A transformation clustering algorithm and its application in polyribosomes structural profiling

Improvements in cryo-electron tomography sample preparation, electron-microscopy instrumentations, and image processing algorithms have advanced the structural analysis of macromolecules in situ. Beyond such analyses of individual macromolecules, the study of their interactions with functionally related neighbors in crowded cellular habitats, i.e. 'molecular sociology', is of fundamental importance in biology. Here we present a NEighboring Molecule TOpology Clustering (NEMO-TOC) algorithm. We optimized this algorithm for the detection and profiling of polyribosomes, which play both constitutive and regulatory roles in gene expression. Our results suggest a model where polysomes are formed by connecting multiple nonstochastic blocks, in which translation is likely synchronized.