# Statistical similarity on sequences

What are the methods that can be used for evaluating the similarity of the sequences? Lets say that we have three sequences; I want to be able to say that "statistically" the sequence number 1 and the sequence number 3 are similar, but the sequence number 2 is "statistically" distinct. Like a p-value for statistical tests.

• What type of sequences are we talking about (continuous variable or discrete? Which length?)
– nico
Jul 25, 2013 at 17:03
• @nico I have a sequence of different sizes from 4 to a hundred, but it has a skewed distribution (the mod is around 20). The variables are discrete (categorical) and I am using the levenshtein distance to calculate the distance between them, but I am not sure how can I say that how these sequences are statistically similar to each other
– POD
Jul 25, 2013 at 17:12
• This sounds quite similar to biological sequence alignment, which has been studied quite a bit. You can find some information here: ncbi.nlm.nih.gov/BLAST/tutorial Jul 25, 2013 at 17:18