Preeclampsia is a multifactorial pregnancy complication diagnosed by hypertension and proteinuria with serious and unpredictable adverse outcomes such as HELLP syndrome, eclampsia, stroke or intra-uterine growth retardation, to name a few. It is one of the major causes worldwide of extreme prematurity and maternal deaths. After initial diagnosis, because of a lack of specific biomarkers or adequate combination of biomarkers, preeclamptic pregnancy outcome is hard to predict, limiting the panel of therapeutic strategies offered to the clinicians and leading to emergencies unexpected situations with adverse effects for mother and baby. Therefore, there is an urgent need to develop efficient methods to help to predict complications of preeclampsia. We propose to investigate the properties of a serum-based diagnostic strategy, which uses a selective array sensing rather than specific, mimicking olfaction mechanism: each sensing element of the array interacts differentially with serum components, generating a unique pattern tied back to the serum composition. From a chemical side, our sensing array relies on fluorescent conjugated cucurbiturils molecules (i.e. barrel shaped supramolecules) which can interact differentially with biomolecules in the serum thanks to a panel of non-specific binding modes (host/guest chemistry, electrostatic, hydrophobic, size exclusion). As a consequence of the binding, fluorescence is modulated which provide a readable output for the assay. The chemical structure of the cucurbituril family makes this molecule a good choice to develop original sensor array since performance of such array is linked not only to the number of sensors but also to its ability to capture diversity in binding. Droplet based high-throughput microfluidic device will be developed to evaluate this high-density probe sensing array with a rapid analysis and detection manner. Microfluidic device only requires small sample volumes, which is consistent with the use of clinical samples. A pre-existing cohort of preeclamptic and non-preeclamptic samples will be assessed with the sensor-array in the microfluidic device. The results of the assay along with available clinical and biological data implementation, including women and foetus outcomes, will be processed by a suitable statistical approach to obtain classifiers, either supervised or unsupervised, to predict preeclampsia outcomes and to explore preeclampsia complexity. In addition, focus on preeclampsia is also of high value to evaluate this kind of methodology thanks to the diversity of both this multifactorial disease and its associated severe outcomes. Indeed, a large diversity of clinical and biological inputs is required to evaluate the arrays ability to explore the space of pathological states. The success of this methodology will pave the way for the study of other multifactorial and complex diseases, such as sepsis or acute respiratory distress symptoms.