Genevention (Germany)

Ambitions towards carbon neutrality and fossil-free production processes lead to challenges for the Swedish electrical grid from increasing demands of industries and the transport sector as well as increasing and more volatile energy production from renewable sources. It is expected that the electricity demand in Sweden will double within the next ~20 years. At the same time, many assets in the electrical grids are reaching the end of their technical lifetime. Therefore, it is essential, that the necessary reinvestment measures are planned proactively with the future capacity requirements in mind.

The electricity grid of KWH Netz is facing the challenges of the German energy transition: combined with ageing assets in the low-voltage and medium-voltage level these challenges lead to increased peak-to-peak reliability values. For the future, an increase in technical risks is expected. Therefore, the target is to describe the structural specifics of KWH Netz's electricity grid and to disclose the specific challenges in supplying consumers and feeders. The combined results of a long-term infrastructure analysis and an electrical analysis deliver various evaluation opportunities: Vulnerabilities regarding reliability on medium-voltage level are revealed and for the low-voltage level, a grid overview and critical assets are discovered. The “Integrated electrical and long-term infrastructure analysis” for the medium- and low-voltage grid has resulted in well-founded and detailed findings. A sound basis for short-term measures and the development of “optimized” medium- and long-term measures (including optimization of the grid structure reflecting the dependencies between medium- and low-voltage level) has been evolved.

We present Scaden, a deep neural network for cell deconvolution that uses gene expression information to infer the cellular composition of tissues. Scaden is trained on single-cell RNA sequencing (RNA-seq) data to engineer discriminative features that confer robustness to bias and noise, making complex data preprocessing and feature selection unnecessary. We demonstrate that Scaden outperforms existing deconvolution algorithms in both precision and robustness. A single trained network reliably deconvolves bulk RNA-seq and microarray, human and mouse tissue expression data and leverages the combined information of multiple datasets. Because of this stability and flexibility, we surmise that deep learning will become an algorithmic mainstay for cell deconvolution of various data types. Scaden's software package and web application are easy to use on new as well as diverse existing expression datasets available in public resources, deepening the molecular and cellular understanding of developmental and disease processes.