I am using snakemake in a workflow for NGS analyses.
In one rule, I make use of the unique (temporary) output from another rule.The output of this one rule is also unique and contributes to the creation of the final output. A simple wildcard {sample} is used over these rules. I do not see any cyclic dependency, but snakemake tells me there is:
CyclicGraphException in line xxx of Snakefile: Cyclic dependency on rule
I understand that there is an option to investigate this problem: --debug-dag.
How do I interpret the output? What is candidate versus selected?
This my (pseudo-) code of the rule:
rule split_fasta:
input:
dataFile="data/path1/{sample}.tab",
scaffolds="data/path2/{sample}.fasta",
database="path/to/db",
output:
onefasta="data/path2/{sample}_one.fasta",
twofasta="data/path2/{sample}_two.fasta",
threefasta="data/path2/{sample}_three.fasta",
conda:
"envs/env.yaml"
log:
"logs/split_fasta_{sample}.log"
benchmark:
"logs/benchmark/split_fasta_{sample}.txt"
threads: 4
shell:
"""
python bin/split_fasta.py {input.dataFile} {input.scaffolds} {input.database} {output.onefasta} {output.twofasta} {output.threefasta}
"""
There is no other connection between input and output than in this rule.
The problem is solved now, further downstream and upstream some subtle dependencies were present.
But, for future reference I would like to know how to interpret the output od the --debug-dag option.
--debug-dag Print candidate and selected jobs (including their wildcards) while inferring DAG. This can help to debug unexpected DAG topology or errors.
It does not seem to have further documentation than this, but I believe the candidate jobs are the jobs that can be made matching to the required string through wildcards. The selected job is the one that is chosen from the candidates (either through wildcard constraints, ruleorder, or the first candidate with the option --allow-ambiguity).
As an example I have a rule that does adapter trimming, and I have a rule for both paired end and single end:
rule trim_SE:
input:
"{sample}.fastq.gz"
output:
"{sample}_trimmed.fastq.gz"
shell:
...
rule trim_PE:
input:
"{sample}_R1.fastq.gz",
"{sample}_R2.fastq.gz"
output:
"{sample}_R1_trimmed.fastq.gz"
"{sample}_R2_trimmed.fastq.gz"
shell:
...
If I now tell snakemake to generate the output exp_R1_trimmed.fastq.gz it complains that it can use either rule.
AmbiguousRuleException:
Rules trim_PE and trim_SE are ambiguous for the file exp_R1_trimmed.fastq.gz.
Consider starting rule output with a unique prefix, constrain your wildcards, or use the ruleorder directive.
Wildcards:
trim_PE: sample=exp
trim_SE: sample=exp_R1
we can solve this problem by for instance placing a ruleorder:
ruleorder: trim_PE > trim_SE
And the file gets generated as we want. If we now use the --debug-dag option we get two candidate rules, and one selected rule (based on our ruleorder).
candidate job trim_PE
wildcards: sample=exp
candidate job trim_SE
wildcards: sample=exp_R1
selected job sra2fastq_PE
wildcards: sample=GSM2837484
If the rule trim_PE and trim_SE depended on other rules downstream, we can use the --debug-dag option to detect in which rule the wildcard expansion goes wrong, instead of only getting an error in the rule where it goes wrong.
Related
In this Snakemake script the rule all defines a target, and there are three other rules that claim this target as an output:
rule all:
input:
"target.txt"
rule from_non_existing_file:
input:
"non_existing_file.txt"
output:
"target.txt"
rule broad_input:
output:
"target.txt"
rule narrow_input:
input:
"optional_input.txt"
output:
"target.txt"
ruleorder: narrow_input > broad_input
The file non_existing_file.txt doesn't exist, so the rule from_non_existing_file should not be regarded by Snakemake. The rule broad_input has no input files, so it always can produce the output, and the rule narrow_input can produce the output whenever the file optional_input.txt exists. To resolve the ambiguity between broad and narrow inputs, the ruleorder is defined.
Whenever the file optional_input.txt exists, the script prefers the rule narrow_input:
Job counts:
count jobs
1 all
1 narrow_input
2
This script works most of the times, but sometimes it fails:
AmbiguousRuleException:
Rules narrow_input and broad_input are ambiguous for the file target.txt.
Consider starting rule output with a unique prefix, constrain your wildcards, or use the ruleorder directive.
Wildcards:
narrow_input:
broad_input:
Expected input files:
narrow_input: optional_input.txt
broad_input: Expected output files:
narrow_input: target.txt
broad_input: target.txt
Here Snakemake ignores the fact that the ruleorder directive is defined, and advises to define it again.
To confirm this behavior I've designed the test script below:
import os
def test_snakemake():
for i in range(100):
rcode = os.system("snakemake --cores=1 --printshellcmds --forceall --dry-run")
assert(rcode == 0)
This test fails within first 20 iterations with high confidence.
