Seeking reproducibility of findings is an essential part of research. This becomes a tedious and cumbersome task when big data with dependant features or multiple potentially interdependent data sources become available. Earlier work has shown that this is already the case for GWAIS compared to easier GWAS. The main problem is that replication studies make the assumption that research is repeated in all its important details, including population representation, but also analytic methodology and the type or number of measurements collected for sample. Here we will focus on study results in genetic epidemiology that can be represented as networks consisting of nodes and edges. By adopting a gene centric approach and by graphically presenting study results in a gene network, we aim to develop a generic strategy to compare results from studies that are analytically heterogeneous and for which different multi-space data panels are available. We will investigate how study specificity can be incorporated in the strategy, borrowing ideas from meta-analyses with random effects.