Snakemake-Aligner

A workflow to map long reads back to an assembly.

This workflow will:

1. Align reads via minimap2, pbmm2, or winnowmap.
2. Optionally, filter unmapped reads and non-primary alignments.

Getting Started

git clone git@github.com:logsdon-lab/Snakemake-Aligner.git
cd Snakemake-Aligner

Configuration

Input

Files can be passed multiple ways in the samples section of config.yaml:

Assemblies

By path.

samples:
-   name: "1"
    asm_fa: "1.fa"

By fofn.

samples:
-   name: "1"
    asm_fofn: "1.fofn"

By directory and file regex.

samples:
-   name: "1"
    asm_dir: "1/"
    asm_rgx: ".*\\.fa.gz$"

Reads

By fofn.

samples:
-   name: "1"
    read_fofn: "1.fofn"

By directory and file extension.

samples:
-   name: "1"
    read_dir: "1/"
    read_rgx: ".*\\.bam$"

By path.

samples:
-   name: "1"
    reads: [
        "1/r1.bam"
    ]

Configuration

General

General configuration can be filled in config.yaml:

# Aligner to use.
# Either "winnowmap", "minimap2", or "pbmm2".
aligner: "winnowmap"
# To override default aligner params.
aligner_opts: "--MD -ax map-pb"
# Output directory
output_dir: "results/align"
# Log directory
logs_dir: "logs/align"
# Benchmarks directory
benchmarks_dir: "benchmarks/align"
# Job resources. Memory in GB.
threads_aln: 8
mem_aln: 30G

Aligner default parameters:

• winnowmap
  • -y -a --eqx --cs -x map-pb -I8g
• pbmm2
  • --log-level DEBUG --preset SUBREAD --min-length 5000
• minimap2
  • -y -a --eqx --cs -x map-hifi -I8g

To keep tags of BAM files:

samples:
-   name: "1"
    keep_tags: ["ML", "MM"]

# Or globally across all samples:
keep_tags: ["ML", "MM"]

To keep reads greater than some length:

samples:
-   name: "1"
    min_read_length: 30000

# Or globally across all samples:
min_read_length: 30000

To output CRAM files:

output_format: cram

If output BAM files and need csi index.

use_bam_csi: true

Usage

snakemake -np -c 1 --configfile config/config.yaml

Module

To incorporate this into a workflow.

SAMPLE_NAMES = ["sample_1"]
CFG = {
    "samples": [
        {
            "name": sm,
            "asm_fa": f"{sm}.fa",
            "read_dir": f"reads/{sm}/",
            "read_ext": "bam",
        }
        for sm in SAMPLE_NAMES
    ],
    **config["align"]
}


module Align:
    snakefile:
        github(
            "logsdon-lab/Snakemake-Aligner",
            path="workflow/Snakefile",
            branch="main"
        )
    config: CFG

use rule * from Align as align_*

rule all:
    input:
        expand(rules.align.input, sm=SAMPLE_NAMES),

Test

To run the dry-run workflow. Workflow with real files is a WIP.

snakemake --configfile test/config.yaml -c 1 -np