I've conducted some experiments and got surprising results:
The test pass if optional_input.txt doesn't exist
The test pass if any of the three rules is removed
This problem is confirmed on two different Windows machines with Snakemake versions 5.7.4 and 6.5.3.
My question is whether that is a Snakemake bug. Is there another explanation of this behavior?
In the first step of my process, I am extracting some hourly data from a database. Because of things data is sometimes missing for some hours resulting in files. As long as the amount of missing files is not too large I still want to run some of the rules that depend on that data. When running those rules I will check how much data is missing and then decide if I want to generate an error or not.
An example below. The Snakefile:
rule parse_data:
input:
"data/1.csv", "data/2.csv", "data/3.csv", "data/4.csv"
output:
"result.csv"
shell:
"touch {output}"
rule get_data:
output:
"data/{id}.csv"
shell:
"Rscript get_data.R {output}"
And my get_data.R script:
output <- commandArgs(trailingOnly = TRUE)[1]
if (output == "data/1.csv")
stop("Some error")
writeLines("foo", output)
How do I force running of the rule parse_data even when some of it's inputs are missing? I do not want to force running any other rules when input is missing.
One possible solution would be to generate, for example, an empty file in get_data.R when the query failed. However, in practice I am also using --restart-times 5 when running snakemake as the query can also fail because of database timeouts. When creating an empty file this mechanism of retrying the queries would no longer work.
You need data-dependent conditional execution.
Use a checkpoint on get_data. Then you replace parse_data's input with a function, that aggregates whatever files do exist.
(note that I am a Snakemake newbie and am just learning this myself, I hope this is helpful)
I am trying to create a pipeline where a small chain of rules are ran on a dynamic number of files output by an earlier rule using output. However, I am getting the following error: "wildcards in input files cannot be determined from output files:".
This suggests to me that what I am trying to do is not currently supported. Here is a pseudo example of what I am trying to do:
rule a:
input: "my static file.txt"
output: dynamic('my/path/{id}.txt')
rule b:
input: dynamic('my/path/{id}.txt')
output: dynamic('my/path/{id}.reprocessed.txt')
rule c:
input: dynamic('my/path/{id}.reprocessed.txt')
output: 'gather.txt'
Running snakemake with
rule all:
input: dynamic('my/path/{id}.txt')
Works without any issues, but when I run snakemake with:
rule all:
input: dynamic('my/path/{id}.reprocessed.txt')
I get the error: "wildcards in input files cannot be determined from output files:"
Is this feature supported? Has anyone successfully made such a chain? Any considerations I need to take into account?
Thanks!
This was resolved by removing the dynamic statement from rule b.
I am currently working on this project where iam struggling with this issue.
My current directory structure is
/shared/dir1/file1.bam
/shared/dir2/file2.bam
/shared/dir3/file3.bam
I want to convert various .bam files to fastq in the results directory
results/file1_1.fastq.gz
results/file1_2.fastq.gz
results/file2_1.fastq.gz
results/file2_2.fastq.gz
results/file3_1.fastq.gz
results/file3_2.fastq.gz
I have the following code:
END=["1","2"]
(dirs, files) = glob_wildcards("/shared/{dir}/{file}.bam")
rule all:
input: expand( "/results/{sample}_{end}.fastq.gz",sample=files, end=END)
rule bam_to_fq:
input: {dir}/{sample}.bam"
output: left="/results/{sample}_1.fastq", right="/results/{sample}_2.fastq"
shell: "/shared/packages/bam2fastq/bam2fastq --force -o /results/{sample}.fastq {input}"
This outputs the following error:
Wildcards in input files cannot be determined from output files:
'dir'
Any help would be appreciated
You're just missing an assignment for "dir" in your input directive of the rule bam_to_fq. In your code, you are trying to get Snakemake to determine "{dir}" from the output of the same rule, because you have it setup as a wildcard. Since it didn't exist, as a variable in your output directive, you received an error.
input:
"{dir}/{sample}.bam"
output:
left="/results/{sample}_1.fastq",
right="/results/{sample}_2.fastq",
Rule of thumb: input and output wildcards must match
rule all:
input:
expand("/results/{sample}_{end}.fastq.gz", sample=files, end=END)
rule bam_to_fq:
input:
expand("{dir}/{{sample}}.bam", dir=dirs)
output:
left="/results/{sample}_1.fastq",
right="/results/{sample}_2.fastq"
shell:
"/shared/packages/bam2fastq/bam2fastq --force -o /results/{sample}.fastq {input}
NOTES
the sample variable in the input directive now requires double {}, because that is how one identifies wildcards in an expand.
dir is no longer a wildcard, it is explicitly set to point to the list of directories determined by the glob_wildcard call and assigned to the variable "dirs" which I am assuming you make earlier in your script, since the assignment of one of the variables is successful already, in your rule all input "sample=files".
