Skip to content
Snippets Groups Projects
Commit 6eb1455d authored by Susheel Busi's avatar Susheel Busi
Browse files

Merge branch 'work' into 'master' 

'rule all' updated

See merge request susheel.busi/ont_pilot_gitlab!7
parents 2c4d1e8e e6f39dee
No related branches found
No related tags found
1 merge request!7'rule all' updated
Hide whitespace changes
Inline Side-by-side
Showing
with 46 additions and 23 deletions
Snakefile
+ 46
− 23
View file @ 6eb1455d
......@@ -55,13 +55,36 @@ rule all:
#input: expand(os.path.join(RESULTS_DIR, "genomecov/sr/{mapper}/{sample}-x-igc.avg_cov.txt"), sample="NEB2_MG_S17", mapper=MAPPERS)
#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, "assembly/flye/lr/merged/{barcode}/assembly.fna"), 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, "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/{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"))
#### ADDITIONAL RULES TO BE RUN #####
input: "basecall_merge_qc.done"
input: "coverage_of_references.done"
input: "annotate.done"
input: expand(os.path.join(RESULTS_DIR, "qc/lr/{run}/{barcode}/{barcode}NanoStats.txt"), run=RUNS, barcode=BARCODES)
#input: expand(os.path.join(RESULTS_DIR, "qc/sr/fastqc/{sr_sample}/{sr_sample}_{suffix}.fastqc.zip"), sr_sample="NEB2_MG_S17",suffix=["R1_001.fastp", "R1_002.fastp"]))
#input: expand(os.path.join(RESULTS_DIR, "assembly/{assembler}/polished/racon/round_{round}/lr/merged/{barcode}/assembly_polished.fna"), barcode=BARCODES, assembler=ASSEMBLERS, round="01")
#input: expand(os.path.join(RESULTS_DIR, "assembly/metaspades_hybrid/lr_{barcode}-sr_{sr_sample}/contigs.fna"), barcode=BARCODES, sr_sample="NEB2_MG_S17")
#input: expand(os.path.join(RESULTS_DIR, "mapping/sr/bwa_mem/{sr_sample}_reads-x-lr_{barcode}_sr_{sr_sample}-{assembler}_contigs.bam"), barcode=BARCODES, assembler=ASSEMBLERS, sr_sample="NEB2_MG_S17")
#input: expand(os.path.join(RESULTS_DIR, "mapping/lr/bwa_mem/{barcode}_reads-x-lr_{barcode}_sr_{sr_sample}-{assembler}_contigs.bam"), barcode=BARCODES, assembler=ASSEMBLERS, sr_sample="NEB2_MG_S17")
#input: expand(os.path.join(RESULTS_DIR, "mapping/polished/{polisher}/round_{round}/lr/merged/{barcode}/{barcode}_reads-x-{barcode}-{assembler}_polished_contigs.bam"), barcode=BARCODES, assembler=ASSEMBLERS, round="01", polisher="medaka")
#input: expand(os.path.join(RESULTS_DIR, "assembly/megahit/{sr_sample}/final.contigs.fna"), sr_sample="NEB2_MG_S17")
#input: expand(os.path.join(RESULTS_DIR, "mapping/lr/bwa_mem/{barcode}_reads-x-{sr_sample}-{assembler}_contigs.bam"), barcode=BARCODES, assembler=ASSEMBLERS, sr_sample="NEB2_MG_S17")
#input: expand(os.path.join(RESULTS_DIR, "annotation/nonpareil/{sr_sample}_R{orientation}_{suffix}.fastp.npo"), sr_sample="NEB2_MG_S17", orientation=["1", "2"], suffix="001")
#input: expand(os.path.join(RESULTS_DIR, "basecalled/merged/{barcode}.fastq.gz.index.readdb"), barcode=BARCODES)
#input: expand(directory(os.path.join(RESULTS_DIR, "assembly/flye/polished/nanopolish/round_{round}/lr/merged/{barcode}/windows"), round="01", barcode=BARCODES))
#input: expand(os.path.join(RESULTS_DIR, "assembly/{assembler}/polished/nanopolish/round_{round}/lr/merged/{barcode}/assembly_polished.fna"), assembler=ASSEMBLERS, round="01", barcode=BARCODES)
#input: expand(os.path.join(RESULTS_DIR, "annotation/methylation/{assembler}/nanopolish/round_{round}/lr/merged/{barcode}/methylation_frequencies.tsv"), assembler=ASSEMBLERS, round="01", barcode=BARCODES)
