CNVKIT_SEGMETRICS(1)

💖 CNVKIT_SEGMETRICS(1) 💖

okay bestie, sit down, hydrate, and let me tell you about cnvkit_segmetrics — the section of the cnvkit.py manual (User Commands, no less 👑) that computes segment-level metrics from bin-level log2 ratios. yes. we're doing math. slay.

---

✨ NAME ✨

cnvkit_segmetrics — compute segment-level metrics from bin-level log2 ratios. it's giving "let me judge your segments, statistically." 🎀

---

🌸 DESCRIPTION 🌸

the way this command wants to be called:

usage: cnvkit.py segmetrics [-h] -s SEGMENTS [--drop-low-coverage]
       [-o FILENAME] [--mean] [--median] [--mode]
              [--t-test] [--stdev] [--sem] [--mad] [--mse] [--iqr] [--bivar]
              [--ci] [--pi] [-a ALPHA] [-b BOOTSTRAP] [--smooth-bootstrap]
              cnarray

yes, that's a lot of flags. no, you don't need all of them. but we're covering all of them anyway because every single detail matters and i'm not about to gatekeep. 💅

---

🦋 positional arguments: 🦋

• cnarray — the bin-level copy ratio data file (*.cnn, *.cnr). this is your input, your little data baby, your main character. feed it in.

---

💜 options: 💜

• -h, --help — shows this help message and exits. the classic. the icon. press it when you're lost (you won't be, because i got you 🫶).
• -s SEGMENTS, --segments SEGMENTS — the segmentation data file (*.cns, which is what the segment command spits out). yes this one is required, no you cannot skip it.
• --drop-low-coverage — yeets the very-low-coverage bins out of the picture before the calculations happen, so poor-quality tumor samples don't get a sad negative bias dragging their numbers down. protecting the vibe, fr fr. 🙈
• -o FILENAME, --output FILENAME — the name of the output table file. name it something cute.

---

⭐ Statistics available: ⭐

like, the whole menu of stats you can ask for. pick your fighter(s):

• --mean — mean log2 ratio, unweighted. the plain one. the reliable bestie.
• --median — median. the middle child (affectionate). 💖
• --mode — mode, a.k.a. the peak density of the bin log2 ratios. the most popular kid in school.
• --t-test — one-sample t-test of the bin log2 ratios versus 0.0. is it different from zero? we're asking. dramatically. 😭
• --stdev — standard deviation. how spread out is everybody.
• --sem — standard error of the mean. not the same as stdev, don't @ me.
• --mad — median absolute deviation, standardized. yes, standardized. it's got range. ✨
• --mse — mean squared error.
• --iqr — inter-quartile range. the middle 50%'s whole deal.
• --bivar — Tukey's biweight midvariance. say it with your chest. 🗣️
• --ci — confidence interval, computed by bootstrap. team effort energy.
• --pi — prediction interval.
• -a ALPHA, --alpha ALPHA — the level used to estimate the confidence and prediction intervals; use it alongside --ci and --pi. default is 0.05. ⭐
• -b BOOTSTRAP, --bootstrap BOOTSTRAP — how many bootstrap iterations to run when estimating the confidence interval; use it alongside --ci. default is 100. 🔥
• --smooth-bootstrap — applies Gaussian noise to the bootstrap samples (a.k.a. smoothed bootstrap) to estimate the confidence interval; use it alongside --ci. sprinkle a lil chaos on it, as a treat. 🌈

heads up bestie: -a/--alpha is a --ci/--pi thing, -b/--bootstrap is a --ci thing, and --smooth-bootstrap is also a --ci thing. pair them up correctly or they just sit there doing nothing. periodt. 💅

---

and that's the whole page, no notes, nothing skipped, nothing invented. every flag, every default, every file extension, exactly as written. i would never do you dirty like that. 🧚

this has been cnvkit.py segmetrics, version 0.9.10, from July 2023, straight out of User Commands. mwah. go compute some segment metrics, you gorgeous genius. 💖✨🎀