Minimal Micro-C preprocessing and QC summarization on Slurm.
v0.1.0 – Minimal runnable release.

Runs alignment → pairing → cooler generation, plus a tiny QC report.
Opinionated defaults; designed for HPC queues (Slurm). Local mode planned.

• Single-sample run via one script: mdp.sh
• Produce BAM / pairs / .cool with sensible paths
• Tiny QC export (QC.tsv + QC.html) via get_qc.py
• Conda environment (env.yml) for quick setup
• Minimal CI (shellcheck + pyflakes)

Planned for v0.2.0: sample sheet orchestration, Nextflow preview, richer QC, container image.

flowchart LR
  A["FASTQ (R1/R2)"] --> B["Align (bwa or bowtie2)"]
  B --> C["Sort/Index (samtools)"]
  C --> D["Pairing (pairtools)"]
  D --> E["Contact matrix (.cool, cooler)"]
  E --> F["QC collect (get_qc.py)"]

  subgraph Outputs
    C
    D
    E
    F
  end

I/O layout (per sample)

input/{SAMPLE}_R1.fastq.gz
input/{SAMPLE}_R2.fastq.gz
results/{SAMPLE}/
  ├─ align/     # BAM + index
  ├─ pairs/     # .pairs / stats
  ├─ cool/      # .cool
  └─ logs/      # step-wise logs
qc/{SAMPLE}/
  ├─ QC.tsv
  └─ QC.html

• System: Linux on HPC (Slurm). Bash, Python ≥3.11.

• Reference: Genome FASTA (mm10/hg38, etc.) + index (.fai, bwa/bowtie2 index).

• Tools (installed via Conda/Mamba):

  • bwa or bowtie2, samtools, pairtools, cooler, bedtools, pigz

• Python libs: pandas, matplotlib (from requirements.txt)

Tip: On Biowulf or similar, point to institute-provided FASTA when possible.

# 1) Clone
git clone [email protected]:Fuki-Kudoh/microC-pipeline.git
cd microC-pipeline

# 2) Environment (conda/mamba)
mamba env create -f env.yml
mamba activate microc

# 3) Put input FASTQs
ls input/
# SAMPLE_R1.fastq.gz  SAMPLE_R2.fastq.gz

# 4) Submit (example: mm10)
sbatch --time=24:00:00 --cpus-per-task=16 --mem=32g \
  mdp.sh SAMPLE mm10 /path/to/mm10.fa

# 5) Collect QC (after run finishes)
python get_qc.py -i results/SAMPLE -o qc/SAMPLE

Minimal Slurm resources that work for smoke runs:

• --cpus-per-task 16 / --mem 32g / --time 24:00:00 Adjust for your read depth and queue policy.

For multiple samples, you can fan out jobs with a tiny launcher:

samples.csv
sample,fastq1,fastq2,genome,fa
S1,input/S1_R1.fastq.gz,input/S1_R2.fastq.gz,mm10,/path/to/mm10.fa

bash run_samples.sh    # submits one Slurm job per row

Usage:
  mdp.sh <SAMPLE_ID> <GENOME> <GENOME_FASTA>

Env:
  INPUT_DIR   default: input
  OUT_DIR     default: results
  LOG_DIR     default: logs

Behavior

• Validates dependencies (bwa/bowtie2, samtools, pairtools, cooler, bedtools, pigz)
• Checks FASTA + .fai, and input FASTQs
• Writes step logs under logs/ and results/<SAMPLE>/logs/
• Idempotent-ish: will not overwrite obvious final outputs unless --force (stub)

--dry-run (stub) prints planned commands without running them.

Usage:
  python get_qc.py -i results/<SAMPLE> -o qc/<SAMPLE>

Emits:

• QC.tsv – compact metrics (reads_total, mapped_rate, dup_rate, etc.)
• QC.html – quick text report + small plots (reads_total.png, …)

mamba env create -f env.yml
mamba activate microc
python -m pip install -r requirements.txt

mamba create -n microc python=3.11 bwa bowtie2 samtools pairtools cooler bedtools pigz -c conda-forge -c bioconda
mamba activate microc
pip install -r requirements.txt

Pin additional versions in env.yml if you need stricter reproducibility.

• Permission denied (publickey) when cloning Configure SSH (e.g., 1Password SSH Agent) and add your public key on GitHub.

• FASTA not found or missing .fai Ensure GENOME_FASTA points to the actual .fa/.fasta; run samtools faidx.

• Memory/timeouts Start with 16 CPU / 32 GB / 24 h and scale based on read depth. Very deep libraries may need --cpus-per-task 32 and --mem 64g+.

• Cooler errors Verify chromosome naming consistency between FASTA and outputs (chr1 vs 1).

If a job fails, check results/<SAMPLE>/logs/ and Slurm output files. Please open an issue with the failing step, command, and log snippet.

• License: MIT (see LICENSE)
• How to cite: see CITATION.cff (v0.1.0)

If you use this pipeline in academic work, please reference the release tag and commit hash.

• v0.1.x (hardening): --dry-run/--force, better logs, sample-sheet UX, parse real QC logs, FAQs
• v0.2.0 (preview): Nextflow wrapper (slurm/local), container image, benchmark notes
• v0.3+: richer QC HTML, automated reports, downstream modules

To fully reproduce results, record:

• Commit: git rev-parse --short HEAD
• Release tag: v0.1.0
• conda list --explicit export
• FASTA source and index build command

PRs and issues are welcome. Please keep changes small and focused. For feature ideas: open a discussion with example inputs/outputs and expected runtime.

::contentReference[oaicite:0]{index=0}