CONFIG.yaml

data_dir: "data"
results_dir: "results"
db_dir: "dbs"
runs:
    first: "20181106_1450_noselection_sizeselection"
    second: "20181107_0906_same"
    third: "20181108_0827_test"
barcodes: ["barcode06", "barcode07", "barcode08", "barcode09", "barcode10"]
assemblers: ["flye"]
p7zip:
    bin: "/home/users/claczny/apps/software/p7zip_16.02/bin/7za"
    threads: 4
ont_fast5_api:
    single_to_multi_fast5:
        bin: "single_to_multi_fast5"
        batch: 8000
        threads: 8
flowcell: "FLO-MIN106"
kit: "SQK-LSK108"
#barcodes: ["barcode06", "barcode07", "barcode08", "barcode09", "barcode10"]
barcodes: ["barcode07"]
guppy_cpu:
    path: "/scratch/users/claczny/ont/apps/software/ont-guppy-cpu-3.1.5_linux64/bin"
    bin: "/scratch/users/claczny/ont/apps/software/ont-guppy-cpu-3.1.5_linux64/bin/guppy_basecaller"
    version: "cpu-3.1.5"
    config: "dna_r9.4.1_450bps_modbases_dam-dcm-cpg_hac.cfg"
    cpu_threads: 28
guppy_gpu:
    path: "/home/users/sbusi/apps/ont-guppy/bin"
    bin: "set +u; source ~/.bashrc; set -u; ml compiler/LLVM system/CUDA && /home/users/sbusi/apps/ont-guppy/bin/guppy_basecaller"
    version: "3.4.5+fb1fbfb"
    config: "dna_r9.4.1_450bps_modbases_dam-dcm-cpg_hac.cfg"
    hac_config: "dna_r9.4.1_450bps_hac.cfg"
    records_per_fastq: 8000
    chunk_size: 1000
    chunks_per_runner: 1000
    num_callers: 4
    runners_per_device: 2
    gpu_device: "cuda:0"
    cpu_threads: 28
guppy_barcoder:
    path: "/home/users/sbusi/apps/ont-guppy/bin"
    bin: "set +u; source ~/.bashrc; set -u; ml compiler/LLVM system/CUDA && /home/users/sbusi/apps/ont-guppy/bin/guppy_barcoder"
    version: "3.4.5+fb1fbfb"
    records_per_fastq: 8000
    threads: 8
nanostats:
#short_reads_prefix: "/scratch/users/claczny/ont/fecal_pilot/data/raw/short_reads"
short_reads_prefix: "/mnt/isilon/projects/lcsb_sequencing/transfer/bioecosystem/Rashi/2019/Apr/fastq"
metaT_prefix: "/scratch/users/sbusi/ONT/cedric_ont_basecalling/metaT/2018_GDB"
#samples: ["Kapa1_MG_S18", "Kapa2_MG_S19", "NEB1_MG_S16", "NEB2_MG_S17", "NEBmod1_MG_Rashi_S22", "NEBmod2_MG_Rashi_S23"]
fastp:
    min_length: 40
minimap2:
    threads: 16
igc:
    uri: "parrot.genomics.cn/gigadb/pub/10.5524/100001_101000/100064/1.GeneCatalogs/IGC.fa.gz"
hg38:
    uri: "ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/001/405/GCF_000001405.38_GRCh38.p12/GCF_000001405.38_GRCh38.p12_genomic.fna.gz"
genomecov:
    bin: "bedtools genomecov"
compute_avg_coverage:
    bin: "scripts/coverage.awk"
bwa:
    threads: 24
    long_reads_index:
        opts: "-aY -A 5 -B 11 -O 2,1 -E 4,3 -k 8 -W 16 -w 40 -r 1 -D 0 -y 20 -L 30,30 -T 2.5"
samtools:
    sort:
        threads: 4