Solution for QC exercises

src/05_qc/Snakefile

+SAMPLE = "TC1-ST2-D0.12"
+
+rule all:
+  input: f"output/{SAMPLE}_peaks.narrowPeak", f"output/{SAMPLE}_control_lambda.bigwig", f"output/{SAMPLE}_treat_pileup.bigwig", f"output/H3K4-{SAMPLE}_fastqc.html", f"output/INPUT-{SAMPLE}_fastqc.html", f"output/alignment_metrics_H3K4-{SAMPLE}.txt", f"output/alignment_metrics_INPUT-{SAMPLE}.txt"
+
+rule mapping:
+  input: "chip-seq/{sample}.fastq.gz"
+  output: "bowtie2/{sample}.bam"
+  params:
+    idx = "reference/Mus_musculus.GRCm38.dna_sm.chromosome.12"
+  log: "logs/bowtie2_{sample}.log"
+  benchmark: "benchmarks/mapping/{sample}.tsv"
+  conda: "envs/bowtie2.yaml"
+  shell:
+    """
+    bowtie2 -x {params.idx} -U {input} 2> {log} | \
+    samtools sort - > {output}
+    samtools index {output}
+    """
+
+rule peak_calling:
+  input:
+    control = "bowtie2/INPUT-{sample}.bam",
+    chip = "bowtie2/H3K4-{sample}.bam"
+  output:
+    peaks = "output/{sample}_peaks.narrowPeak",
+    control_bdg = "macs2/{sample}_control_lambda.bdg",
+    chip_bdg = "macs2/{sample}_treat_pileup.bdg"
+  conda: "envs/macs2.yaml"
+  shell:
+    """
+    macs2 callpeak -t {input.chip} -c {input.control} -f BAM -g mm -n {wildcards.sample} -B -q 0.01 --outdir macs2
+
+    cp macs2/{wildcards.sample}_peaks.narrowPeak {output.peaks}
+    """
+
+rule bigwig:
+  input: "macs2/{sample}.bdg"
+  output: "output/{sample}.bigwig"
+  params:
+    idx = "reference/Mus_musculus.GRCm38.dna_sm.chromosome.12.fa.fai"
+  conda: "envs/ucsc.yaml"
+  shell:
+    """
+    bedGraphToBigWig {input} {params.idx} {output}
+    """
+
+rule fastqc:
+  input: "chip-seq/{sample}.fastq.gz"
+  output:
+    html = "output/{sample}_fastqc.html",
+    zip = "chip-seq/{sample}_fastqc.zip"
+  conda: "envs/fastqc.yaml"
+  shell:
+    """
+    fastqc {input}
+    cp chip-seq/{wildcards.sample}_fastqc.html {output.html}
+    """
+
+rule alignment_metrics:
+  input: "bowtie2/{sample}.bam"
+  output: "output/alignment_metrics_{sample}.txt"
+  params:
+    idx = "reference/Mus_musculus.GRCm38.dna_sm.chromosome.12.fa"
+  conda: "envs/picard.yaml"
+  shell:
+    """
+    picard CollectAlignmentSummaryMetrics R={params.idx} I={input} O={output}
+    """

src/05_qc/envs/fastqc.yaml

+name: fastqc
+channels:
+  - bioconda
+dependencies:
+  - fastqc

src/05_qc/envs/picard.yaml

+name: fastqc
+channels:
+  - bioconda
+dependencies:
+  - picard