How to use Snakemake container for htslib (bgzip + tabix) - docker

Question

I have a pipeline which uses a global singularity image and rule-based conda wrappers.
However, some of the tools don't have wrappers (i.e. htslib's bgzip and tabix).
Now I need to learn how to run jobs in containers.
In the official documentation link it says:
"Allowed image urls entail everything supported by singularity (e.g., shub:// and docker://)."
Now I've tried the following image from singularity hub but I get an error:
minimal reproducible example:
config.yaml
# Files
REF_GENOME: "c_elegans.PRJNA13758.WS265.genomic.fa"
GENOME_ANNOTATION: "c_elegans.PRJNA13758.WS265.annotations.gff3"
Snakefile
# Directories------------------------------------------------------------------
configfile: "config.yaml"
# Setting the names of all directories
dir_list = ["REF_DIR", "LOG_DIR", "BENCHMARK_DIR", "QC_DIR", "TRIM_DIR", "ALIGN_DIR", "MARKDUP_DIR", "CALLING_DIR", "ANNOT_DIR"]
dir_names = ["refs", "logs", "benchmarks", "qc", "trimming", "alignment", "mark_duplicates", "variant_calling", "annotation"]
dirs_dict = dict(zip(dir_list, dir_names))
GENOME_INDEX=config["REF_GENOME"]+".fai"
VEP_ANNOT=config["GENOME_ANNOTATION"]+".gz"
VEP_ANNOT_INDEX=config["GENOME_ANNOTATION"]+".gz.tbi"
# Singularity with conda wrappers
singularity: "docker://continuumio/miniconda3:4.5.11"
# Rules -----------------------------------------------------------------------
rule all:
input:
expand('{REF_DIR}/{GENOME_ANNOTATION}{ext}', REF_DIR=dirs_dict["REF_DIR"], GENOME_ANNOTATION=config["GENOME_ANNOTATION"], ext=['', '.gz', '.gz.tbi']),
expand('{REF_DIR}/{REF_GENOME}{ext}', REF_DIR=dirs_dict["REF_DIR"], REF_GENOME=config["REF_GENOME"], ext=['','.fai']),
rule download_references:
params:
ref_genome=config["REF_GENOME"],
genome_annotation=config["GENOME_ANNOTATION"],
ref_dir=dirs_dict["REF_DIR"]
output:
os.path.join(dirs_dict["REF_DIR"],config["REF_GENOME"]),
os.path.join(dirs_dict["REF_DIR"],config["GENOME_ANNOTATION"]),
os.path.join(dirs_dict["REF_DIR"],VEP_ANNOT),
os.path.join(dirs_dict["REF_DIR"],VEP_ANNOT_INDEX)
resources:
mem=80000,
time=45
log:
os.path.join(dirs_dict["LOG_DIR"],"references","download.log")
singularity:
"shub://biocontainers/tabix"
shell: """
cd {params.ref_dir}
wget ftp://ftp.wormbase.org/pub/wormbase/releases/WS265/species/c_elegans/PRJNA13758/c_elegans.PRJNA13758.WS265.genomic.fa.gz
bgzip -d {params.ref_genome}.gz
wget ftp://ftp.wormbase.org/pub/wormbase/releases/WS265/species/c_elegans/PRJNA13758/c_elegans.PRJNA13758.WS265.annotations.gff3.gz
bgzip -d {params.genome_annotation}.gz
grep -v "#" {params.genome_annotation} | sort -k1,1 -k4,4n -k5,5n -t$'\t' | bgzip -c > {params.genome_annotation}.gz
tabix -p gff {params.genome_annotation}.gz
"""
rule index_reference:
input:
os.path.join(dirs_dict["REF_DIR"],config["REF_GENOME"])
output:
os.path.join(dirs_dict["REF_DIR"],GENOME_INDEX)
resources:
mem=2000,
time=30,
log:
os.path.join(dirs_dict["LOG_DIR"],"references", "faidx_index.log")
wrapper:
"0.64.0/bio/samtools/faidx"
Error
Building DAG of jobs...
Pulling singularity image shub://biocontainers/tabix.
WorkflowError:
Failed to pull singularity image from shub://biocontainers/tabix:
ESC[31mFATAL: ESC[0m While pulling shub image: failed to get manifest for: shub://biocontainers/tabix: the requested manifest was not found in singularity hub
File "/home/moldach/anaconda3/envs/snakemake/lib/python3.7/site-packages/snakemake/deployment/singularity.py", line 88, in pull
~
It appears this is a problem with the container?
(snakemake) [moldach#arc CONTAINER_TROUBLESHOOT]$ singularity pull shub://biocontainers/tabix
FATAL: While pulling shub image: failed to get manifest for: shub://biocontainers/tabix: the requested manifest was not found in singularity hub
In fact, I experience this problem with other biocontainers containers.
For example, I also need to use a container to do bowtie2 indexing and this is the error I get from the biocontainers/bowtie2 versus another developers container of the same tool comics/bowtie2:
^C(snakemake) [moldach#arc CONTAINER_TROUBLESHOOT]$ singularity pull docker://biocontainers/bowtie2
FATAL: While making image from oci registry: failed to get checksum for docker://biocontainers/bowtie2: Error reading manifest latest in docker.io/biocontainers/bowtie2: manifest unknown: manifest unknown
(snakemake) [moldach#arc CONTAINER_TROUBLESHOOT]$ singularity pull docker://comics/bowtie2
INFO: Converting OCI blobs to SIF format
INFO: Starting build...
Getting image source signatures
Copying blob a02a4930cb5d done
Does anyone know why?

Answer 1

Biocontainers does not allow latest as tag for their containers, and therefore you will need to specify the tag to be used.
From their doc:
The BioContainers community had decided to remove the latest tag. Then, the following command docker pull biocontainers/crux will fail. Read more about this decision in Getting started with Docker
When no tag is specified, it defaults to latest tag, which of course is not allowed here. See here for bowtie2's tags. Usage like this will work:
singularity pull docker://biocontainers/bowtie2:v2.4.1_cv1

Answer 2

Using another container solves the issue; however, the fact I'm getting errors from biocontainers is troubling given that these are both very common and used as examples in the literature so I will award the top-answer to whomever can solve that specific issue.
As it were, the use of stackleader/bgzip-utility solve the issue of actually running this rule in a container.
container:
"docker://stackleader/bgzip-utility"
Once again, for those coming to this post, it's probably best to test any container first before running snakemake, e.g. singularity pull docker://stackleader/bgzip-utility.