Biology subjects

Holzinger, A.

Publications and source records attributed to Holzinger, A..

4 recordsLinked to original sources

bioRxiv2022

Interactive explainable AI platform for graph neural networks

BackgroundLack of trust in artificial intelligence (AI) models in medicine is still the key blockage for the use of AI in clinical decision support systems (CDSS). Although AI models are already performing excellently in systems medicine, their black-box nature entails that patient-specific decisions are incomprehensible for the physician. This is especially true for very complex models such as graph neural networks (GNNs), a common state-of-the-art approach to model biological networks such as protein-protein-interaction graphs (PPIs) to predict clinical outcomes. The aim of explainable AI (XAI) algorithms is to "explain" to a human domain expert, which input features, such as genes, influenced a specific recommendation. However, in the clinical domain, it is essential that these explanations lead to some degree of causal understanding by a clinician in the context of a specific application. ResultsWe developed the CLARUS platform, aiming to promote human understanding of GNN predictions by allowing the domain expert to validate and improve the decision-making process. CLARUS enables the visualisation of the patient-specific biological networks used to train and test the GNN model, where nodes and edges correspond to gene products and their interactions, for instance. XAI methods, such as GNNExplainer, compute relevance values for genes and interactions. The CLARUS graph visualisation highlights gene and interaction relevances by color intensity and line thickness, respectively. This enables domain experts to gain deeper insights into the biological network by identifying the most influential sub-graphs and molecular pathways crucial for the decision-making process. More importantly, the expert can interactively alter the patient-specific PPI network based on the acquired understanding and initiate re-prediction or retraining. This interactivity allows to ask manual counterfactual questions and analyse the resulting effects on the GNN prediction. ConclusionTo the best of our knowledge, we present the first interactive XAI platform prototype, CLARUS, that allows not only the evaluation of specific human counterfactual questions based on user-defined alterations of patient PPI networks and a re-prediction of the clinical outcome but also a retraining of the entire GNN after changing the underlying graph structures. The platform is currently hosted by the GWDG on https://rshiny.gwdg.de/apps/clarus/.

bioinformatics↗

bioRxiv2022

Effects of microplastic ingestion on hydrogen production and microbiomes in the gut of the terrestrial isopod Porcellio scaber

Microplastic (MP) pollution is an environmental burden. MP enters food webs via ingestion by macrofauna, including isopods (Porcellio scaber) in terrestrial ecosystems. However, MP-effects on the host and its gut microbiome are largely unknown. We tested the hypothesis that biodegradable (polylactic acid, PLA) and non-biodegradable (polyethylene terephthalate, PET; polystyrene, PS) MP have contrasting effects on P. scaber mediated by changes of the associated gut microbiome. Although the isopods avoided food containing PS, isopod fitness after eight-week MP-exposure was unaffected. Qualitative and quantitative 16S rRNA gene and 16S rRNA analyses of gut microbiomes indicated general MP effects, MP-type specific indicator taxa, and stimulation by PLA compared to MP-free controls. Isopods emitted hydrogen, and its production increased and decreased after PLA-food and PET- or PS-food ingestion, respectively, relative to controls as indicated by microsensor measurements. Gut pH was unaffected by MP. We identified the gut of P. scaber as significant mobile source of reductant for soil microbiomes likely due to Enterobacteriaceae related fermentation activities that were stimulated by lactate generated during PLA-degradation. The findings suggest negative effects of PET and PS on gut fermentation, modulation of isopod hydrogen emissions by MP pollution, and the potential of MP to affect terrestrial food webs.

microbiology↗

bioRxiv2022

Comparison of fitness effects in the earthworm Eisenia fetida after exposure to single or multiple anthropogenic pollutants

Terrestrial ecosystems are exposed to many anthropogenic pollutants. Non-target effects of pesticides and fertilizers have put agricultural intensification in the focus as a driver for biodiversity loss. However, other pollutants, such as heavy metals, particulate matter, or microplastic also enter the environment, e.g. via traffic and industrial activities in urban areas. As soil acts as a potential sink for such pollutants, soil invertebrates like earthworms may be particularly affected by them. Under natural conditions soil invertebrates will likely be confronted with combinations of pollutants simultaneously, which may result in stronger negative effects if pollutants act synergistically. Within this work we study how multiple pollutants affect the soil-dwelling, substrate feeding earthworm Eisenia fetida. We compared the effects of the single stressors, polystyrene microplastic fragments, polystyrene fibers, brake dust and soot, with the combined effect of these pollutants when applied as a mixture. Endpoints measured were survival, growth, reproductive fitness, and changes in three oxidative stress markers. We found that among single pollutant treatments, brake dust imposed the strongest negative effects on earthworms in all measured endpoints including increased mortality rates. Sub-lethal effects were found for all pollutants. Exposing earthworms to all four pollutants simultaneously led to effects on mortality and oxidative stress markers that were smaller than expected by the respective null models. These antagonistic effects are likely a result of the adsorption of toxic substances found in brake dust to the other pollutants. With this study we show that effects of combinations of pollutants cannot necessarily be predicted from their individual effects and that combined effects will likely depend on identity and concentration of the pollutants.

ecology↗

bioRxiv2022

GNN-SubNet: disease subnetwork detection with explainable Graph Neural Networks

The tremendous success of graphical neural networks (GNNs) has already had a major impact on systems biology research. For example, GNNs are currently used for drug target recognition in protein-drug interaction networks as well as cancer gene discovery and more. Important aspects whose practical relevance is often underestimated are comprehensibility, interpretability, and explainability. In this work, we present a graph-based deep learning framework for disease subnetwork detection via explainable GNNs. In our framework, each patient is represented by the topology of a protein-protein network (PPI), and the nodes are enriched by molecular multimodal data, such as gene expression and DNA methylation. Therefore, our novel modification of the GNNexplainer for model-wide explanations can detect potential disease subnetworks, which is of high practical relevance. The proposed methods are implemented in the GNN-SubNet Python program, which we have made freely available on our GitHub for the international research community (https://github.com/pievos101/GNN-SubNet).

bioinformatics↗