Snakemake can't identify the rule - bioinformatics

Question

I'm writing a pipeline with Snakemake and the program can't identify the rule stringtie. I can't find what I'm doing wrong. I already runned the rule fastp and star, the problem is specific with the stringtie rule.
include:
'config.py'
rule all:
input:
expand(FASTP_DIR + "{sample}R{read_no}.fastq",sample=SAMPLES ,read_no=['1', '2']), #fastp
expand(STAR_DIR + STAR_DIR + "output/{sample}/{sample}Aligned.sortedByCoord.out.bam",sample=SAMPLES), #STAR
expand(STRINGTIE_DIR + "/{sample}/{sample}Aligned.sortedByCoord.out.gtf", sample=SAMPLES),
GTF_DIR + "path_samplesGTF.txt"
rule fastp:
input:
R1= DATA_DIR + "{sample}R1_001.fastq.gz",
R2= DATA_DIR + "{sample}R2_001.fastq.gz"
output:
R1out= FASTP_DIR + "{sample}R1.fastq",
R2out= FASTP_DIR + "{sample}R2.fastq"
params:
data_dir = DATA_DIR,
name_sample = "{sample}"
log: FASTP_LOG + "{sample}.html"
message: "Executando o programa FASTP"
run:
shell('fastp -i {input.R1} -I {input.R2} -o {output.R1out} -O {output.R2out} \
-h {log} -j {log}')
shell("find {params.data_dir} -type f -name '{params.name_sample}*' -delete ")
rule star:
input:
idx_star = IDX_DIR,
R1 = FASTP_DIR + "{sample}R1.fastq",
R2 = FASTP_DIR + "{sample}R2.fastq",
parameters = "parameters.txt",
params:
outdir = STAR_DIR + "output/{sample}/{sample}",
star_dir = STAR_DIR,
star_sample = '{sample}'
# threads: 18
output:
out = STAR_DIR + "output/{sample}/{sample}Aligned.sortedByCoord.out.bam"
#run_time = STAR + "log/star_run.time"
# log: STAR_LOG
# benchmark: BENCHMARK + "star/{sample_star}"
run:
shell("STAR --runThreadN 12 --genomeDir {input.idx_star} \
--readFilesIn {input.R1} {input.R2} --outFileNamePrefix {params.outdir}\
--parametersFiles {input.parameters} \
--quantMode TranscriptomeSAM GeneCounts \
--genomeChrBinNbits 12")
# shell("find {params.star_dir} -type f ! -name
'{params.star_sample}Aligned.sortedByCoord.out.bam' -delete")
rule stringtie:
input:
star_output = STAR_DIR + "output/{sample}/{sample}Aligned.sortedByCoord.out.bam"
output:
stringtie_output = STRINGTIE_DIR + "/{sample}/{sample}Aligned.sortedByCoord.out.gtf"
run:
shell("stringtie {input.star_output} -o {output.stringtie_output} \
-v -p 12 ")
rule grep_gtf:
input:
list_gtf = STRINGTIE_DIR
output:
paths = GTF_DIR + "path_samplesGTF.txt"
shell:
"find {input.list_gtf} | grep .gtf > {output.paths}"
This is the output I get with the option dry-run (flag -n)
Building DAG of jobs...
Job counts:
count jobs
1 all
1 grep_gtf
2
[Fri Apr 17 15:59:24 2020]
rule grep_gtf:
input: /homelocal/boralli/workdir/pipeline_v4/STRINGTIE/
output: /homelocal/boralli/workdir/pipeline_v4/GTF/path_samplesGTF.txt
jobid: 1
find /homelocal/boralli/workdir/pipeline_v4/STRINGTIE/ | grep .gtf >
/homelocal/boralli/workdir/pipeline_v4/GTF/path_samplesGTF.txt
[Fri Apr 17 15:59:24 2020]
localrule all:
input: /homelocal/boralli/workdir/pipeline_v4/GTF/path_samplesGTF.txt
jobid: 0
Job counts:
count jobs
1 all
1 grep_gtf
2
This was a dry-run (flag -n). The order of jobs does not reflect the order of execution.
I really don't know whats going on. The same pipeline worked before.

Answer 1

In addition to Troy's comment:
You specify as input of your rule grep_gtf a directory. Since that directory probably already exists, the rule stringtie does not need to be executed before running grep_gtf.
Using a directory as input isn't really a good idea. If you need the outputs of rule stringtie before executing rule grep_gtf, i suggest you specify the output files of rule stringtie as input of rule grep_gtf.
So your rule grep_gtf should be something like:
rule grep_gtf:
input:
expand(STRINGTIE_DIR + "/{sample}/{sample}Aligned.sortedByCoord.out.gtf", sample=SAMPLES)
output:
paths = GTF_DIR + "path_samplesGTF.txt"
shell:
"find {STRINGTIE_DIR} | grep .gtf > {output.paths}"
EDIT:
I think there's a bad copy/paste in rule all where there is twice STAR_DIR:
expand(STAR_DIR + STAR_DIR + "output/{sample}/{sample}Aligned.sortedByCoord.out.bam",sample=SAMPLES), #STAR
I also think there is a misunderstanding on the snakemake "workflow" concept. You do not need to specify the outputs of all rules in rule all. You only need to specify the last file of the workflow. Snakemake will decide which rules need to be run in order to achieve the creation of the final file. I don't really understand why your snakemake does not want to build the gtf files since you ask for them in rule all but I do see why rule grep_gtf does not need the output of rule stringtie to run.