Skip to content

Latest commit

 

History

History
256 lines (227 loc) · 11.8 KB

README.md

File metadata and controls

256 lines (227 loc) · 11.8 KB
Raw

Pipelines Format data, plink in vcf, vcf in various format, summary statistics and change of build

Script : vcf_in_plink.nf

Requirement :

  • python3, plink, bcftools, bash, nextflow
  • singularity / dockers image : no test yet

what script done :

  • select for each chromosome on quality of imputation : min info
  • convert each vcf in plink
  • rename duplicate rs or . by chro
  • added cm in bim files if file genetic_map give in argumen,
  • merge all chromosome in plink format
  • give a report with analyse of frequencie and score

arguments :

  • file_listvcf : file contains each bgzip vcf files to merge, one by line [default : none]
  • min_scoreinfo : what score info minimum do you want : [default :0.6]
  • output_pat : pattern of output for bed final file [default : out]
  • output_dir : directory of output : [default : plink]
  • score_imp : header of score imputation [default : INFO], for score info depend of software used on software of imptuation
  • PBWT : INFO
  • Filters done by plink :
  • cut_maf : default 0
  • cut_hwe : default 0
  • cut_geno: default 0
  • cut_mind :
  • file to extract rsinfomation with position :
    • file_ref_gzip : must be in gzip example of file used : here
      • poshead_chro_inforef psotion of column chromosome in file [default : 0]
      • poshead_bp_inforef : position of column where bp in file [default : 1]
      • poshead_rs_inforef : position of column where rs in file [default : 2]
      • poshead_a1_inforef : position of column where rs in file [default : 3]
      • poshead_a2_inforef : position of column where rs in file [default : 4]
  • do_stat : by default true make stat using frequencies and score
  • statfreq_vcf :
  • pattern used in Info to computed frequencies ([default : "%AN %AC" with AN total and AC alternative number])
  • can be two value NAll NAlt, where frequencies computed as Nalt/NAll
  • can be one value frequencies
  • genetic_maps : genetics maps to added map in bim file, if not provided, map doesn't added in bim, must be not compressed :
  • file for hg19
  • file for hg17
  • file for hg18
  • file for hg38
  • memory and cpu :
  • max_plink_cores_merge, plink_mem_req_merge are now two news parameter can defined memory for merge process
  • plink_mem_req : memory used for plink, bcftools
  • other_mem_req : other memory
  • max_plink_cores : cpus for plink and bcftools
chr position COMBINED_rate(cM/Mb) Genetic_Map(cM)
1 55550 0 0
1 568322 0 0
1 568527 0 0

Example

Script : vcf_in_bgen.nf / vcf_in_bgen_merge.nf vcf_in_bgen_merge_chro.nf

three srategy of conversion :

  • vcf_in_bgen.nf : convert each vcf that was filter in bgen, without merge with one process
  • vcf_in_bgen_merge.nf :
    • vcf are filter and merge and merge and format in bgen
  • vcf_in_bgen_merge_chro.nf :
  • vcf are filter
  • vcf are format in bgen
  • bgen are merge

Requirement :

  • plink, bcftools, bash, qctools, nextflow
  • singularity / dockers image : no test yet

what script done :

  • Intial data : format of Sanger imputation format vcf file
  • select for each chromosome on quality of imputation : min info
  • convert each vcf in impute2 format used by boltlmmm
  • output for each chromosome is basename of initial files with .impute2.gz

arguments :

  • file_listvcf : file contains each bgzip vcf files to merge, one by line [default : none]
  • min_scoreinfo : what score info minimum do you want : [default :0.6]
  • output_dir : directory of output : [default : bgen]
  • qctoolsv2_bin : bin file for qctools
  • genotype_field : genotype field to transform [degault : GP]
  • bgen_type : bgen type see [qctool manual] (https://www.well.ox.ac.uk/~gav/qctool_v2/documentation/alphabetical_options.html) :
  • default bgen (other : "bgen_v1.2", "bgen_v1.1")
  • other_option : other option to give to qctools

for instance for bolt lmm format bgen must be :

~/nextflow ~/Travail/git/h3agwas/formatdata/vcf_in_bgen_merge.nf --file_listvcf listvcf --output_pat  exampledata2_imp --output_dir ./ -profile slurmSingularity -resume --bgen_type bgen_v1.2

Script : vcf_in_impute2.nf

Requirement :

  • plink, bcftools, bash, nextflow
  • singularity / dockers image : no test yet

what script done :

  • Intiial data : format of Sanger imputation format vcf file
  • select for each chromosome on quality of imputation : min info
  • convert each vcf in impute2 format used by boltlmmm
  • output for each chromosome is basename of initial files with .impute2.gz

arguments :

  • file_listvcf : file contains each bgzip vcf files to merge, one by line [default : none]
  • min_scoreinfo : what score info minimum do you want : [default :0.6]
  • output_dir : directory of output : [default : impute2]

format a gwas file format_gwasfile.nf

Requirement :

  • plink, bash, nextflow, python3 (library : panda)
  • singularity / dockers image : no test yet

what script done :

  • initial data of gwas format transform in other files
  • search rs on file to added new rs at each position (if not found add chro:pos)
  • added N and frequencies values if need and plink file gave
  • Change header, separator... etc

arguments :

  • file_gwas : one file gwas :

