-
Thomas Weber authored
Split into files ; define modes & associated conda envs ; run on a single chromosome by keeping full initial BAM ; functionnal until strandphaser
Thomas Weber authoredSplit into files ; define modes & associated conda envs ; run on a single chromosome by keeping full initial BAM ; functionnal until strandphaser
config.yaml 2.66 KiB
mode: "count"
plot: False
check_sm_tag_in_bam_header: "False"
config_df_location: "config/config_df.tsv"
# Input BAM location
input_bam_ena: "/g/korbel2/weber/MosaiCatcher_files/bam_KG_full/"
input_bam_location: "/g/korbel2/weber/MosaiCatcher_files/bam_KG_full/"
output_location: "/g/korbel2/weber/MosaiCatcher_output/MosaiCatcher_output_sample_KG_dryrun/"
snv_sites_to_genotype: "/g/korbel2/weber/MosaiCatcher_files/snv_sites_to_genotype/ALL.chr1-22plusX_GRCh38_sites.20170504.renamedCHR.vcf.gz"
# FIXME : TMP debuging
# snv_sites_to_genotype: "/g/korbel2/weber/MosaiCatcher_files/snv_sites_to_genotype/split_by_chr/1000G_chr21.vcf.gz"
reference: "/g/korbel2/weber/MosaiCatcher_files/refgenomes_human_local/GCA_000001405.15_GRCh38_no_alt_analysis_set.fna"
# reference: "/g/korbel2/weber/MosaiCatcher_files/refgenomes_human_local/chr21.fna"
# TODO : add a SMK rule to clone & compile mosaic
mosaicatcher: "/g/korbel2/weber/Gits/mosaicatcher/build/mosaic"
# FIXME : TMP solution for strandphaser location
strandphaser: "/home/tweber/.conda/envs/strandseqnation/lib/R/library/StrandPhaseR/R/StrandPhaseR"
Rscript: "/home/tweber/.conda/envs/strandseqnation/bin/Rscript"
# TODO : remove hard coded paths for R scripts coming from MC
plot_script: "/g/korbel2/weber/Gits/mosaicatcher/R/qc.R"
sv_plot_script: "/g/korbel2/weber/Gits/mosaicatcher/R/chrom.R"
norm_script: "/g/korbel2/weber/Gits/mosaicatcher/R/norm.R"
# CHECKME : check implicit natural path through Conda env
samtools: "/g/funcgen/bin/samtools"
# bcftools: "/biolo/ngs/bin/bcftools"
# TODO : path reference
variable_bins:
"50000": "/g/korbel2/weber/workspace/pipeline/utils/variable_bins.GRCh38.50kb.bed"
"100000": "/g/korbel2/weber/workspace/pipeline/utils/variable_bins.GRCh38.100kb.bed"
# TODO : find another solution with conda for R_ref
R_reference: "BSgenome.Hsapiens.UCSC.hg38"
# CHECKME : Parameters optimization ?
exclude_list:
- BM510x3PE20490
chromosomes:
- chr21
# chromosomes:
# - chr1
# - chr2
# - chr3
# - chr4
# - chr5
# - chr6
# - chr7
# - chr8
# - chr9
# - chr10
# - chr11
# - chr12
# - chr13
# - chr14
# - chr15
# - chr16
# - chr17
# - chr18
# - chr19