Disordered proteins PONDRs

The success of the initial PONDRs based on small databases of disordered protein motivated attempts to improve predictor accuracy. The main limitation for such attempts has been and continues to be the lack of low-noise structural data for both ordered and disordered protein, where noise means ordered regions misclassified as disordered and vice versa. 

The accuracies of the various PONDRs were estimated (Table IV) by applying them to the ordered sequences in 0 PDBS25 as summarized in Table II and to the merged set of disordered proteins described in Table I. Overall, the prediction accuracy of each PONDR was much better on the 222,116 ordered residues of 0 PDBS25 than on the 18,833 residues of the merged disorder set. Thus, prediction of order generalized much better than prediction of disorder. 

III. PONDR Estimations of the Commonness of Intrinsically Disordered Proteins... 

The likelihood of multiple domains in a protein can also be estimated from the distribution of compositionally biased or predicted disordered regions, which tend to function as interdomain linkers. This type of subsequences can be identified using tools such as SEG [106], CAST [107], DisEMBL [108], PONDR [109], Disopred 2 [110], GlobPlot [111], DISPro, CARD [112], andNORsp [113], Armadillo, DLP2 [114], or DomCut can also be used to predict the linker regions of multidomain proteins from their primary sequence. 

---