From Keypoints to Object Landmarks via Self-Training Correspondence: A Novel Approach to Unsupervised Landmark Discovery.

This paper proposes a novel paradigm for the unsupervised learning of object landmark detectors. Contrary to existing methods that build on auxiliary tasks such as image generation or equivariance, we propose a self-training approach where, departing from generic keypoints, a landmark detector and descriptor is trained to improve itself, tuning the keypoints into distinctive landmarks. To this end, we propose an iterative algorithm that alternates between producing new pseudo-labels through feature clustering and learning distinctive features for each pseudo-class through contrastive learning. With a shared backbone for the landmark detector and descriptor, the keypoint locations progressively converge to stable landmarks, filtering those less stable. Compared to previous works, our approach can learn points that are more flexible in terms of capturing large viewpoint changes. We validate our method on a variety of difficult datasets, including LS3D, BBCPose, Human3.6M and PennAction, achieving new state of the art results. Code and models can be found at https://github.com/dimitrismallis/KeypointsToLandmarks/.

Lessons from past pandemics: a systematic review of evidence-based, cost-effective interventions to suppress COVID-19.

BACKGROUND: In an unparalleled global response, during the COVID-19 pandemic, 90 countries asked 3.9 billion people to stay home. Yet other countries avoided lockdowns and focused on other strategies, like contact tracing. How effective and cost-effective are these strategies? We aimed to provide a comprehensive summary of the evidence on past pandemic controls, with a focus on cost-effectiveness. METHODS: Following PRISMA guidelines, MEDLINE (1946 to April week 2, 2020) and EMBASE (1974 to April 17, 2020) were searched using a range of terms related to pandemic control. Articles reporting on the effectiveness or cost-effectiveness of at least one intervention were included. RESULTS: We found 1653 papers; 62 were included. The effectiveness of hand-washing and face masks was supported by randomized trials. These measures were highly cost-effective. For other interventions, only observational and modelling studies were found. They suggested that (1) the most cost-effective interventions are swift contact tracing and case isolation, surveillance networks, protective equipment for healthcare workers, and early vaccination (when available); (2) home quarantines and stockpiling antivirals are less cost-effective; (3) social distancing measures like workplace and school closures are effective but costly, making them the least cost-effective options; (4) combinations are more cost-effective than single interventions; and (5) interventions are more cost-effective when adopted early. For 2009 H1N1 influenza, contact tracing was estimated to be 4363 times more cost-effective than school closure ($2260 vs. $9,860,000 per death prevented). CONCLUSIONS AND CONTRIBUTIONS: For COVID-19, a cautious interpretation suggests that (1) workplace and school closures are effective but costly, especially when adopted late, and (2) scaling up as early as possible a combination of interventions that includes hand-washing, face masks, ample protective equipment for healthcare workers, and swift contact tracing and case isolation is likely to be the most cost-effective strategy.

Variability in Enteric Methane Emissions among Dairy Cows during Lactation.

The aim of this study was to investigate variability in enteric CH4 emission rate and emissions per unit of milk across lactations among dairy cows on commercial farms in the UK. A total of 105,701 CH4 spot measurements were obtained from 2206 mostly Holstein-Friesian cows on 18 dairy farms using robotic milking stations. Eleven farms fed a partial mixed ration (PMR) and 7 farms fed a PMR with grazing. Methane concentrations (ppm) were measured using an infrared CH4 analyser at 1s intervals in breath samples taken during milking. Signal processing was used to detect CH4 eructation peaks, with maximum peak amplitude being used to derive CH4 emission rate (g/min) during each milking. A multiple-experiment meta-analysis model was used to assess effects of farm, week of lactation, parity, diet, and dry matter intake (DMI) on average CH4 emissions (expressed in g/min and g/kg milk) per individual cow. Estimated mean enteric CH4 emissions across the 18 farms was 0.38 (s.e. 0.01) g/min, ranging from 0.2 to 0.6 g/min, and 25.6 (s.e. 0.5) g/kg milk, ranging from 15 to 42 g/kg milk. Estimated dry matter intake was positively correlated with emission rate, which was higher in grazing cows, and negatively correlated with emissions per kg milk and was most significant in PMR-fed cows. Mean CH4 emission rate increased over the first 9 weeks of lactation and then was steady until week 70. