Reputation: 747
I am new to snakemake, and I am using it to merge steps from two pipelines into a single larger pipeline. The issue that several steps create similarly named files, and I cannot find a way to limit the wildcards, so I am getting a Missing input files for rule
error on one step that I just cannot resolve.
The full snakefile is long, and is available here: https://mdd.li/snakefile
The relevant sections are (sections of the file are missing below):
wildcard_constraints:
sdir="[^/]+",
name="[^/]+",
prefix="[^/]+"
# Several mapping rules here
rule find_intersecting_snps:
input:
bam="hic_mapping/bowtie_results/bwt2/{sdir}/{prefix}_hg19.bwt2pairs.bam"
params:
hornet=os.path.expanduser(config['hornet']),
snps=config['snps']
output:
"hic_mapping/bowtie_results/bwt2/{sdir}/{prefix}_hg19.bwt2pairs.remap.fq1.gz",
"hic_mapping/bowtie_results/bwt2/{sdir}/{prefix}_hg19.bwt2pairs.remap.fq2.gz",
"hic_mapping/bowtie_results/bwt2/{sdir}/{prefix}_hg19.bwt2pairs.keep.bam",
"hic_mapping/bowtie_results/bwt2/{sdir}/{prefix}_hg19.bwt2pairs.to.remap.bam",
"hic_mapping/bowtie_results/bwt2/{sdir}/{prefix}_hg19.bwt2pairs.to.remap.num.gz"
shell:
dedent(
"""\
python2 {params.hornet}/find_intersecting_snps.py \
-p {input.bam} {params.snps}
"""
)
# Several remapping steps, similar to the first mapping steps, but in a different directory
rule wasp_merge:
input:
"hic_mapping/bowtie_results/bwt2/{sdir}/{prefix}_hg19.bwt2pairs.keep.bam",
"hic_mapping/wasp_results/{sdir}_{prefix}_filt_hg19.remap.kept.bam"
output:
"hic_mapping/wasp_results/{sdir}_{prefix}_filt_hg19.bwt2pairs.filt.bam"
params:
threads=config['job_cores']
shell:
dedent(
"""\
{module}
module load samtools
samtools merge --threads {params.threads} {output} {input}
"""
)
# All future steps use the name style wildcard, like below
rule move_results:
input:
"hic_mapping/wasp_results/{name}_filt_hg19.bwt2pairs.filt.bam"
output:
"hic_mapping/wasp_results/{name}_filt_hg19.bwt2pairs.bam"
shell:
dedent(
"""\
mv {input} {output}
"""
)
This pipeline is essentially doing some mapping steps in one directory structure that looks like hic_mapping/bowtie_results/bwt2/<subdir>/<file>
, (where subdir is three different directories) then filtering the results, and doing another almost identical mapping step in hic_remap/bowtie_results/bwt2/<subdir>/<file>
, before merging the results into an entirely new directory and collapsing the subdirectories into the file name: hic_mapping/wasp_results/<subdir>_<file>
.
The problem I have is that the wasp_merge
step breaks the find_intersecting_snps
step if I collapse the subdirectory name into the filename. If I just maintain the subdirectory structure, everything works fine. Doing this would break future steps of the pipeline though.
The error I get is:
MissingInputException in line 243 of /godot/quertermous/PooledHiChip/pipeline/Snakefile:
Missing input files for rule find_intersecting_snps:
hic_mapping/bowtie_results/bwt2/HCASMC5-8_HCASMC-8-CAGAGAGG-TACTCCTT_S8_L006/001_hg19.bwt2pairs.bam
The correct file is:
hic_mapping/bowtie_results/bwt2/HCASMC5-8/HCASMC-8-CAGAGAGG-TACTCCTT_S8_L006_001_hg19.bwt2pairs.bam
But it is looking for:
hic_mapping/bowtie_results/bwt2/HCASMC5-8_HCASMC-8-CAGAGAGG-TACTCCTT_S8_L006/001_hg19.bwt2pairs.bam
Which is not created anywhere, nor defined by any rule. I think it is somehow getting confused by the existence of the file created by the wasp_merge
step:
hic_mapping/wasp_results/HCASMC5-8_HCASMC-8-CAGAGAGG-TACTCCTT_S8_L006_001_filt_hg19.bwt2pairs.filt.bam
Or possibly a downstream file (after the target that creates this error):
hic_mapping/wasp_results/HCASMC5-8_HCASMC-8-CAGAGAGG-TACTCCTT_S8_L006_001_filt_hg19.bwt2pairs.bam
However, I have no idea why either of those files would confuse the find_intersecting_snps
rule, because the directory structures are totally different.
I feel like I must be missing something obvious, because this error is so absurd, but I cannot figure out what it is.
Upvotes: 2
Views: 2618
Reputation: 747
The problem is that both the directory name and the file name contain underscores, and in the final file name I separate the two components by underscores.
By either changing that separation character, or replacing the rule with a python function that get the names from elsewhere, I can solve the issue.
This works:
rule wasp_merge:
input:
"hic_mapping/bowtie_results/bwt2/{sdir}/{prefix}_hg19.bwt2pairs.keep.bam",
"hic_mapping/wasp_results/{sdir}{prefix}_filt_hg19.remap.kept.bam"
output:
"hic_mapping/wasp_results/{sdir}{prefix}_filt_hg19.bwt2pairs.filt.bam"
params:
threads=config['job_cores']
shell:
dedent(
"""\
{module}
module load samtools
samtools merge --threads {params.threads} {output} {input}
"""
)
Upvotes: 2