#input: expand(os.path.join(RESULTS_DIR, "annotation/diamond/lr/merged/{barcode}.diamond.daa"), barcode=BARCODES)
#input: expand(os.path.join(RESULTS_DIR, "annotation/proteins/{prefix}.faa"))
#input: expand(os.path.join(RESULTS_DIR, "annotation/diamond/{prefix}.daa"))
#input: expand(os.path.join(RESULTS_DIR, "annotation/diamond/{prefix}.tsv"))
######
......@@ -652,27 +675,27 @@ rule polish_long_read_contigs_w_racon_initial_round:
#polished_contigs=os.path.join(RESULTS_DIR, "assembly/operams/lr_{barcode}-sr_{sr_sample}/contigs.polished.fna")
#ln -s contigs.polished.fasta {output.polished_contigs} &&\
rule assemble_long_and_short_reads_w_operams:
input:
lr=rules.get_single_fastq_per_barcode.output,
sr_r1=rules.run_fastp_on_short_reads.output.r1,
sr_r2=rules.run_fastp_on_short_reads.output.r2,
sr_contigs=os.path.join(RESULTS_DIR, "assembly/megahit/{sr_sample}/final.contigs.fna")
output:
contigs=os.path.join(RESULTS_DIR, "assembly/operams/lr_{barcode}-sr_{sr_sample}/contigs.fna")
params:
tmp_lr=os.path.join(RESULTS_DIR, "assembly/operams/lr_{barcode}-sr_{sr_sample}/tmp_lr_{barcode}.fna")
threads: config["operams"]["threads"]
log: os.path.join(RESULTS_DIR, "assembly/operams/lr_{barcode}-sr_{sr_sample}/opera_ms.log")
shell:
"""
(date &&\
zcat {input.lr} > {params.tmp_lr} &&\
{config[operams][bin]} --long-read-mapper minimap2 --num-processors {threads} --short-read1 {input.sr_r1} --short-read2 {input.sr_r2} --long-read {params.tmp_lr} --contig-file {input.sr_contigs} --out-dir $(dirname {output.contigs}) &&\
ln -s contigs.fasta {output.contigs} &&\
rm {params.tmp_lr} &&\
date) 2> >(tee {log}) > >(tee {log})
"""
#rule assemble_long_and_short_reads_w_operams:
# input:
# lr=rules.get_single_fastq_per_barcode.output,
# sr_r1=rules.run_fastp_on_short_reads.output.r1,
# sr_r2=rules.run_fastp_on_short_reads.output.r2,
# sr_contigs=os.path.join(RESULTS_DIR, "assembly/megahit/{sr_sample}/final.contigs.fna")
# output:
# contigs=os.path.join(RESULTS_DIR, "assembly/operams/lr_{barcode}-sr_{sr_sample}/contigs.fna")
# params:
# tmp_lr=os.path.join(RESULTS_DIR, "assembly/operams/lr_{barcode}-sr_{sr_sample}/tmp_lr_{barcode}.fna")
# threads: config["operams"]["threads"]
# log: os.path.join(RESULTS_DIR, "assembly/operams/lr_{barcode}-sr_{sr_sample}/opera_ms.log")
# shell:
# """
# (date &&\
# zcat {input.lr} > {params.tmp_lr} &&\
# {config[operams][bin]} --long-read-mapper minimap2 --num-processors {threads} --short-read1 {input.sr_r1} --short-read2 {input.sr_r2} --long-read {params.tmp_lr} --contig-file {input.sr_contigs} --out-dir $(dirname {output.contigs}) &&\
# ln -s contigs.fasta {output.contigs} &&\
# rm {params.tmp_lr} &&\
# date) 2> >(tee {log}) > >(tee {log})
# """
#zcat {input.lr} > {params.tmp_lr} &&\
rule assemble_long_and_short_reads_w_metaspades:
......@@ -782,7 +805,7 @@ rule map_long_reads_to_long_read_polished_contigs:
prefix=os.path.join(RESULTS_DIR, "mapping/polished/{polisher}/round_{round}/lr/merged/{barcode}/{barcode}_reads-x_{barcode}_polished_contigs")
conda: "envs/minimap2.yaml"
threads: config["minimap2"]["threads"]
log: os.path.join(RESULTS_DIR, "mapping/polished/{polisher}/round_{round}/lr/merged/{barcode}/{barcode}_reads-x-{barcode}-{assembler}_polished_con tigs.minimap2_samtools")
log: os.path.join(RESULTS_DIR, "mapping/polished/{polisher}/round_{round}/lr/merged/{barcode}/{barcode}_reads-x-{barcode}-{assembler}_polished_contigs.minimap2_samtools")
shell:
"""
(date &&\
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment