Distance-Metric Learning for Personalized Survival Analysis.

Personalized time-to-event or survival prediction with right-censored outcomes is a pervasive challenge in healthcare research. Although various supervised machine learning methods, such as random survival forests or neural networks, have been adapted to handle such outcomes effectively, they do not provide explanations for their predictions, lacking interpretability. In this paper, an alternative method for survival prediction by weighted nearest neighbors is proposed. Fitting this model to data entails optimizing the weights by learning a metric. An individual prediction of this method can be explained by providing the user with the most influential data points for this prediction, i.e., the closest data points and their weights. The strengths and weaknesses in terms of predictive performance are highlighted on simulated data and an application of the method on two different real-world datasets of breast cancer patients shows its competitiveness with established methods.

The importance of diet-related effects of the antibiotic ciprofloxacin on the leaf-shredding invertebrate Gammarus fossarum (Crustacea; Amphipoda).

Antibiotics may constitute a risk for aquatic detritivorous macroinvertebrates (i.e., shredders) via waterborne and dietary antibiotic exposure. In addition, antibiotics can alter the food quality for shredders mediated by shifts in leaf-associated decomposer (i.e., aquatic fungi and bacteria) communities. However, little is known about the relative importance of the waterborne and dietary effect pathway. Therefore, we followed a tiered testing approach aimed at assessing the relative importance of these effect pathways. We employed the antibiotic ciprofloxacin (CIP) and the shredder Gammarus fossarum as model stressor and test species, respectively. In a first step, we assessed the short-term waterborne toxicity of CIP using survival and leaf consumption of G. fossarum as response variables. Alterations in the leaf-associated decomposer community, which may be reflected by their palatability, were assessed using food choice assays. Finally, we conducted a 2 × 2-factorial experiment over 24 days assessing the pathways individually and combined using energy processing (i.e., leaf consumption and feces production), growth and energy storage (i.e., neutral lipid fatty acids) as variables. Short term waterborne exposure indicated low toxicity with LC50 and EC50 values of 13.6 and 6.4 mg CIP/L, respectively. At the same time, shredders did not prefer any leaf material during the food choice assay. However, the fungal community was significantly affected in the highest CIP-treatments (0.5 and 2.5 mg/L) suggesting an altered food quality for shredders. This assumption is supported by the results of the long-term assay. At 0.5 mg CIP/L, gammarids' leaf consumption, growth and energy storage were increased when subjected via the dietary pathway, which was linked to changes in the leaf-associated microbial community. Our data highlight the importance of dietary effect pathways for effects on shredders, potentially impacting energy dynamics in detritus-based stream ecosystems.