Global healthcare systems are challenged by the COVID-19 pandemic. There is a need to optimize allocation of treatment and resources in intensive care, as clinically established risk assessments such as SOFA and APACHE II scores show only limited performance for predicting the survival of severely ill COVID-19 patients. Additional tools are also needed to monitor treatment, including experimental therapies in clinical trials. Comprehensively capturing human physiology, we speculated that proteomics in combination with new data-driven analysis strategies could produce a new generation of prognostic discriminators. We studied two independent cohorts of patients with severe COVID-19 who required intensive care and invasive mechanical ventilation. SOFA score, Charlson comorbidity index, and APACHE II score showed limited performance in predicting the COVID-19 outcome. Instead, the quantification of 321 plasma protein groups at 349 timepoints in 50 critically ill patients receiving invasive mechanical ventilation revealed 14 proteins that showed trajectories different between survivors and non-survivors. A predictor trained on proteomic measurements obtained at the first time point at maximum treatment level (i.e. WHO grade 7), which was weeks before the outcome, achieved accurate classification of survivors (AUROC 0.81). We tested the established predictor on an independent validation cohort (AUROC 1.0). The majority of proteins with high relevance in the prediction model belong to the coagulation system and complement cascade. Our study demonstrates that plasma proteomics can give rise to prognostic predictors substantially outperforming current prognostic markers in intensive care.

Genetic and non-genetic factors contribute to breast cancer development. An epigenome-based signature capturing these components in easily accessible samples could identify women at risk. Here, we analyse the DNA methylome in 2,818 cervical, 357 and 227 matched buccal and blood samples respectively, and 42 breast tissue samples from women with and without breast cancer. Utilising cervical liquid-based cytology samples, we develop the DNA methylation-based Women’s risk IDentification for Breast Cancer index (WID-BC-index) that identifies women with breast cancer with an AUROC (Area Under the Receiver Operator Characteristic) of 0.84 (95% CI: 0.80–0.88) and 0.81 (95% CI: 0.76–0.86) in internal and external validation sets, respectively. CpGs at progesterone receptor binding sites hypomethylated in normal breast tissue of women with breast cancer or in BRCA mutation carriers are also hypomethylated in cervical samples of women with poor prognostic breast cancer. Our data indicate that a systemic epigenetic programming defect is highly prevalent in women who develop breast cancer. Further studies validating the WID-BC-index may enable clinical implementation for monitoring breast cancer risk.

Background: Earlier detection of pancreatic ductal adenocarcinoma (PDAC) is key to improving patient outcomes, as it is mostly detected at advanced stages which are associated with poor survival. Developing non-invasive blood tests for early detection would be an important breakthrough.Methods: The primary objective of the work presented here is to use a dataset that is prospectively collected, to quantify a set of cancer-associated proteins and construct multi-marker models with the capacity to predict PDAC years before diagnosis. The data used is part of a nested case-control study within the UK Collaborative Trial of Ovarian Cancer Screening and is comprised of 218 samples, collected from a total of 143 post-menopausal women who were diagnosed with pancreatic cancer within 70 months after sample collection, and 249 matched non-cancer controls. We develop a stacked ensemble modelling technique to achieve robustness in predictions and, therefore, improve performance in newly collected datasets.Results: Here we show that with ensemble learning we can predict PDAC status with an AUC of 0.91 (95% CI 0.75–1.0), sensitivity of 92% (95% CI 0.54–1.0) at 90% specificity, up to 1 year prior to diagnosis, and at an AUC of 0.85 (95% CI 0.74–0.93) up to 2 years prior to diagnosis (sensitivity of 61%, 95% CI 0.17–0.83, at 90% specificity).Conclusions: The ensemble modelling strategy explored here outperforms considerably biomarker combinations cited in the literature. Further developments in the selection of classifiers balancing performance and heterogeneity should further enhance the predictive capacity of the method.

Parenclitic networks provide a powerful and relatively new way to coerce multidimensional data into a graph form, enabling the application of graph theory to evaluate features. Different algorithms have been published for constructing parenclitic networks, leading to the question—which algorithm should be chosen? Initially, it was suggested to calculate the weight of an edge between two nodes of the network as a deviation from a linear regression, calculated for a dependence of one of these features on the other. This method works well, but not when features do not have a linear relationship. To overcome this, it was suggested to calculate edge weights as the distance from the area of most probable values by using a kernel density estimation. In these two approaches only one class (typically controls or healthy population) is used to construct a model. To take account of a second class, we have introduced synolytic networks, using a boundary between two classes on the feature-feature plane to estimate the weight of the edge between these features. Common to all these approaches is that topological indices can be used to evaluate the structure represented by the graphs. To compare these network approaches alongside more traditional machine-learning algorithms, we performed a substantial analysis using both synthetic data with a priori known structure and publicly available datasets used for the benchmarking of ML-algorithms. Such a comparison has shown that the main advantage of parenclitic and synolytic networks is their resistance to over-fitting (occurring when the number of features is greater than the number of subjects) compared to other ML approaches. Secondly, the capability to visualise data in a structured form, even when this structure is not a priori available allows for visual inspection and the application of well-established graph theory to their interpretation/application, eliminating the “black-box” nature of other ML approaches.

Fetal growth restriction (FGR) is a leading cause of perinatal morbidity and mortality. Altered placental formation and functional capacity are major contributors to FGR pathogenesis. Relating placental structure to function across the placenta in healthy and FGR pregnancies remains largely unexplored but could improve understanding of placental diseases. We investigated integration of these parameters spatially in the term human placenta using predictive modelling. Systematic sampling was able to overcome heterogeneity in placental morphological and molecular features. Defects in villous development, elevated fibrosis, and reduced expression of growth and functional marker genes (IGF2, VEGA, SLC38A1, and SLC2A3) were seen in age-matched term FGR versus healthy control placentas. Characteristic histopathological changes with specific accompanying molecular signatures could be integrated through computational modelling to predict if the placenta came from a healthy or FGR pregnancy. Our findings yield new insights into the spatial relationship between placental structure and function and the etiology of FGR.

This paper introduces a novel approach for classifying multidimensional physiological and clinical data using Synolitic Graph Neural Networks (SGNNs). SGNNs are particularly good for addressing the challenges posed by high-dimensional datasets, particularly in healthcare, where traditional machine learning and Artificial Intelligence methods often struggle to find global optima due to the “curse of dimensionality”. To apply Geometric Deep Learning we propose a synolitic or ensemble graph representation of the data, a universal method that transforms any multidimensional dataset into a network, utilising only class labels from training data. The paper demonstrates the effectiveness of this approach through two classification tasks: synthetic and fMRI data from cognitive tasks. Convolutional Graph Neural Network architecture is then applied, and the results are compared with established machine learning algorithms. The findings highlight the robustness and interpretability of SGNNs in solving complex, high-dimensional classification problems.