updated Snakefile

Snakefile

@@ -38,6 +38,7 @@ rule all:
     #input: expand("{data_dir}/multifast5/{run}", data_dir=DATA_DIR, run=RUNS)
     #input: expand(os.path.join(RESULTS_DIR, "basecalled/guppy/cpu/{run}/basecalling.done"), run=RUNS)
     #input: expand(os.path.join(RESULTS_DIR, "basecalled/{run}/basecalling.done"), run=RUNS)
+    input: expand(os.path.join(RESULTS_DIR, "basecalled_NO_MOD/{run}/NO_MOD_basecalling.done"), run=RUNS)
     #input: expand(os.path.join(RESULTS_DIR, "basecalled/{run}/demux.done"), run=RUNS),
     #input: expand(os.path.join(RESULTS_DIR, "basecalled/{run}/{barcode}/{barcode}.fastq.gz"), run=RUNS, barcode=BARCODES),
     #input: expand(os.path.join(RESULTS_DIR, "basecalled/merged/{barcode}.fastq.gz"), barcode=BARCODES)
@@ -55,10 +56,13 @@ rule all:
     #input: expand(os.path.join(RESULTS_DIR, "genomecov/lr/merged/{barcode}/{barcode}-x-igc.avg_cov.txt"), barcode=BARCODES)
     #input: expand(os.path.join(RESULTS_DIR, "mapping/lr/merged/{sample}-x-{barcode}.bam"), sample=SAMPLES, barcode=BARCODES)
     #input: expand(os.path.join(RESULTS_DIR, "assembly/{assembler}/lr/merged/{barcode}/assembly.fna"), assembler=ASSEMBLERS, barcode=BARCODES)
-    input: expand(os.path.join(RESULTS_DIR, "mapping/lr/merged/{barcode}/{barcode}_reads-x-{barcode}-{assembler}_contigs.bam"), barcode=BARCODES, assembler=ASSEMBLERS)
+    #input: expand(os.path.join(RESULTS_DIR, "mapping/lr/merged/{barcode}/{barcode}_reads-x-{barcode}-{assembler}_contigs.bam"), barcode=BARCODES, assembler=ASSEMBLERS)
     #input: expand(os.path.join(RESULTS_DIR, "genomecov/lr/merged/{barcode}/{barcode}_reads-x-{barcode}-{assembler}_contigs.avg_cov.txt"), barcode=BARCODES, assembler=ASSEMBLERS)
     #input: expand(os.path.join(RESULTS_DIR, "mapping/sr/{mapper}/{sample}_reads-x-{barcode}-{assembler}_contigs.bam"), sample=SAMPLES, barcode=BARCODES, assembler=ASSEMBLERS, mapper=MAPPERS)
-    #input: expand(os.path.join(RESULTS_DIR, "genomecov/sr/{sample}_reads-x-{barcode}-{assembler}_contigs.avg_cov.txt"), sample=SAMPLES, barcode=BARCODES, assembler=ASSEMBLERS)
+    #input: expand(os.path.join(RESULTS_DIR, "genomecov/sr/{mapper}/{sample}_reads-x-{barcode}-{assembler}_contigs.avg_cov.txt"), sample=SAMPLES, barcode=BARCODES, assembler=ASSEMBLERS, mapper=MAPPERS)
+    #input: "assemble_and_coverage.done"
+    #input: expand(os.path.join(RESULTS_DIR, "assembly/{assembler}/polished/medaka/round_{round}/lr/merged/{barcode}/assembly_polished.fna", assembler=ASSEMBLERS, barcode=BARCODES, round="01"))
+
 
