Skip to main content

Table 1 Binary patterns for nucleotides, A, U, C, G, and a loop/bulge structure, denoted by L, in PHDCleav [14] and LBSizeCleav with k ones based on sequences and predicted secondary structures

From: LBSizeCleav: improved support vector machine (SVM)-based prediction of Dicer cleavage sites using loop/bulge length

Mapping

 

Sequence

Structure

PHDCleav

A

[1,0,0,0]

[1,0,0,0]

 

U

[0,1,0,0]

[0,1,0,0]

 

C

[0,0,1,0]

[0,0,1,0]

 

G

[0,0,0,1]

[0,0,0,1]

 

L

−

[0,0,0,0]

Extended PHDCleav

A

−

[1,0,0,0,0]

 

U

 

[0,1,0,0,0]

 

C

 

[0,0,1,0,0]

 

G

 

[0,0,0,1,0]

 

L

 

[0,0,0,0,1]

LBSizeCleav

A

−

[1,0,0,0,0, … 0]

 

U

 

[0,1,0,0,0, … 0]

 

C

 

[0,0,1,0,0, … 0]

 

G

 

[0,0,0,1,0, … 0]

 

L

 

\([0, 0, 0, 0, 0, \dots, 0, \overbrace {1, \dots, 1}^{k}, 0, \dots ]\)

  1. In PHDCleav binary patterns each nucleotide is represented by a 4-dimensional vector, and in PHDCleav Extended patterns each nucleotide is represented by a 5-dimensional vector, while in LBSizeCleav the dimension of the vector is 3+k+N, in which N denotes the maximum number of length of loop/bulges among all the pre-miRNAs in the training dataset