Table 1

Asymptotic runtimes. Asymptotic runtimes for the unfiltered cmsearch procedure and three filters, including our Sifter approach. R is the number of families in the current Rfam release. L is the target sequence length, N is the window size. |f| is the number of sequences in all families. rp is the remaining set of covariance models that have to be searched after the application of the filtering program p, r R.

Program

Asymptotic runtime


CM

O(R·L·N3)

HMM

O(R·L·N2 + rHMM·L·N3)

BLAST

O(R·|fN + rBLAST·L·N3)

Sifter

O(N3 + rSifter·L·N3)


Janssen et al. BMC Bioinformatics 2008 9:131   doi:10.1186/1471-2105-9-131

Open Data