Mutate DNA
Random substitutions, insertions and deletions at the rate you set, with every change listed.
Random substitutions, insertions and deletions at the rate you set, with every change listed.
It walks the sequence once and draws for every position: with probability equal to the substitution rate the residue is replaced, with probability equal to the indel rate a residue is inserted in front of it or the residue itself is deleted, and otherwise nothing happens. Every change is listed with its position on the original sequence, so the mutated copy and the list of what was done to it come out together.
Mutated copies are used as decoys and negative controls, as input for testing how far an aligner, a variant caller or a primer design survives divergence, and for building a small family of sequences at a known distance from a common ancestor.
The counts are random around those numbers, not fixed at them. On 720 bases a 2% rate means 14.4 substitutions on average, and single runs of 9 or 20 are ordinary. A 1% rate is roughly the divergence between human and chimpanzee genomes; 10% and above is far enough that a local aligner starts to lose short matches.
"Try an example" loads the 720 nt EGFP coding sequence. With the default 2% substitutions, 0.5% indels, Ti/Tv 2 and seed 1, the draw applies 12 mutations: 9 substitutions, 1 insertion and 2 deletions, so the copy comes back 719 nt long. The first three substitutions are A71G, G109A and C249T, all three of them transitions, and the full list with positions sits under Details. About 14.4 substitutions and 3.6 indels were expected, and the result says so next to what it applied: an independent draw per position scatters around its rate.
For DNA and RNA each base has one transition partner (A and G, C and T or U) and two transversion partners. The ratio you set decides how often the transition is taken; the two transversions then share what is left equally. Bases that are not A, C, G, T or U, such as N, are left alone and do not count towards the rate.
For protein, "Similar residues" weighs the 19 possible replacements by the BLOSUM62 score of the pair:
So isoleucine goes to valine or leucine most of the time and to tryptophan almost never, which is what conservative drift looks like. "Any residue" spreads the draw evenly over all 19 and gives a harsher, less realistic control.
For a sequence with no relationship to the original, useshuffle DNA or random DNA. To plan a specific change rather than a random one, usemutate for digest, and to see what a mutated coding sequence now translates to, translate DNA to protein.
Every position is drawn on its own, so the count is random around the rate you set: a 2% substitution rate on 720 bases gives 14.4 expected substitutions, and any single run lands a few either side of that. The tool prints what it actually applied next to what was expected. For an exact number of changes, raise the rate until the count you want is the average and keep the runs you like, or use one of the site-directed tools.
It sets how often a substitution stays within the purines (A to G) or the pyrimidines (C to T) rather than crossing between them. Each base has one transition partner and two transversion partners, so with a ratio of R the transition is chosen with probability R ÷ (R + 1). Real point mutations are transition-biased, with genomic estimates near 2 and higher in mitochondrial DNA, which is why 2 is the default rather than an even 1.
With "Similar residues" the replacement is drawn with probability proportional to 2 to the power of half the BLOSUM62 score for that pair, so the swaps that alignments accept most often are the ones drawn most often: I becomes V about 20% of the time and W only rarely. With "Any residue" all 19 other amino acids are equally likely, which is the harsher control.
At the same per-position rate, split evenly: half the indel events insert one random residue in front of the position, half delete the residue at it. Positions in the list are numbered on the sequence you pasted, before any change, so they still point at the original coordinates after an indel has shifted everything downstream.
Yes. The seed drives a deterministic generator, so the same sequence, rates and seed always produce the same mutations. Which positions change is drawn from one stream and what they change to from another, so editing the Ti/Tv ratio or the replacement model changes the new residues without moving the mutated positions. Press Mutate again for a fresh draw.