A Bayesian model for the identification of differentially expressed genes in Daphnia magna exposed to munition pollutants.

In this article we propose a Bayesian hierarchical model for the identification of differentially expressed genes in Daphnia magna organisms exposed to chemical compounds, specifically munition pollutants in water. The model we propose constitutes one of the very first attempts at a rigorous modeling of the biological effects of water purification. We have data acquired from a purification system that comprises four consecutive purification stages, which we refer to as "ponds," of progressively more contaminated water. We model the expected expression of a gene in a pond as the sum of the mean of the same gene in the previous pond plus a gene-pond specific difference. We incorporate a variable selection mechanism for the identification of the differential expressions, with a prior distribution on the probability of a change that accounts for the available information on the concentration of chemical compounds present in the water. We carry out posterior inference via MCMC stochastic search techniques. In the application, we reduce the complexity of the data by grouping genes according to their functional characteristics, based on the KEGG pathway database. This also increases the biological interpretability of the results. Our model successfully identifies a number of pathways that show differential expression between consecutive purification stages. We also find that changes in the transcriptional response are more strongly associated to the presence of certain compounds, with the remaining contributing to a lesser extent. We discuss the sensitivity of these results to the model parameters that measure the influence of the prior information on the posterior inference.

Bayesian feature selection for radiomics using reliability metrics.

Introduction: Imaging of tumors is a standard step in diagnosing cancer and making subsequent treatment decisions. The field of radiomics aims to develop imaging based biomarkers using methods rooted in artificial intelligence applied to medical imaging. However, a challenging aspect of developing predictive models for clinical use is that many quantitative features derived from image data exhibit instability or lack of reproducibility across different imaging systems or image-processing pipelines. Methods: To address this challenge, we propose a Bayesian sparse modeling approach for image classification based on radiomic features, where the inclusion of more reliable features is favored via a probit prior formulation. Results: We verify through simulation studies that this approach can improve feature selection and prediction given correct prior information. Finally, we illustrate the method with an application to the classification of head and neck cancer patients by human papillomavirus status, using as our prior information a reliability metric quantifying feature stability across different imaging systems.