    • intial header of your file :
      • head_pval [optional]
      • head_freq [optional]
      • head_bp
      • head_chr
      • head_beta [optional]
      • head_se [optional]
      • head_A1 [optional]
      • head_A2 [optional]
      • head_N [optional]
    • sep separator default space or tab, [optional], for comma : COM, tabulation : TAB and space "WHI"

  • header of your output :

    • out_gc : prepared data for submission of gwas catalog
    • if not initialise : using output of your initial files
    • headnew_pval [optional]
    • headnew_freq [optional]
    • headnew_bp [optional]
    • headnew_chr [optional]
    • headnew_beta [optional]
    • headnew_se [optional]
    • headnew_A1 [optional]
    • headnew_A2 [optional]
    • headnew_N [optional]
    • sepout : separator for your summary stat output [optional : default " "]

  • file to extract rsinfomation with position :

    • file_ref_gzip : must be in gzip example of file used : here
      • poshead_chro_inforef psotion of column chromosome in file [default : 0]
      • poshead_bp_inforef : position of column where bp in file [default : 1]
      • poshead_rs_inforef : position of column where rs in file [default : 2]

  • others option :

    • added some N and frequencie in gwas file :
      • used plink information to compute freq and N and added in gwas file if head_N and/or head_freq not intialise
      • input_dir : plink directory
      • input_pat : plink basename
    • mem_req : memory request for processes>

Pipeline : convert position hg38 in hg19

nextflow run convert_posversiongenome.nf

what is doing?

  • if no file give download gwas catalog
  • extract positions of interest
  • used rs to search position see args file_ref_gzip
  • used crossmap to defined position s not found previously and strand : see bin_crossmap and data_crossmap
  • return file with new position

arguments

  • output_dir : direction of output [default : output]

  • output : output : [default : out]

  • file_toconvert : file to convert if empty download gwas catalog

    • link_gwas_cat : link to download gwas catalog [default : https://www.ebi.ac.uk/gwas/api/search/downloads/alternative ]
    • head_rs : head rs to file to convert [default SNPS (gwas catalog)]
    • head_bp : head bp to file to convert [default SNPS (gwas catalog)]
    • head_chro : head bp to file to convert [default SNPS (gwas catalog)]
    • sep : separator used TAB, SPACE, "," [default TAB] (not allowed : ;)

  • file to extract rsinfomation with position :

  • file_ref_gzip : must be in gzip example of file used : here

    • poshead_chro_inforef psotion of column chromosome in file [default : 0]
    • poshead_bp_inforef : position of column where bp in file [default : 1]
    • poshead_rs_inforef : position of column where rs in file [default : 2]

  • bin_crossmap : crossmap [default ~/.local/bin/CrossMap.py]

  • data_crossmap : data to convert [default : "" ]

###output ;

  • {out}.tsv : final file
  • {out}.multi.tsv : more that one position have been found
  • {out}.detail.tsv : file before cleaning
  • {out}.notfound.tsv : fileswhere position not found
  • folder datai : contains files contains files download
  • folder datatmp : contains temporary file (extract of rs file)

installation :

R : library pip3.6 install CrossMap --user pip3.6 install numpy==1.16.1 --user chmod +x ~/.local/bin/CrossMap.py

Pipeline : vcf_in_bimbam.nf

transform vcf in bimbam format after filters for quality. ###arguments

  • file_listvcf : file contains each bgzip vcf files to merge, one by line [default : none]
  • min_scoreinfo : what score info minimum do you want : [default :0.6]
  • output_dir : directory of output : [default : impute2]
  • genotype_field : genotype field in vcf file [default : GP]
  • qctoolsv2_bin : qctools v2 binary [default :qctool_v2]
  • bcftools_bin : bcftools bin [default : bcftools]

Prepare data for imputation : plk_in_vcf_imp.nf

argument :

  • input_dir
  • input_pat
  • output_dir : direction of output [default : output]
  • file to extract rsinfomation with position :
  • file_ref_gzip : must be in gzip example of file used : here
    • poshead_chro_inforef psotion of column chromosome in file [default : 0]
    • poshead_bp_inforef : position of column where bp in file [default : 1]
    • poshead_rs_inforef : position of column where rs in file [default : 2]
  • deleted_notref : deleted position s not found in file_ref_gzip
  • reffasta : fasta reference, if present do control of vcf file :
  • checkVCF.py
  • bcftools : used plugin of +fixref see BCFTOOLS_PLUGINS=bcftools/plugins/
  • michigan_qc : apply migigan qc [default 0]
  • see : prepare you data
  • dataref_michigan : data ref used by michigan, if empty dowload [default : ""]
  • ftp_dataref_michigan : dl data from michigan [default : ftp://ngs.sanger.ac.uk/production/hrc/HRC.r1-1/HRC.r1-1.GRCh37.wgs.mac5.sites.tab.gz]
  • bin_checkmich perl script [default : "HRC-1000G-check-bim.pl"]

General requirement

  • bcftools
  • plink
  • R
  • python
  • qctools (v2)
  • samtools
  • for control of vcf
  • checkVCF.py is present in binary of nextflow pipeline (https://github.com/zhanxw/checkVCF)
  • if michigan qc apply, need data set of michigan and perl script (see here)

Example

see h3agwas-example github

 nextflow run h3abionet/h3agwas/formatdata/vcf_in_plink.nf --file_listvcf utils/listvcf --output_pat  kgp_imputed --output_dir plink_imputed/   --reffasta utils_data/Homo_sapiens.GRCh37.dna.primary_assembly.fa.gz  -profile singularity