Skip to main content

Table 11 LEA classification rule set induced by supervised learning

From: LEAping to conclusions: A computational reanalysis of late embryogenesis abundant proteins and their possible roles

Group

Rule

2a

H < = 0.15 and aromatic > = 0.077 and min_hyph < = -1.97 and charged < = 0.42

2b

L > = 0.23 and H < = 0.3 and ave_hyph > = -1.233 and ave_hyph < = -0.978

2c

aromatic > = 0.077 and min_hyph < = -2.743 and charged > = 0.4

3

H > = 0.34

1

E > = 0.02 and ave_hyph < = -1.241

LE5

max_hyph > = 1.0 and ave_hyph < = -0.3

LE14

aliphatic > = 0.25

6

H > = 0.25 and max_hyph > = 0.5

4

Otherwise

  1. The rules are to be applied in a top-down, if .. else if, manner, so, for example, if the percentage of predicted helical conformation (expressed as a number in the range 0 .. 1.0) is greater than or equal to 0.34 then the protein is classified as a Group 3 LEA protein, but only if each of the rules above has failed, e.g. because the percentage of aromatic residues is less than 0.077 in the Group 2 rules. min_hyph and max_hyph are, respectively, the values of minimum and maximum hydrophobicity windows, while ave_hyph is the average across all the hydrophobicity windows. H, E and L refer, respectively, to the percentage composition of amino acids that are found by ProteinPredict (in four-state mode) to be alpha-helical, beta-sheet or loop.