config.yaml

outpath:
    output

seed:
    42

samples:
   ERR3588905:
        reads: samples/ERR3588905.fastq # This is where you should put the reads OR
        SRA: ERR3588905                 #    add SRA ID here to for Snakemake to download the reads to the path specified by reads
        ref: drosophila_melanogaster  # The ref must match the refs fields below
        degradation_level: non        # The degradation_level must match the degradation_level fields below
        chroms: # specify the name of contigs to simulate; add All_chroms to include all chroms
            - All_chroms
        genes:  # specify the ENSEMBL ID or gene name to simulate; add All_genes to include all chroms
            - All_genes
   22Rv1:
        reads: samples/22Rv1.fastq
        ref: homo_sapiens
        degradation_level: non
        chroms:
            - All_chroms
        genes:
            - All_genes

exec:
    polyA      : scripts/polyA.py
    degrade    : scripts/degradation.py
    split_file : scripts/split_file.py
    badread    : extern/Badread/badread-runner.py

degradation_level:
    non:
        slope     : 0
        intercept : 0
    low:
        slope     : 5
        intercept : 400
    high:
        slope     : 25
        intercept : 300

polyA_config:
    mean   : 60
    stddev : 20
    mincut : 20
    maxcut : 150

refs:
    homo_sapiens:
        dna.fa    : http://ftp.ensembl.org/pub/release-97/fasta/homo_sapiens/dna/Homo_sapiens.GRCh38.dna.primary_assembly.fa.gz
        cdna.fa   : http://ftp.ensembl.org/pub/release-97/fasta/homo_sapiens/cdna/Homo_sapiens.GRCh38.cdna.all.fa.gz
        annot.gtf : http://ftp.ensembl.org/pub/release-97/gtf/homo_sapiens/Homo_sapiens.GRCh38.97.gtf.gz
        dna_contig_filter:
            - dna:scaffold # ignore dna contigs with this tag, useful to keep only fully assembled contigs
    drosophila_melanogaster:
        dna.fa    : http://ftp.ensembl.org/pub/release-97/fasta/drosophila_melanogaster/dna/Drosophila_melanogaster.BDGP6.22.dna.toplevel.fa.gz
        cdna.fa   : http://ftp.ensembl.org/pub/release-97/fasta/drosophila_melanogaster/cdna/Drosophila_melanogaster.BDGP6.22.cdna.all.fa.gz
        annot.gtf : http://ftp.ensembl.org/pub/release-97/gtf/drosophila_melanogaster/Drosophila_melanogaster.BDGP6.22.97.gtf.gz
        dna_contig_filter:
            - dna:scaffold 

badread:
    batches  : 64
    url      : https://github.com/baraaorabi/Badread.git
    tag      : v0.1.5