Personal Data Management Systems (PDMS) are flourishing, boosted by legal and technical means like smart disclosure, data portability and data altruism. A PDMS allows its owner to easily collect, store and manage data, directly generated by his devices, or resulting from his interactions with companies or administrations. PDMSs unlock innovative usages by crossing multiple data sources from one or many users, thus requiring aggregation primitives. Indeed, they are essential to compute statistics on user data, but are also a fundamental building block for machine learning algorithms. This paper proposes a protocol allowing secure aggregation in a massively distributed PDMS environment which adapts to selective participation and PDMSs characteristics and is reliable w.r.t. failure, with no compromise on accuracy. Preliminary experiments show the effectiveness of our protocol which can handle PDMSs characteristics in terms of communication speed or CPU resources and adjust the aggregation strategy to the estimated participation.