CNVKIT_SEGMENT(1)

✨ CNVKIT_SEGMENT ✨

okay so bestie, this is cnvkit_segment from the User Commands section of the manual (section 1, iykyk) — and it's literally the part of the pipeline where we infer copy number segments from the coverage table you already built. it's giving detective work for your genome 💅

---

🎀 DESCRIPTION 🎀

the vibe check, in its natural habitat:

usage: cnvkit.py segment [-h] [-o FILENAME] [-d DATAFRAME]
       [-m {cbs,flasso,haar,none,hmm,hmm-tumor,hmm-germline}]
              [-t THRESHOLD] [--drop-low-coverage] [--drop-outliers FACTOR]
              [--rscript-path PATH] [-p [PROCESSES]] [--smooth-cbs] [-v
              FILENAME] [-i SAMPLE_ID] [-n NORMAL_ID] [--min-variant-depth
              MIN_VARIANT_DEPTH] [-z [ALT_FREQ]] filename

---

🌸 positional arguments

• filename — the bin-level log2 ratios, aka the .cnr file, exactly as produced by fix. that's your main character input, no substitutions 💖

---

🦋 options

• -h, --help — show this help message and dip. classic.
• -o FILENAME, --output FILENAME — output table file name: a CNR-like table of segments, .cns. this is where your beautiful segments go 🎀
• -d DATAFRAME, --dataframe DATAFRAME — file name to save the raw R dataframe emitted by CBS or Fused Lasso. highkey useful for debugging 👀
• -m {cbs,flasso,haar,none,hmm,hmm-tumor,hmm-germline}, --method {cbs,flasso,haar,none,hmm,hmm-tumor,hmm-germline} — segmentation method (see docs), or none for chromosome arm-level averages as segments. default: cbs — pick your fighter, bestie ⭐
• -t THRESHOLD, --threshold THRESHOLD — significance threshold (p-value or FDR, depending on method) to accept breakpoints during segmentation. for HMM methods, this is the smoothing window size. so like, how confident do we need to be before we call it 👑
• --drop-low-coverage — drop very-low-coverage bins before segmentation to avoid false-positive deletions in poor-quality tumor samples. protecting you from lies, fr fr 🛡️
• --drop-outliers FACTOR — drop outlier bins more than this many multiples of the 95th quantile away from the average within a rolling window. set to 0 for no outlier filtering. default: 10 — bye bye weirdo bins 🙈
• --rscript-path PATH — path to the Rscript executable to use for running R code. use this option to specify a non-default R installation. default: Rscript
• -p [PROCESSES], --processes [PROCESSES] — number of subprocesses to segment in parallel. give 0 or a negative value to use the maximum number of available CPUs. default: use 1 process — go parallel, go fast, slay ✨
• --smooth-cbs — perform an additional smoothing before CBS segmentation, which in some cases may increase the sensitivity. used only for CBS method. extra smoothing, extra sensitive, no cap 🌈

💜 To additionally segment SNP b-allele frequencies:

• -v FILENAME, --vcf FILENAME — VCF file name containing variants for segmentation by allele frequencies.
• -i SAMPLE_ID, --sample-id SAMPLE_ID — specify the name of the sample in the VCF (-v/--vcf) to use for b-allele frequency extraction and as the default plot title.
• -n NORMAL_ID, --normal-id NORMAL_ID — corresponding normal sample ID in the input VCF (-v/--vcf). this sample is used to select only germline SNVs to plot b-allele frequencies.
• --min-variant-depth MIN_VARIANT_DEPTH — minimum read depth for a SNV to be displayed in the b-allele frequency plot. default: 20
• -z [ALT_FREQ], --zygosity-freq [ALT_FREQ] — ignore VCF's genotypes (the GT field) and instead infer zygosity from allele frequencies. default if used without a number: 0.25 💀

---

ngl bestie, read the descriptions before you yeet flags at your data — --drop-outliers 0 and --drop-low-coverage are doing very different jobs 💖

okay that's the whole page, mwah 😭🫶 this was cnvkit.py segment 0.9.10, dated July 2023, and it's filed under the manual name CNVKIT_SEGMENT. periodt. 💅✨