I like and recommend easily differentiable variable names. I'm not a huge fan of the usage of variable names "dir", and "dirs". This makes you prone to pedantic spelling errors. Consider changing it to "dirLIST" and "dir"... or anything really. I just fear one day someone will miss an 's' somewhere and it's going to be frustrating to debug. I'm personally guilty, an thus a slight hypocrite, as I do use "sample=samples" in my core Snakefile. It has caused me minor stress, thus why I make this recommendation. Also makes it easier for others to read your code as well.
EDIT 1; Adding to response as I had initially missed the requirement for key-value matching of the dir and sample
I recommend keeping separate the path and the sample name in different variables. Two approaches I can think of:
Keep using glob_wildcards to make a blanket search for all possible variables, and then use a python function to validate which path+file combinations are legit.
Drop the usage of glob_wildcards. Propagate the directory name as a wildcard variable, {dir}, throughout your rules. Just set it as a sub-directory of "results". Use pandas to pass known, key-value pairs listed in a file to the rule all. Initially I suggest generating the key-value pairs file manually, but eventually, it's generation could just be a rule upstream of others.
Generalizing bam_to_fq a little bit... utilizing an external config, something like....
from pandas import read_table
rule all:
input:
expand("/results/{{sample[1][dir]}}/{sample[1][file]}_{end}.fastq.gz", sample=read_table(config["sampleFILE"], " ").iterrows(), end=['1','2'])
rule bam_to_fq:
input:
"{dir}/{sample}.bam"
output:
left="/results/{dir}/{sample}_1.fastq",
right="/results/{dir}/{sample}_2.fastq"
shell:
"/shared/packages/bam2fastq/bam2fastq --force -o /results/{sample}.fastq {input}
sampleFILE
dir file
dir1 file1
dir2 file2
dir3 file3
New to snakemake and I've been trying to transform my shell script based pipeline into snakemake based today and run into a lot of syntax issues.. I think most of the trouble I have is around getting all the files in a particular directories and infer output names from input names since that's how I use shell script (for loop), in particular, I tried to use expand function in the output section and it always gave me an error.
After checking some example Snakefile, I realized people never use expand in the output section. So my first question is: is output the only section where expand can't be used and if so, why? What if I want to pass a prefix defined in config.yaml file as part of the output file and that prefix can not be inferred from input file names, how can I achieve that, just like what I did below for the log section where {runid} is my prefix?
Second question about syntax: I tried to pass a user defined id in the configuration file (config.yaml) into the log section and it seems to me that here I have to use expand in the following form, is there a better way of passing strings defined in config.yaml file?
log:
expand("fastq/fastqc/{runid}_fastqc_log.txt",runid=config["run"])
where in the config.yaml
run:
"run123"
Third question: I initially tried the following 2 methods but they gave me errors so does it mean that inside log (probably input and output) section, Python syntax is not followed?
log:
"fastq/fastqc/"+config["run"]+"_fastqc_log.txt"
log:
"fastq/fastqc/{config["run"]}_fastqc_log.txt"
Here is an example of small workflow:
# Sample IDs
SAMPLES = ["sample1", "sample2"]
CONTROL = ["sample1"]
TREATMENT = ["sample2"]
rule all:
input: expand("{treatment}_vs_{control}.bed", treatment=TREATMENT, control=CONTROL)
rule peak_calling:
input: control="{control}.sam", treatment="{treatment}.sam"
output: "{treatment}_vs_{control}.bed"
shell: "touch {output}"
rule mapping:
input: "{samples}.fastq"
output: "{samples}.sam"
shell: "cp {input} {output}"
I used the expand function only in my final target. From there, snakemake can deduce the different values of the wildcards used in the rules "mapping" and "peak_calling".
As for the last part, the right way to put it would be the first one:
log:
"fastq/fastqc/" + config["run"] + "_fastqc_log.txt"
But again, snakemake can deduce it from your target (the rule all, in my example).
rule mapping:
input: "{samples}.fastq"
output: "{samples}.sam"
log: "{samples}.log"
shell: "cp {input} {output}"
Hope this helps!
You can use f-strings:
If this is you folder_with_configs/some_config.yaml:
var: value
Then simply
configfile:
"folder_with_configs/some_config.yaml"
rule one_to_rule_all:
output:
f"results/{config['var']}.file"
shell:
"touch {output[0]}"
Do remember about python rules related to nesting different types of apostrophes.
config in the smake rule is a simple python dictionary.
If you need to use additional variables in a path, e.g. some_param, use more curly brackets.
rule one_to_rule_all:
output:
f"results/{config['var']}.{{some_param}}"
shell:
"touch {output[0]}"
enjoy