 ######
 # WRAPPER RULES
@@ -122,9 +126,9 @@ rule ASSEMBLE_AND_COVERAGE:
         #expand(os.path.join(RESULTS_DIR, "assembly/operams/lr_{barcode}-sr_{sample}/contigs.fna"), sample=SAMPLES, barcode=BARCODES),
         #expand(os.path.join(RESULTS_DIR, "mapping/lr/merged/{barcode}/{barcode}_reads-x-{barcode}-{assembler}_contigs.bam"), barcode=BARCODES, assembler="flye"),
         # short reads on hybrid contigs
-        expand(os.path.join(RESULTS_DIR, "genomecov/sr/bwa_mem/{sample}_reads-x-lr_{barcode}_sr_{sample}-{assembler}_contigs.avg_cov.txt"), sample=SAMPLES, barcode=BARCODES, assembler="operams"),
+        #expand(os.path.join(RESULTS_DIR, "genomecov/sr/bwa_mem/{sample}_reads-x-lr_{barcode}_sr_{sample}-{assembler}_contigs.avg_cov.txt"), sample=SAMPLES, barcode=BARCODES, assembler="operams"),
         # long reads on hybrid contigs
-        expand(os.path.join(RESULTS_DIR, "genomecov/lr/bwa_mem/{barcode}_reads-x-lr_{barcode}_sr_{sample}-{assembler}_contigs.avg_cov.txt"), barcode=BARCODES, sample=SAMPLES, assembler="operams"),
+        #expand(os.path.join(RESULTS_DIR, "genomecov/lr/bwa_mem/{barcode}_reads-x-lr_{barcode}_sr_{sample}-{assembler}_contigs.avg_cov.txt"), barcode=BARCODES, sample=SAMPLES, assembler="operams"),
         #expand(os.path.join(RESULTS_DIR, "assembly/flye/lr/merged/{barcode}/assembly.fna"), barcode=BARCODES)
         # short reads on short read contigs
         expand(os.path.join(RESULTS_DIR, "genomecov/sr/bwa_mem/{sample}_reads-x-{sample}-{assembler}_contigs.avg_cov.txt"), sample=SAMPLES, assembler="megahit"),
@@ -151,7 +155,7 @@ rule ANNOTATE:
 rule POLISH_AND_COVERAGE:
     input:
         # short read-polished contigs from OPERA-MS
-        expand(os.path.join(RESULTS_DIR, "assembly/{assembler}/lr_{barcode}-sr_{sample}/contigs.fna"), sample=SAMPLES, barcode=BARCODES, assembler="operams"),
+        #expand(os.path.join(RESULTS_DIR, "assembly/{assembler}/lr_{barcode}-sr_{sample}/contigs.fna"), sample=SAMPLES, barcode=BARCODES, assembler="operams"),
         #long read polished contigs
         #expand(os.path.join(RESULTS_DIR, "assembly/{assembler}/polished/{polisher}/round_{round}/lr/merged/{barcode}/assembly_polished.fna"), assembler="flye", polisher="medaka", round="01", barcode=BARCODES),
         # map long reads to long read polished contigs
@@ -238,6 +242,19 @@ rule guppy_gpu_basecall:
         date) 2> >(tee {log}.stderr) > >(tee {log}.stdout)
         """
 
+rule guppy_gpu_basecall_NO_MOD:
+    input: rules.create_multifast5s.output
+    output: os.path.join(RESULTS_DIR, "basecalled_NO_MOD/{run}/NO_MOD_basecalling.done")
+    log: os.path.join(RESULTS_DIR, "basecalled_NO_MOD/{run}/NO_MOD_basecalling")
+    threads: config["guppy_gpu"]["cpu_threads"]
+    shell: 
+        """
+        (date &&\
+        {config[guppy_gpu][bin]} --input_path {input} --save_path $(dirname {output}) --config {config[guppy_gpu][hac_config]} --cpu_threads_per_caller {threads} -x "auto" --disable_pings --compress_fastq --records_per_fastq {config[guppy_gpu][records_per_fastq]} && \
+        touch {output} &&\
+        date) 2> >(tee {log}.stderr) > >(tee {log}.stdout)
+        """
+
 # Demultiplex.
 # Currently requires dummy target, but would be nice to have this be a "real" file target. However, it seems that files are filled incrementally, hence there is no check if the whole process was successful or not if checking only for file presence.
 rule guppy_gpu_demux: