastq  --un

19.2.

TIPs detection

Software

Value	Label
epiTEome	NAME
Josquin Daron	DEVELOPER
http://github.com/jdaron/epiTEome	LINK
1.0	VERSION

Citation

Josquin Daron & R. Keith Slotkin EpiTEome: Simultaneous detection of transposable element insertion sites and their DNA methylation levels Genome Biology http://doi.org/10.1186/s13059-017-1232-0

idxEpiTEome.pl -l 100 -gff genome_ref.gff -t /path/to/TE_LIBRARY -fasta genome_ref.fa

epiTEome.pl -gff genome_ref.gff -ref genome_ref.epiTEome.masked.fasta -un unmapped_reads.fastq -t /path/to/TE_LIBRARY

Bioinformatics manual for population epigenomics combining whole-genome and target genome sequencing

Abstract

Steps

Whole Genome Sequencing - Removing false C/T mutations

Whole Genome Bisulfite Sequencing - Detecting mC whole genome and Identifying outlier DMCs

ScaffoldID position allele1 allele2

SeqCapBis - Detecting mC Target regions

Transposon insertion polymorphisms (TIPs)

推荐阅读