What it does
A k-mer is a word of k letters taken from a sequence: the read ACGTAC holds the 3-mers ACG, CGT, GTA and TAC. Counting every k-mer in a set of reads is a quick, assembly-free way to see what is in them:
- The spectrum plots how many distinct k-mers were seen once, twice, three times and so on. For good whole-genome reads it has a tall wall at 1 (k-mers carrying a sequencing error), a valley, then a hump at the typical depth of the genome, with a tail of repeats to the right.
- Most common k-mers shows the top 20 with their share, which is where adapters, primers and repeats stand out.
- Look up a k-mer gives the count of any word you type.
- Genome size (rough) is the number of k-mers counted from the valley on, divided by the depth at the peak. It assumes one copy of each chromosome set and even coverage, so treat it as a first guess, not a measurement.
Options
- k is the word length, from 1 to 25. Longer words are more likely to be unique in a genome; 21 is a common choice. Small k (3 to 8) is useful for word use and composition in short sequences.
- Count a word and its reverse complement together merges ACG with CGT. Leave it on for reads, which come from both strands; turn it off for strand-specific data.
Letters other than A, C, G and T (such as N) break the sequence, and no word is counted across them. Up to 5 million distinct k-mers are kept; the page says so if a file has more.
Worked example
The example is 10,000 reads of 100 bases simulated from a made-up 40,000-base genome that holds a 600-base repeat three times, with 1 % random errors. With k = 21 and reverse complements together, the 1,000,000 bases give 800,000 k-mers, 188,111 of them distinct. 145,234 are seen only once: nearly all of them hold an error. The valley is at 4×, the main peak at 15× (25× read coverage, less the k-mers lost to read ends and errors), and the genome size estimate is about 43,100 bases, within 8 % of the true 40,000. The repeat’s k-mers form a small second hump around 45 to 50×, about three times the peak. Turn off reverse complements and the peak halves, because each strand is now counted on its own.
FAQ
Is my data uploaded? No. Files are read piece by piece in your browser, so even large or private sequencing data stays on your device. There are no ads on this page.
Why is my genome size estimate off? Heterozygous genomes have a second, half-height peak, polyploid ones more, and low coverage (under about 15×) blurs the peak. Dedicated tools fit a model to the whole spectrum for that; this page gives the simple peak estimate.
Can I use an assembly? Yes, for counting and the most common k-mers, but an assembly has no error wall or depth hump, so no genome size is given.
Where the numbers come from
- k-mer counting, spectrum and the peak-based genome size written by fibub. Counts, spectrum and most common k-mers checked in the tests and against a separate Python Counter script for k = 5, 21 and 25, with and without reverse complements.
- The example reads are simulated: a seeded random 40,000-base genome with a made-up repeat, 1 % random errors.
Standard scientific values, such as constants, atomic weights and melting points, are facts found in every textbook; we name where ours were checked. If you are an author or publisher and want a value, a credit or a table changed or removed, tell us and we will act quickly.
Last updated 2026-11-09.