Developing large-scale datasets with research-quality annotations is challenging due to the high cost of refining clinically generated markup into high precision annotations. We evaluated the direct use of a large dataset with only clinically generated annotations in development of high-performance segmentation models for small research-quality challenge datasets.

We used a large retrospective dataset from our institution comprised of 1,620 clinically generated segmentations, and two challenge datasets (PROMISE12 50 patients, ProstateX-2 99 patients). We trained a 3D U-Net convolutional neural network (CNN) segmentation model using our entire dataset, and used that model as a template to train models on the challenge datasets. We also trained versions of the template model using ablated proportions of our dataset, and evaluated the relative benefit of those templates for the final models. Finally, we trained a version of the template model using an out-of-domain brain cancer dataset, and evaluated the relned a state-of-the-art model using unrefined clinical prostate annotations and found that its use as a template model significantly improved performance in other prostate segmentation tasks, even when trained with only 5% of the original dataset.The triangle structure, being a fundamental and significant element, underlies many theories and techniques in studying complex networks. The formation of triangles is typically measured by the clustering coefficient, in which the focal node is the centre-node in an open triad. In contrast, the recently proposed closure coefficient measures triangle formation from an end-node perspective and has been proven to be a useful feature in network analysis. Here, we extend it by proposing the directed closure coefficient that measures the formation of directed triangles. By distinguishing the direction of the closing edge in building triangles, we further introduce the source closure coefficient and the target closure coefficient. Then, by categorising particular types of directed triangles (e.g., head-of-path), we propose four closure patterns. Through multiple experiments on 24 directed networks from six domains, we demonstrate that at network-level, the four closure patterns are distinctive features in classifying network types, while at node-level, adding the source and target closure coefficients leads to significant improvement in link prediction task in most types of directed networks.The objective of this research was to comparatively investigate the effect of hot air drying (HA) and hybrid microwave-infrared drying (MI) on physico-chemical characteristics of Thai fermented fish viscera, Tai-Pla, curry powder (TCP). HA was carried out at 60°C, 70°C, and 80°C and MI was carried out at a microwave power of 740, 780, and 810 W with a constant infrared heating power (500 W) for different drying times to obtain the final moisture content ≤ 12.0% and the water activity (aw) ≤ 0.6. The quality characteristics of TCP were governed by HA temperature and MI output power. TCP dried using HA and MI at all conditions had similar contents of protein, lipid, ash, fiber, and carbohydrate (p>0.05). The fastest drying rate was detected when MI at 810 W for 40 min was applied (p0.05). RK-33 mouse Similar Fourier transform infrared (FTIR) spectra with different peak intensities were observed in all samples, indicating the same functional groups with different contents were found. The bulk density of all TCP ranged from 0.51 g/mL to 0.61 g/mL and the wettability ranged from 24.02% to 26.70%. MI at 810 W for 40 min effectively reduced the drying time (5-fold faster) and lowered the specific energy consumption (18-fold lower) compared to the HA at 60°C for 210 min. Therefore, MI is a promising drying technique to reduce the drying time and improve the overall quality of TCP.The APOBEC3 (A3) genes encode cytidine deaminase proteins with potent antiviral and anti-retroelement activity. This locus is characterized by duplication, recombination, and deletion events that gave rise to the seven A3s found in primates. These include three single deaminase domain A3s (A3A, A3C, and A3H) and four double deaminase domain A3s (A3B, A3D, A3F, and A3G). The most potent of the A3 proteins against HIV-1 is A3G. However, it is not clear if double deaminase domain A3s have a generalized functional advantage to restrict HIV-1. In order to test whether superior restriction factors could be created by genetically linking single A3 domains into synthetic double domains, we linked A3C and A3H single domains in novel combinations. We found that A3C/A3H double domains acquired enhanced antiviral activity that is at least as potent, if not better than, A3G. Although these synthetic double domain A3s package into budding virions more efficiently than their respective single domains, this does not fully explain their gain of antiviral potency. The antiviral activity is conferred both by cytidine-deaminase dependent and independent mechanisms, with the latter correlating to an increase in RNA binding affinity. T cell lines expressing this A3C-A3H super restriction factor are able to control replicating HIV-1ΔVif infection to similar levels as A3G. Together, these data show that novel combinations of A3 domains are capable of gaining potent antiviral activity to levels similar to the most potent genome-encoded A3s, via a primarily non-catalytic mechanism.[This corrects the article DOI 10.1371/journal.pone.0250078.].Small airways are difficult to access. Exhaled droplets, also referred to as particles, provide a sample of small airway lining fluid and may reflect inflammatory responses. We aimed to explore the effect of smoking on the composition and number of exhaled particles in a smoker-enriched study population. We collected and chemically analyzed exhaled particles from 102 subjects (29 never smokers, 36 former smokers and 37 current smokers) aged 39 to 83 years (median 63). A breathing maneuver maximized the number exhaled particles, which were quantified with a particle counter. The contents of surfactant protein A and albumin in exhaled particles was quantified with immunoassays and the contents of the phospholipids dipalmitoyl- and palmitoyl-oleoyl- phosphatidylcholine with mass spectrometry. Subjects also performed spirometry and nitrogen single breath washout. Associations between smoking status and the distribution of contents in exhaled particles and particle number concentration were tested with quantile regression, after adjusting for potential confounders.