Computational approaches such as Principal Component Analysis (PCA) and Elastic Network Model-Normal Mode Analysis (ENM-NMA) are proving to be of great value in investigating relevant biological problems linked to slow motions with no demand in computer power. In this study, these approaches have been coupled to the graph theory-based Protein Structure Network (PSN) analysis to dissect functional dynamics and structural communication in the Dbl family of Rho Guanine Nucleotide Exchange Factors (RhoGEFs). They are multidomain proteins whose common structural feature is a DH-PH tandem domain deputed to the GEF activity that makes them play a central role in cell and cancer biology. While their common GEF action is accomplished by the DH domain, their regulatory mechanisms are highly variegate and depend on the PH and the additional domains as well as on interacting proteins. Major evolutionary-driven deformations as inferred from PCA concern the α6 helix of DH that dictates the orientation of the PH domain. Such deformations seem to depend on the mechanisms adopted by the GEF to prevent Rho binding, i.e. functional specialization linked to autoinhibition. In line with PCA, ENM-NMA indicates α6 and the linked PH domain as the portions of the tandem domain holding almost the totality of intrinsic and functional dynamics, with the α6/β1 junction acting as a hinge point for the collective motions of PH. In contrast, the DH domain holds a static scaffolding and hub behavior, with structural communication playing a central role in the regulatory actions by other domains/proteins. Possible allosteric communication pathways involving essentially DH were indeed found in those RhoGEFs acting as effectors of small or heterotrimeric RasGTPases. The employed methodology is suitable for deciphering structure/dynamics relationships in large sets of homologous or analogous proteins.