Table 5.

fasta36 statistical options (-z)

-z	Statistics options
1	Default linear regression fit to average unrelated score vs. log (length) of library sequences
2	Censored (5%) maximum-likelihood estimation of λ, K for extreme-value distribution
3	Altschul-Gish pre-calculated λ, K
4	Variation of -z 1 that does additional iterations to censor high scores
5	Variation of -z 1 that also does regression of variance with log(length) of library sequences
6	Censored maximum-likelihood estimates that include a composition term Mott (1992)
11-16	Identical to 1–2, 4–6, but fits against scores from shuffled library sequences (required when libraries of related sequences are searched)
21-26	Identical to 1–2, 4–6, but calculates a second E()-value ( E2() based on shuffles of the top alignment scores
-Z#	calculate E()-value based on specified database size