Add multiqc

[rroutsong]

Add multiqc into ngsqc pipeline.

Addresses #12 #7

[rroutsong]

@skchronicles , example report at /data/RTB_GRS/dev/Dmux/test_ngsqc2/GRS_0212_Bhasym/230907_NS500353_0215_AHLTNVBGXM/multiqc/Run-230907_NS500353_0215_AHLTNVBGXM-Project-GRS_0212_Bhasym_multiqc_report.html

[skchronicles]

We need to add an option to point to a sif cache and dynamically resolve one of the following: a local SIF on the file-system or a URI to pull an image from Dockerhub.

[rroutsong]

I have a solution to this issue in the next coming PR. I have serialized the server-centric SIF directories and dynamically adding the specific server configuration at initialization time.

Ends up like:

containerized: server_config["sif"] + "dmux_ngsqc_0.0.1.sif"

SIF cache is always specified at execution time through environmental variables and subprocess.

[skchronicles]

We need a method to dynamically resolve the reference files.

[rroutsong]

Also addressed this in the next PR. I just kind of saved all the server resolution methods until I moved onto bigsky.

[skchronicles]

At some point, we may want to point to a MutliQC config file to clean up the general statistics table, create two sections for fastqc, and create a preferred module order in the final report.

[rroutsong]

This is outlined in #15

[skchronicles]

We need to change how adapter sequences are being removed. Currently, there is a bug where the barcode sequences from Illumina's sample sheet (i7/i5) sequences are being passed to fastqc and fastp. These barcode sequences should be removed after bcl2fastq step and do not represent traditional library-prep-kit-specific adapter sequences that need to removed. With that being said, let's make use of fastp's auto-detect-adapter-sequences feature to remove them. We can also make use of fastqc's internal contaminates/adapters list to identify sequencing adapters.

[rroutsong]

Here's fastp rule in new branch master_job_and_bigsky:

    shell:
        """
        fastp \
        --detect_adapter_for_pe \
        --in1 {input.in_read1} --in2 {input.in_read2} \
        --out1 {output.out_read1} \
        --out2 {output.out_read2} \
        --html {output.html} \
        --json {output.json} \
        """

Fastqc:

    shell:
        """
        mkdir -p {params.output_dir}
        fastqc -o {params.output_dir} -t {threads} {input.samples}
        """

FastQC before trim depends on demuxed reads, after trimmed depends on trimmed reads file.

[skchronicles]

Will address some of these comments/issues in the next PR.