GT-GENOMEDIFF(1)

✨ GT-GENOMEDIFF ✨ (that's the gt-genomediff page, bestie 💅)

okay so you opened the GenomeTools Manual and landed on this one — it's literally the girl that calculates Kr: pairwise distances between genomes. no cap. 🧬💖

---

🌸 NAME 🌸

gt-genomediff — Calculates Kr: pairwise distances between genomes. that's it, that's the whole job, and she does it gorgeously. periodt. ✨

---

🦋 SYNOPSIS 🦋

gt genomediff [option ...] (INDEX | -indexname NAME SEQFILE SEQFILE
[...])

like, you either hand her a ready-made INDEX or you make her build one with -indexname NAME plus a whole parade of SEQFILEs. choose your fighter. 👑

---

💜 DESCRIPTION 💜

she's got options, she's got range, let's get into it:

• -indextype [...] — specify type of index, one of: esa|pck|encseq. and heads up, encseq is an encoded sequence and an enhanced suffix array will be constructed only in memory. (default: encseq)
• -indexname [string] — Basename of encseq to construct. (default: undefined)
• -unitfile [filename] — specifies genomic units, see below for description. (default: undefined)
• -mirrored [yes|no] — virtually append the reverse complement of each sequence (default: no)
• -pl [value] — specify prefix length for bucket sort recommendation: use without argument; then a reasonable prefix length is automatically determined. (default: 0)
• -dc [value] — specify difference cover value (default: 0)
• -memlimit [string] — specify maximal amount of memory to be used during index construction (in bytes, the keywords MB and GB are allowed) (default: undefined)
• -scan [yes|no] — do not load esa index but scan it sequentially. (default: yes)
• -thr [value] — Threshold for difference (du, dl) in divergence calculation. default: 1e-9
• -abs_err [value] — absolute error for expected shulen calculation. default: 1e-5
• -rel_err [value] — relative error for expected shulen calculation. default: 1e-3
• -M [value] — threshold for minimum logarithm. default: DBL_MIN
• -v [yes|no] — be verbose (default: no)
• -help — display help for basic options and exit
• -help+ — display help for all options and exit
• -version — display version information and exit

---

The genomediff tool only accepts DNA input. yes, only DNA. don't try to feed her anything else, she will simply not. 🙈

When used with sequence files or encseq, an enhanced suffix array will be built in memory. The ESA will not be created completely, but construction will use -memlimit as a threshold and build it partwise, calculating the Shu-length for each part. so she chunks it up, no biggie. 💅

🎀 File format for option -unitfile 🎀 (in Lua syntax):

units = {
 genome1 = { "path/file1.fa", "file2.fa" },
 genome2 = { "file3.fa", "path/file4.fa" }
}

Give the path to the files as they were given to the encseq tool! as they were given, okay? don't get cute with it. You can use

$ gt encseq info INDEXNAME

to get a list of files in an encoded sequence. iconic little peek-behind-the-curtain moment. 👀

Comment lines in Lua start with -- and will be ignored.

See GTDIR/testdata/genomediff/unitfile1.lua for an example.

Options -pl, -dc and -memlimit are options to influence ESA construction.

---

📣 REPORTING BUGS 📣

Report bugs to <willrodt@zbh.uni-hamburg.de>. (yes, actually email them, they wanna know 🫶)

---

okay and for the little sign-off moment — this whole thing is GenomeTools 1.6.6, dated 11/09/2025, and that's the vibe check, bestie. go compute some pairwise distances, you absolute main character. 🦋🌸💖👑✨