Fig. 1

Flowchart of the CirPred method. After obtaining the input data inclusive of a native template structure, a CP site, and a target sequence, seven major steps are performed. (1) The native template is processed to restore missing atoms and obtain the native amino acid sequence. (2) A pseudo-CP template structure and sequence are created according to the CP site. (3) A linker design procedure is carried out when necessary. (4) Sequence alignment between the pseudo-CP template and the target sequence is performed with native termini connected by the designed linker. (5) A coarse model is produced by comparative structure modeling. (6) The coarse model is refined by an algorithm that uses the CP site as a hinge to find the optimal orientation of domains (see “Methods” section). Each arrow represents a vector from the hinge to the center of mass of a domain. (7) Energy minimization and MD simulations are carried out to make the optimized model. In the web implementation, the linker design algorithm is activated only when the user does not provide the target sequence. In this situation, the target sequence is directly obtained from the template structure according to the CP site. The blue and red colors represent the N- and C-terminal proportions of a protein, respectively. The solid and dotted underlines indicate sequence fragments corresponding to the N- and C-terminal proportions of the native protein