Skip to main content

Table 3 Class I vs genes: The four most discriminating dinucleotides when comparing the HMM trained on class I TEs and the HMM trained on genes for each state of the HMM. E1, E2, E3 are the three coding states; I is the non-coding state. Log2 of the frequency ratio is given in parentheses. A value of +1 (resp. -1) thus indicates a twofold increase (resp. decrease) of the frequency in the TE model.

From: Detection of transposable elements by their compositional bias

HMM state

A. thaliana

E1

CG (-0.79)

CC (-0.32)

AC (0.20)

TC (0.17)

E2

CA (0.28)

TA (0.26)

AC (0.17)

TC (-0.14)

E3

CG (-0.31)

GC (-0.30)

TC (-0.23)

CC (-0.18)

I

CG (2.23)

CC (1.76)

TT (-1.62)

TA (-1.35)

HMM state

C. elegans

E1

TG (-1.71)

CT (1.12)

CC (0.96)

AG (-0.85)

E2

GA (-1.66)

GG (-1.17)

TA (0.80)

TC (0.73)

E3

TA (1.03)

CC (0.90)

TT (-0.68)

AT (-0.67)

I

TT (-1.73)

CG (1.22)

TA (-0.87)

CC (0.87)

HMM state

D. melanogaster

E1

AG (0.89)

AA (0.87)

TA (0.66)

GG (-0.58)

E2

AG (0.59)

TG (0.38)

GA (-0.25)

GG (-0.24)

E3

TA (1.06)

CA (0.73)

GG (0.72)

TG (-0.68)

I

GG (-1.03)

GC (-0.56)

AA (0.55)

CG (-0.51)