config.yml

#####################
# Editing mandatory #
#####################

# Run details. These primarily determine where the output ends up (see output_root below)
sample: '25m'
stype: '1D'
version: 'v2'

# Path of the read (*.fasta) & sequencing summary files you want to analyze
input_fasta: test/reads.fasta
input_summary: test/read_summary.txt

# Path where you would like pipeline results to be written.
# Run outputs will end up in in the directory: <output_root>/<sample>/<stype>/<version>/
output_root: test/test-output/

max_threads: 32 # Maximum number of threads the pipeline will attempt to use
pre_filter: DTR # Only process reads with DTR (None: all reads, Virsorter: virsorter)
                #  virsorter is purely experimental at this point

# Medaka model depends on whether MinION or PromethION used for sequencing and which Guppy version
# was used for basecalling. Please see Medaka documentation for list of available models.
MEDAKA:
    model: 'r941_min_high_g303'
    threads: 4 # Should be << max_threads so that multiple genomes can be polished in parallel

KAIJU:
    db: /media/sfodata/databases/kaiju/nr_euk # dbs can be downloaded from http://kaiju.binf.ku.dk/server
    switch: '-a greedy -e 5' # max 5 substitutions and allows substitutions; slower but more sensitive than default params
    run: True                # Set to False to skip kaiju
                             # will also be skipped if DB is missing

# Specify which k-mer bin to skip for each sample/run/version. HDBSCAN assigns all unbinned reads to bin_id = -1,
# so this bin should always be skipped. But occasionally other 'bad' bins require skipping if they generate downstream 
# errors in the pipeline.
SKIP_BINS: 
    25m:
        1D: 
            v1: [-1]
            v2: [-1]

####################
# Editing optional #
####################

DTR_FIND:
    ovlp: 20 # Percent of read start/end to check for alignment
    chunksize: 5000

VIRSORTER:
    db: 1             # Either "1" (DEFAULT Refseqdb) or "2" (Viromedb)
    diamond: True     # False: blastp, True: diamond
    virome: False     # True: input is a virome, don't train on data
    prophage: False   # False: only use phage, True: use prohpage, too
    data_dir: '/mnt/arginine/galaxydata/seqdbs/virsorter-data'
                      # location of virsorter-data

KMER_CALC:
    k: 5      # k-mer size to use for k-mer binning
    cli: '-c' # Use raw k-mer counts instead of normalized k-mer frequencies (best for genome-spanning reads)

UMAP:
    min_rl: 5000       # Minimum readlength to consider in analysis
    max_cb_rl: 60000   # Maximum read length to include in colorbar for UMAP scatterplot annotated by read length
    min_d: 0.1         # min_dist paramter for UMAP
    neighbors: 15      # n_neighbors parameter for UMAP
    min_q: 9           # Filtering out low qscore reads prior to running UMAP improves resolution of bins
    bin_min_reads: 10  # Don't create a bin if the bin contains < bin_min_reads
    scatter_size: 7    # Marker size for UMAP scatter plots
    scatter_alpha: 0.2 # Marker alpha for UMAP scatter plots

BIN_AVA_ALIGN:
    MINIMAP2:
        cli: '-x ava-ont --no-long-join -r100'
        threads: 1
    CLUST:
        min_score_frac: 0.8
        min_reads: 5
        max_recursion: -1 # set to non-zero value to override python's max recursion depth

DTR_ALIGN:
    ovlp: 20   # Percent of read start/end to check for alignment
    threads: 1 # Single thread is optimum for aligning single read start/end

NUCMER:
    mincov: 98.0 # Genome-genome percent coverage threshold for deduplicating polished genomes
    minpct: 98.0 # Genome-genome percent identity thresholf for deduplicating polished genomes

MINIMAP2:
    STANDARD:
        switches: '-x map-ont'
        threads: 4

RACON:
    repeat: 3 # Number of iterations to run
    include-unpolished: True
    window-length: 500
    quality-threshold: 9
    error-threshold: 0.3
    match: 5
    mismatch: -4
    gap: -8
    threads: 4
    fragment-correction: True ## should not contain dual/self overlaps

PORECHOP:
    switches: '--no_split'

PRODIGAL:
    threads: 4

LIN_CONCAT:
    ovlp_len: 3000
    minlen: 15000