Recognition and reconstruction of cell differentiation patterns with deep learning.

Cell lineage decisions occur in three-dimensional spatial patterns that are difficult to identify by eye. There is an ongoing effort to replicate such patterns using mathematical modeling. One approach uses long ranging cell-cell communication to replicate common spatial arrangements like checkerboard and engulfing patterns. In this model, the cell-cell communication has been implemented as a signal that disperses throughout the tissue. On the other hand, machine learning models have been developed for pattern recognition and pattern reconstruction tasks. We combined synthetic data generated by the mathematical model with spatial summary statistics and deep learning algorithms to recognize and reconstruct cell fate patterns in organoids of mouse embryonic stem cells. Application of Moran's index and pair correlation functions for in vitro and synthetic data from the model showed local clustering and radial segregation. To assess the patterns as a whole, a graph neural network was developed and trained on synthetic data from the model. Application to in vitro data predicted a low signal dispersion value. To test this result, we implemented a multilayer perceptron for the prediction of a given cell fate based on the fates of the neighboring cells. The results show a 70% accuracy of cell fate imputation based on the nine nearest neighbors of a cell. Overall, our approach combines deep learning with mathematical modeling to link cell fate patterns with potential underlying mechanisms.

Impact of the COVID-19 Lockdown Measures on Noise Levels in Urban Areas-A Pre/during Comparison of Long-Term Sound Pressure Measurements in the Ruhr Area, Germany.

BACKGROUND: A major source of noise pollution is traffic. In Germany, the SARS-CoV-2 lockdown caused a substantial decrease in mobility, possibly affecting noise levels. The aim is to analyze the effects of the lockdown measures on noise levels in the densely populated Ruhr Area. We focus on the analysis of noise levels before and during lockdown considering different land use types, weekdays, and time of day. METHODS: We used data from 22 automatic sound devices of the SALVE (Acoustic Quality and Health in Urban Environments) project, running since 2019 in Bochum, Germany. We performed a pre/during lockdown comparison of A-weighted equivalent continuous sound pressure levels. The study period includes five weeks before and five weeks during the SARS-CoV-2 induced administrative lockdown measures starting on 16 March 2020. We stratified our data by land use category (LUC), days of the week, and daytime. RESULTS: We observed highest noise levels pre-lockdown in the 'main street' and 'commercial areas' (68.4 ± 6.7 dB resp. 61.0 ± 8.0 dB), while in 'urban forests' they were lowest (50.9 ± 6.6 dB). A distinct mean overall noise reduction of 5.1 dB took place, with noise reductions occurring in each LUC. However, the magnitude of noise levels differed considerably between the categories. Weakest noise reductions were found in the 'main street' (3.9 dB), and strongest in the 'urban forest', 'green space', and 'residential area' (5.9 dB each). CONCLUSIONS: Our results are in line with studies from European cities. Strikingly, all studies report noise reductions of about 5 dB. Aiming at a transformation to a health-promoting urban mobility can be a promising approach to mitigating health risks of noise in cities. Overall, the experiences currently generated by the pandemic offer data for best practices and policies for the development of healthy urban transportation-the effects of a lower traffic and more tranquil world were experienced firsthand by people during this time.