Next-generation cancer and oncology research needs to take full advantage of the multimodal structured, or graph, information, with the graph data types ranging from molecular structures to spatially resolved imaging and digital pathology, biological networks, and knowledge graphs. Graph Neural Networks (GNNs) efficiently combine the graph structure representations with the high predictive performance of deep learning, especially on large multimodal datasets. In this review article, we survey the landscape of recent (2020–present) GNN applications in the context of cancer and oncology research, and delineate six currently predominant research areas. We then identify the most promising directions for future research. We compare GNNs with graphical models and “non-structured” deep learning, and devise guidelines for cancer and oncology researchers or physician-scientists, asking the question of whether they should adopt the GNN methodology in their research pipelines.

Bayesian Network (BN) modeling is a prominent and increasingly popular computational systems biology method. It aims to construct network graphs from the large heterogeneous biological datasets that reflect the underlying biological relationships. Currently, a variety of strategies exist for evaluating BN methodology performance, ranging from utilizing artificial benchmark datasets and models, to specialized biological benchmark datasets, to simulation studies that generate synthetic data from predefined network models. The last is arguably the most comprehensive approach; however, existing implementations often rely on explicit and implicit assumptions that may be unrealistic in a typical biological data analysis scenario, or are poorly equipped for automated arbitrary model generation. In this study, we develop a purely probabilistic simulation framework that addresses the demands of statistically sound simulations studies in an unbiased fashion. Additionally, we expand on our current understanding of the theoretical notions of causality and dependence / conditional independence in BNs and the Markov Blankets within.

Cancer immunotherapy, specifically immune checkpoint blockade, has been found to be effective in the treatment of metastatic cancers. However, only a subset of patients achieve clinical responses. Elucidating pretreatment biomarkers predictive of sustained clinical response is a major research priority. Another research priority is evaluating changes in the immune system before and after treatment in responders vs. nonresponders. Our group has been studying immune networks as an accurate reflection of the global immune state. Flow cytometry (FACS, fluorescence-activated cell sorting) data characterizing immune cell panels in peripheral blood mononuclear cells (PBMC) from gastroesophageal adenocarcinoma (GEA) patients were used to analyze changes in immune networks in this setting. Here, we describe a novel computational pipeline to perform secondary analyses of FACS data using systems biology/machine learning techniques and concepts. The pipeline is centered around comparative Bayesian network analyses of immune networks and is capable of detecting strong signals that conventional methods (such as FlowJo manual gating) might miss. Future studies are planned to validate and follow up the immune biomarkers (and combinations/interactions thereof) associated with clinical responses identified with this computational pipeline.

Bayesian Network (BN) modeling is a prominent methodology in computational systems biology. However, the incommensurability of datasets frequently encountered in life science domains gives rise to contextual dependence and numerical irregularities in the behavior of model selection criteria (such as MDL, Minimum Description Length) used in BN reconstruction. This renders model features, first and foremost dependency strengths, incomparable and difficult to interpret. In this study, we derive and evaluate a model selection principle that addresses these problems. The objective of the study is attained by (i) approaching model evaluation as a misspecification problem, (ii) estimating the effect that sampling error has on the satisfiability of conditional independence criterion, as reflected by Mutual Information, and (iii) utilizing this error estimate to penalize uncertainty with the novel Minimum Uncertainty (MU) model selection principle. We validate our findings numerically and demonstrate the performance advantages of the MU criterion. Finally, we illustrate the advantages of the new model evaluation framework on real data examples. The new BN model selection principle successfully overcomes performance irregularities observed with MDL, offers a superior average convergence rate in BN reconstruction, and improves the interpretability and universality of resulting BNs, thus enabling direct inter-BN comparisons and evaluations.

Next-generation cancer and oncology research needs to take full advantage of the multi-modal structured, or graph, information, with the graph datatypes ranging from molecular structures to spatially resolved imaging and digital pathology to biological networks to knowledge graphs. Graph Neural Networks (GNNs) efficiently combine the graph structure representations with the high predictive performance of deep learning, especially on the large multi-modal datasets. In this review article, we survey the landscape of recent (2020-present) GNN applications in the context of cancer and oncology research, and delineate six currently predominant research areas. Subsequently, we identify the most promising directions for future research. We compare GNNs with graphical models and "non-structured" deep learning, and devise the guidelines for cancer and oncology researchers or physician-scientists asking the question of whether they should adopt the GNN methodology in their research pipelines.

Identifying target-specific drugs remains a challenge in pharmacology, especially for highly homologous proteins such as dopamine receptors D ₂ R and D ₃ R. Differences in target-specific cryptic druggable sites for such receptors arise from the distinct conformational ensembles underlying their dynamic behavior. While Molecular Dynamics (MD) simulations has emerged as a powerful tool for dissecting protein dynamics, the sheer volume of MD data requires scalable and unbiased data analysis strategies to pinpoint residue communities regulating conformational state ensembles. We present the Dynamically Resolved Universal Model for BayEsiAn network Tracking (DRUMBEAT) interpretable machine learning algorithm and validate it by identifying residue communities that enable the deactivation of the β ₂ -adrenergic receptor. Further, upon analyzing dopamine receptor dynamics we identify distinct and non-conserved residue communities around the contacts F170 ^4.62 _F172 ^ECL2 and S146 ^4.38 _G141 ^34.56 that are specific to D ₃ R conformational transitions compared to D ₂ R. This information can be tapped to design subtype-specific drugs for neuropsychiatric and substance use disorders.