Skip to main content

Table 5 Characteristics for different inclusion control strategies.

From: Subdivision of the MDR superfamily of medium-chain dehydrogenases/reductases through iterative hidden Markov model refinement

Strategy

Families

Sequences

Reiterations

Subsets

I, exclusive

92 (15)

10401

6

22 {16}

II, intermediate

86 (2)

11579

4

34 {14}

III, inclusive

85 (2)

11657

2

36 {15}

  1. Three different strategies for inclusion control were employed, affecting the number of resulting HMMs as well as their composition and relations. Roman numerals denote the different strategies, in increasing order of inclusiveness. Parenthesised numbers show the number of HMMs that were not affixed with the "reliable" qualifier, due to having too few non-spurious sequences in their seed sets. Numbers in braces denote the number of families having such subsets.
  2. In strategy I, all seed sequences failing the leave-one-out check were excluded. In strategy II, only seed sequences with domain scores lower than noise level were excluded. Additionally, for a left-out seed sequence to be excluded in strategy III, its domain score must fall below 90% of the lowest domain score among the remaining seed sequences.