ESTIMATEHYPERPAR(1)

💖 ESTIMATEHYPERPAR 💖

okay bestie, so this is estimateHyperPar from the estimateHyperPar manual (section 1, user commands, iykyk) and she is literally here to ✨estimate expression dependent hyperparameters for bitseq✨. no cap. let's get into it 🫶

---

✨ SYNOPSIS ✨

estimateHyperPar -o <outFileName>  [OPTIONS] [sampleFiles]

like. that's the whole vibe. <outFileName> is mandatory, everything else is negotiable 💅

---

🌸 DESCRIPTION 🌸

so what she does is: estimate expression dependent hyperparameters from the dataset. that's the whole job. main character energy honestly 👑

now, <sampleFiles> — bestie, listen — should contain transposed MCMC samples from replicates. and to distinguish conditions you separate them with a single C between the groups. like this, fr fr:

samplesC1-R1.rpkm samplesC1-R2.rpkm C samplesC2-R1.rpkm samplesC2-R2.rpkm

that lone C is the vibe check between conditions. do not skip it, she will not know what you want 😭

---

🎀 OPTIONS 🎀

• --help — shows you this help information. the classic. she's helpful like that 🫶
• --exThreshold=<exT> — the threshold of the lowest expression for which the estimation actually gets done. (default: -5)
• -g <groupsN>, --groupsNumber=<groupsN> — how many groups of transcript of similar size you want. (default: 200)
• -l <lambda0>, --lambda0=<lambda0> — precision scaling parameter lambda0. yes that is the technical name, no I will not apologise. (default: 2)
• --lowess-f=<lowess-f> — parameter F for lowess smoothing, which specifies the amount of smoothing ✨ (default: 0.2)
• --lowess-steps=<lowess-steps> — parameter Nsteps for lowess smoothing, which specifies the number of iterations. (default: 5)
• --meanFile=<meanFileName> — name of the file containing the joint mean and variance. 👑
• --noforce — do not force smoothing of the parameters. (default: Off)
• --norm=<normalization> — normalization constants for each input file, handed over as a comma separated list of doubles (e.g. 1.0017,1.0,0.9999).
• -o <outFileName>, --outFile=<outFileName> — name of the output file. this one is required bestie, she said what she said 💅
• -p <paramsAllFileName>, --paramsAllFile=<paramsAllFileName> — name of the file to which to store all parameter values generated prior to lowess smoothing.
• -s <samplesN>, --samplesNumber=<samplesN> — number of samples generated for each group. (default: 2)
• --seed=<seed> — random initialization seed. for when you need the chaos to be reproducible 🌈
• -S, --smoothOnly — the input file contains previously sampled hyperparameters which should be smoothed only. (default: Off)
• -v, --verbose — verbose output. (default: Off)
• -V, --veryVerbose — more verbose output. she's chatty in ascending order, iconic. (default: Off)

---

🦋 SIGN-OFF 🦋

and that's the whole page, bestie!! this was estimateHyperPar version 0.7.5+dfsg, dated September 2021. go estimate those hyperparameters, you absolute legend 💖✨🎀