Clustering

At present, there is no command that directly accesses the clustering facility of PLOP, which is based on the K-means algorithm. However, the clustering algorithm is employed by both the "loop predict" and "helix predict" commands, and the parameters are identical, so the clustering parameters have been "nested" inside the loop and helix prediction parameters (see below). Here are the relevant clustering parameters:

iter[ations] integer

nclust integer

emax real

nretry integer

add_clust yes/no

The basic idea is that the clustering algorithm takes as input a series of different conformations (whether loop or helix positions), and then groups them into "nclust" different conformations.