Predicting the in-game status in soccer with machine learning using spatiotemporal player tracking data.

An important structuring feature of a soccer match is the in-game status, whether a match is interrupted or in play. This is necessary to calculate performance indicators relative to the effective playing time or to find standard situations, ball actions, and other tactical structures in spatiotemporal data. Our study explores the extent to which the in-game status can be determined using time-continuous player positions. Therefore, to determine the in-game status we tested four established machine learning methods: logistic regression, decision trees, random forests, and AdaBoost. The models were trained and evaluated using spatiotemporal data and manually annotated in-game status of 102 matches in the German Bundesliga. Results show up to 92% accuracy in predicting the in-game status in previously unknown matches on frame level. The best performing method, AdaBoost, shows 81% precision for detecting stoppages (longer than 2 s). The absolute time shift error at the start was ≤ 2 s for 77% and 81% at the end for all correctly predicted stoppages. The mean error of the in-game total distance covered per player per match using the AdaBoost in-game status prediction was - 102 ± 273 m, which is 1.3% of the mean value of this performance indicator (7939 m). Conclusively, the prediction quality of our model is high enough to provide merit for performance diagnostics when teams have access to player positions (e.g., from GPS/LPM systems) but no human-annotated in-game status and/or ball position data, such as in amateur or youth soccer.

The GEF Trio controls endothelial cell size and arterial remodeling downstream of Vegf signaling in both zebrafish and cell models.

Arterial networks enlarge in response to increase in tissue metabolism to facilitate flow and nutrient delivery. Typically, the transition of a growing artery with a small diameter into a large caliber artery with a sizeable diameter occurs upon the blood flow driven change in number and shape of endothelial cells lining the arterial lumen. Here, using zebrafish embryos and endothelial cell models, we describe an alternative, flow independent model, involving enlargement of arterial endothelial cells, which results in the formation of large diameter arteries. Endothelial enlargement requires the GEF1 domain of the guanine nucleotide exchange factor Trio and activation of Rho-GTPases Rac1 and RhoG in the cell periphery, inducing F-actin cytoskeleton remodeling, myosin based tension at junction regions and focal adhesions. Activation of Trio in developing arteries in vivo involves precise titration of the Vegf signaling strength in the arterial wall, which is controlled by the soluble Vegf receptor Flt1.

Neuronal sFlt1 and Vegfaa determine venous sprouting and spinal cord vascularization.

Formation of organ-specific vasculatures requires cross-talk between developing tissue and specialized endothelial cells. Here we show how developing zebrafish spinal cord neurons coordinate vessel growth through balancing of neuron-derived Vegfaa, with neuronal sFlt1 restricting Vegfaa-Kdrl mediated angiogenesis at the neurovascular interface. Neuron-specific loss of flt1 or increased neuronal vegfaa expression promotes angiogenesis and peri-neural tube vascular network formation. Combining loss of neuronal flt1 with gain of vegfaa promotes sprout invasion into the neural tube. On loss of neuronal flt1, ectopic sprouts emanate from veins involving special angiogenic cell behaviours including nuclear positioning and a molecular signature distinct from primary arterial or secondary venous sprouting. Manipulation of arteriovenous identity or Notch signalling established that ectopic sprouting in flt1 mutants requires venous endothelium. Conceptually, our data suggest that spinal cord vascularization proceeds from veins involving two-tiered regulation of neuronal sFlt1 and Vegfaa via a novel sprouting mode.