winnowmap/winnowmap_align

Alignment

Long-read

Mapping

Repetitive regions

Winnowmap

Description

Map long reads to a reference genome using Winnowmap, a long-read aligner
optimised for repetitive regions such as centromeres and segmental duplications.

Winnowmap extends the minimap2 framework with weighted minimizers, which
down-weight frequently occurring k-mers so that mappings in repeat-rich regions
are more accurate. K-mer frequencies are pre-computed with
meryl; if no pre-computed file is provided
this component will run meryl automatically, in which case --kmer_size must
be given (meryl has no default k, and winnowmap's -k has to match it).

By default the component outputs alignments in SAM format. Use --bam to
produce a coordinate-sorted, indexed BAM file; samtools is bundled in the
engine image, so no extra setup is needed.

Two winnowmap options are deliberately not exposed: -a (SAM output) is always
passed, which makes the PAF-only -c meaningless, and -d (dump index and
exit) is an index-building mode rather than an alignment mode.

Inputs

Name	Type & Properties	Description
--reference	file required	Reference genome in FASTA format. Used both for k-mer frequency counting (meryl) and for read alignment (winnowmap).
--query -q	file required	Query sequences in FASTQ or FASTA format. Compressed files (`.gz`) are supported. Supported formats: - FASTQ (`.fastq`, `.fq`, `.fastq.gz`, `.fq.gz`) - FASTA (`.fasta`, `.fa`, `.fasta.gz`, `.fa.gz`)
--repetitive_kmers -W	file	Pre-computed repetitive k-mers file generated by `meryl`. If omitted, this component will run the following meryl commands automatically, and `--kmer_size` then becomes required: ```bash meryl count k=<kmer_size> output merylDB <reference> meryl print greater-than distinct=0.9998 merylDB > repetitive_kmers.txt ``` The `distinct=0.9998` threshold (top 0.02% most frequent k-mers) is the one prescribed by [Winnowmap's own documentation](https://github.com/marbl/Winnowmap#usage). Providing a pre-computed file avoids repeated meryl runs when aligning multiple query files against the same reference.

Name

Type & Properties

--reference

file

required

--query

-q

file

required

--repetitive_kmers

-W

file

Outputs

Name	Type & Properties	Description
--bam	boolean_true	Output a coordinate-sorted and indexed BAM file instead of SAM. When enabled the component pipes winnowmap output through `samtools sort` and then runs `samtools index` to create a `.bai` index. `samtools` ships in the engine image, so this needs no extra setup. Use `--output_index` to control where the index is written. The output is written as BAM regardless of the `--output` file extension.
--output -o	file required output	Path for the output alignment file. - SAM format (default): e.g. `alignment.sam` - BAM format (with `--bam`): e.g. `alignment.bam`
--output_index	file output	Path for the BAM index. Only produced together with `--bam`; ignored otherwise. If omitted, the index is written next to `--output` as `<output>.bai`. Declaring it explicitly is what makes the index a tracked output, so it is published by the Nextflow runner instead of being left behind in the work directory.

Name

Type & Properties

--bam

boolean_true

--output

-o

file

required

output

--output_index

file

output

Preset

Name	Type & Properties	Description
--preset -x	string	Alignment preset for the sequencing technology, passed to winnowmap `-x`. \| Preset \| Technology \| \|---\|---\| \| `map-ont` \| Oxford Nanopore genomic reads \| \| `map-pb` \| PacBio HiFi / CCS genomic reads \| \| `map-pb-clr` \| PacBio CLR genomic reads (sets `--sv-off`) \| \| `splice` \| Long-read spliced alignment (sets `--sv-off`) \| \| `splice:hq` \| PacBio CCS spliced alignment (sets `--sv-off`) \| \| `asm5` \| Assembly-to-reference; ~0.1% divergence \| \| `asm10` \| Assembly-to-reference; ~1% divergence \| \| `asm20` \| Assembly-to-reference; ~5% divergence \| Winnowmap's preset names are not identical to minimap2's: there is no `map-hifi` (HiFi reads use `map-pb`, which winnowmap's own help documents as "hifi-to-ref"), CLR reads use `map-pb-clr`, and there is no short-read `sr` preset. Like winnowmap itself, no preset is applied by default. The asm presets set their own k (19). Any option set explicitly on this component still wins over the preset's value for it, so `--kmer_size` stays in step with the `-W` k-mer list whichever preset is chosen.

Name

Type & Properties

--preset

-x

string

Indexing options

Name	Type & Properties	Description
--kmer_size -k	integer	K-mer size, no larger than 28. Winnowmap's own default is 15. Required when `--repetitive_kmers` is not given: the component then builds the k-mer list with `meryl`, which needs an explicit k, and the same k is passed to winnowmap's `-k`. Winnowmap refuses to run when the `-W` list was built with a different k than it is using (`input list of k-mers and winnowmap parameter k are inconsistent`). When passing a pre-computed `--repetitive_kmers` file, set this to the k that file was built with. If left unset there, winnowmap applies its own default k and validates the list against it.
--window_size -w	integer	Minimizer window size. Default is 50.
--homopolymer_compressed -H	boolean_true	Use homopolymer-compressed k-mers.
--split_index -I	string	Split the index for every ~NUM input bases. Default is `4G`.

