FITGCP(1)

💖 FITGCP(1) — the manual page, but make it cute 💖

okay bestie, pull up a chair, because this is fitgcp, a User Commands situation from section 1 of the manual, and she is literally out here fitting mixtures of probability distributions to genome coverage profiles. no cap. let's get into it. 🦋

✨ NAME ✨

fitgcp — fit mixtures of probability distributions to genome coverage profiles

🌸 SYNOPSIS 🌸

fitgcp [options] NAME

🎀 DESCRIPTION 🎀

So what is she doing, exactly? She fits mixtures of probability distributions to genome coverage profiles using an EM-like iterative algorithm. that's the whole vibe. 💅

The script uses a SAM file as input, and it parses the mapping information and creates a Genome Coverage Profile (GCP). The GCP gets written to a file, which is the slay part, because it means this step can be skipped the next time. then the user provides a mixture model, and that model gets fitted to the GCP. and like, you can also specify initial parameters for each model if you want to be extra. 👑

As output, the script generates a text file containing the final set of fit parameters plus additional information about the fitting process. There's also a log file, and it contains the current set of parameters in each step of the iteration. And if you request it? A plot of the GCP and the fitted distributions can be created. iconic behavior honestly. ✨

💅 OPTIONS 💅

• NAME — Name of SAM file to analyze. that's the one required thing, bestie, don't forget her. 🫶
• -h, --help — show this help message and exit
• -d DIST, --distributions=DIST — Distributions to fit. Here's the little legend: z -> zero; n: nbinom (MOM); N: nbinom (MLE); p: binom; t: tail. Default: zn
• -i STEPS, --iterations=STEPS — Maximum number of iterations. Default: 50
• -t THR, --threshold=THR — Set the convergence threshold for the iteration. She stops if the change between two iterations is less than THR. Default: 0.01
• -c CUTOFF, --cutoff=CUTOFF — Specifies a coverage cutoff quantile such that only coverage values below this quantile are considered. Default: 0.95
• -p, --plot — Create a plot of the fitted mixture model. Default: False
• -m MEAN, --means=MEAN — Specifies the initial values for the mean of each Poisson or Negative Binomial distribution. Usage: -m 12.4 -m 16.1 will specify the means for the first two non-zero/tail distributions. The default is calculated from the data.
• -a ALPHA, --alpha=ALPHA — Specifies the initial values for the proportion alpha of each distribution. Usage: for three distributions, -a 0.3 -a 0.3 specifies the proportions 0.3, 0.3 and 0.4. The default is equal proportions for all distributions.
• -l, --log — Enable logging. Default: False
• --view — Only view the GCP. Do not fit any distribution. Respects cutoff (-c). Default: False

real talk though: --view really does mean only view, okay? no fitting, no main character fitting arc, just looking. and it respects the -c cutoff, so behave accordingly. 🙈

🌈 a lil sign-off 🌈

This page is fitgcp, dated 2013-04-18, and it lives in the February 2014 release — FITGCP(1), User Commands, periodt. go fit some distributions, bestie, you're gonna ate. 💜✨🍓