Supplementary Information to "Nucleotides of Transcription Factor Binding
Sites Exert Interdependent Effects on the Binding Affinities of Transcription
Factors"
(Martha L. Bulyk, Philip L. F. Johnson, George M. Church)
Web Table 1
Correlation
coefficients between four types of interdependence control datasets - all
mononucleotide (1NN x N2N x NN3), first dinucleotide (12N x NN3),
second dinucleotide (1NN x N23), and trinucleotide (123)
– and values calculated by all three of our methods. The first type of
control had mononucleotide interdependence generated by making a random
mononucleotide weight matrix and multiplying the entries (1NN x N2N
x NN3) to calculate the Kaapp for each triplet.
The second type of control depended on the first dinucleotide (12N)
and the third mononucleotide (NN3); it was created from a random
matrix with one column of dinucleotides and another of mononucleotides.
Again, the entries were multiplied (12N x NN3) to yield the Kaapp’s.
The third type depended on the first mononucleotide (1NN) and the
second dinucleotide (N23) and was created using the same method
(1NN x N23). The fourth type of control (123) exhibited interdependence
at the triplet level.Each triplet was assigned
a random number, and then these numbers were normalized such that the sum
of the 64 triplets would equal one. Coefficients represent mean values
over 1000 sets of independent random controls.
|
Control type
|
1*2*3
|
12*3
|
1*23
|
|
1NNxN2NxNN3
|
1.00
|
1.00
|
1.00
|
|
12NxNN3
|
0.782
|
1.00
|
0.782
|
|
1NNxN23
|
0.780
|
0.780
|
1.00
|
|
123
|
0.370
|
0.529
|
0.527
|