Name

Type & Properties

--kmer_size

-k

integer

--window_size

-w

integer

--homopolymer_compressed

-H

boolean_true

--split_index

-I

string

Mapping options

Name	Type & Properties	Description
--filter_fraction -f	double	Filter out the top FLOAT (<1) fraction of repetitive minimizers. Default is 0.0, since winnowmap down-weights repetitive minimizers via `-W` rather than discarding them.
--max_chain_gap -g	integer	Stop chain elongation if there are no minimizers within this many bp. Default is 5000.
--max_intron_length -G	string	Maximum intron length, effective with `--preset splice`; changes `--bandwidth`. Default is `200k`.
--max_fragment_length -F	string	Maximum fragment length in fragment mode. Default is `800`.
--bandwidth -r	integer	Bandwidth used in chaining and DP-based alignment. Default is 500.
--min_minimizers -n	integer	Minimal number of minimizers on a chain. Default is 3.
--min_chaining_score -m	integer	Minimal chaining score, matching bases minus the log gap penalty. Default is 40.
--skip_self_dual -X	boolean_true	Skip self and dual mappings, for the all-vs-all mode.
--secondary_ratio -p	double	Minimal secondary-to-primary score ratio. Default is 0.8.
--sv_off	boolean_true	Turn off SV-aware mode (`--sv-off`). Some presets set this themselves, see `--preset`.

Name

Type & Properties

--filter_fraction

-f

double

--max_chain_gap

-g

integer

--max_intron_length

-G

string

--max_fragment_length

-F

string

--bandwidth

-r

integer

--min_minimizers

-n

integer

--min_chaining_score

-m

integer

--skip_self_dual

-X

boolean_true

--secondary_ratio

-p

double

--sv_off

boolean_true

Alignment options

Name	Type & Properties	Description
--match_score -A	integer	Matching score. Default is 2.
--mismatch_penalty -B	integer	Mismatch penalty. Default is 4.
--gap_open_penalty -O	string	Gap open penalty, as `INT` or `INT,INT` for the two-piece affine gap cost. Default is `4,24`.
--gap_extension_penalty -E	string	Gap extension penalty, as `INT` or `INT,INT`. A k-long gap costs `min{O1+kE1,O2+kE2}`. Default is `2,1`.
--zdrop -z	string	Z-drop score and inversion Z-drop score, as `INT` or `INT,INT`. Default is `400,200`.
--min_peak_score -s	integer	Minimal peak DP alignment score. Default is 80.
--splice_strand -u	string	How to find GT-AG splice sites. \| Value \| Meaning \| \|---\|---\| \| `f` \| Transcript strand \| \| `b` \| Both strands \| \| `n` \| Do not match GT-AG (default) \|

Name

Type & Properties

--match_score

-A

integer

--mismatch_penalty

-B

integer

--gap_open_penalty

-O

string

--gap_extension_penalty

-E

string

--zdrop

-z

string

--min_peak_score

-s

integer

--splice_strand

-u

string

Input/output options

Name	Type & Properties	Description
--read_group -R	string	SAM read group line, in a format like `'@RG\tID:foo\tSM:bar'`. Required by most downstream variant callers.
--cigar_bam -L	boolean_true	Write CIGARs with more than 65535 operations to the `CG` tag. BAM's `n_cigar_op` field is 16 bit, so without this flag long reads spanning centromeric or segmental-duplication repeats can produce alignments that cannot be converted to BAM (`CIGAR has more than 65535 operations`).
--md_tag	boolean_true	Output the `MD` tag (`--MD`).
--cs_tag	string	Output the `cs` tag in the given form (`--cs`). Not emitted by default.
--eqx	boolean_true	Write `=`/`X` CIGAR operators instead of `M` (`--eqx`).
--soft_clipping -Y	boolean_true	Use soft clipping for supplementary alignments.
--minibatch_size -K	string	Minibatch size for mapping. Default is `1000M`.

Name

Type & Properties

--read_group

-R

string

--cigar_bam

-L

boolean_true

--md_tag

boolean_true

--cs_tag

string

--eqx

boolean_true

--soft_clipping

-Y

boolean_true

--minibatch_size

-K

string

Run this component

Run the following command to execute this component with Nextflow:

cat > params.yaml <<'EOM'  
output: "$id.$key.output.sam"  
output_index: "$id.$key.output_index.bai"  
id: "run"  
publish_dir: "output/"  
EOM

nextflow run https://packages.viash-hub.com/vsh/biobox.git \  
  -revision v0.5.0 \  
  -main-script target/nextflow/winnowmap/winnowmap_align/main.nf \  
  -params-file params.yaml

winnowmap/winnowmap_align

Inputs

Outputs

Preset

Indexing options

Mapping options

Alignment options

Input/output options

Run this component

Relationships

Used by

Uses