An artificial neural network and Random Forest identify glyphosate-impacted brackish communities based on 16S rRNA amplicon MiSeq read counts
Marine pollution bulletin
Machine learning algorithms can be trained on complex data sets to detect, predict, or model specific aspects. Aim of this study was to train an artificial neural network in comparison to a Random Forest model to detect induced changes in microbial communities, in order to support environmental monitoring efforts of contamination events. Models were trained on taxon count tables obtained via next-generation amplicon sequencing of water column samples originating from a lab microcosm incubation experiment conducted over 140 days to determine the effects of glyphosate on succession within brackish-water microbial communities. Glyphosate-treated assemblages were classified correctly; a subsetting approach identified the taxa primarily responsible for this, permitting the reduction of input features. This study demonstrates the potential of artificial neural networks to predict indicator species for glyphosate contamination. The results could empower the development of environmental monitoring strategies with applications limited to neither glyphosate nor amplicon sequence data.
Using Direct-touch Interaction for the Visual Exploration of Profiling Sensor Data
OCEANS 2015 MTS/IEEE Washington
MTS/IEEE Oceans Conference and Exhibition (OCEANS) <2015, Genova, Italy>
We present a novel software approach for the interactive and integrated visual exploration of sensor data acquired by profiling platforms. The solution combines decent web-technologies and linked interactive 2D and 3D data presentations utilizing a direct-touch interaction metaphor. The user can point with his finger on specific data points and slide through the depth profiles and the time scale. For the visualization and interaction design, widely recognized visualization and interaction principles and techniques like "focus & context" and "overview & detail" were incorporated. Our goal is to facilitate the dynamic exploration of spatio-temporal sensor data, to reduce visual clutter in the user interface and to allow the user to focus in on specific time and depth situations. We present first results of a user study with domain experts (marine researchers). The goal of the study was to prove the design considerations and to investigate the practicability of web-based interactive visualization and direct-touch interaction for analytical tasks. Additionally we tested the suitability of our approach for data exploration tasks on different multi-touch display setups. The study results show, that through the intuitive direct-touch interaction and the linked 2D and 3D visualizations of the data sets, the proposed solution offers a better support to explore profiling data sets, compared to currently available nonintegrated tool sets. The presented approach also reduces the efforts of analytical tasks. While the proposed 3D view of ADCP data was considered very useful for a fast assessment of current conditions in the water column, most of the participants also had problems to fully understand the offered navigation and interaction within the 3D space. Creating an intuitive view navigation in 3D space with direct-touch interaction proved to be rather challenging and needs a careful design.