Abstract Large neuroimaging datasets, including information about structural (SC) and functional connectivity (FC), play an increasingly important role in clinical research, where they guide the design of algorithms for automated stratification, diagnosis or prediction. A major obstacle is, however, the problem of missing features (e.g., lack of concurrent DTI SC and resting-state fMRI FC measurements for many of the subjects). We propose here to address the missing connectivity features problem by introducing strategies based on computational whole-brain network modeling. Using two datasets, the ADNI dataset and a healthy aging dataset, for proof-of-concept, we demonstrate the feasibility of virtual data completion (i.e., inferring “virtual FC” from empirical SC or “virtual SC” from empirical FC), by using self-consistent simulations of linear and nonlinear brain network models. Furthermore, by performing machine learning classification (to separate age classes or control from patient subjects) we show that algorithms trained on virtual connectomes achieve discrimination performance comparable to when trained on actual empirical data; similarly, algorithms trained on virtual connectomes can be used to successfully classify novel empirical connectomes. Completion algorithms can be combined and reiterated to generate realistic surrogate connectivity matrices in arbitrarily large number, opening the way to the generation of virtual connectomic datasets with network connectivity information comparable to the one of the original data. Significance statement Personalized information on anatomical connectivity (“structural connectivity”, SC) or coordinated resting state activation patterns (“functional connectivity’, FC) is a source of powerful neuromarkers to detect and track the development of neurodegenerative diseases. However, there are often “gaps” in the available information, with only SC (or FC) being known but not FC (or SC). Exploiting whole-brain modelling, we show that gap in databases can be filled by inferring the other connectome through computational simulations. The generated virtual connectomic data carry information analogous to the one of empirical connectomes, so that machine learning algorithms can be trained on them. This opens the way to the release in the future of cohorts of “virtual patients”, complementing traditional datasets in data-driven predictive medicine.