Skip to main content

Table 4 Comparison of cumulative contact and readout position weight matrices derived from two comparative models for the FNR and Giant transcription factors

From: Prediction of TF target sites based on atomistic models of protein-DNA complexes

SCOP v1.73
superfamily
TF [PDB id] Interface
identity
Resol.
(Å)
Robs E-valuecontacts E-valuereadout
Winged helix FNR [1cgp] 6/10 3 0.24 1.40E-01 6.03E-01
(6.35E-02)
[1.46E-04]
Leucine zipper domain Giant [1gu4] 6/12 1.8 0.23 7.10E-03 3.65E-02
(4.52E-04)
  1. In the first column we include the name of the structural superfamily of each TF according to SCOP [37]. The name of each model is followed by the PDB identifier of the template used to guide model building, enclosed in brackets. The following three columns contain the ratio of contact interface identity, the resolution and the R-value of the crystals respectively. The last two columns report the E-values resulting from STAMP [38] local alignments between the structure-based PWMs obtained with the method of Morozov and Siggia (E-valuecontacts) or our approach (E-valuereadout) and the corresponding cognate PWMs, extracted from the literature [49, 54]. Parenthesized E-values were calculated after sampling interface rotamers while bracketed E-values correspond to cumulative contact PWMs computed over the refined DNAPROT complex.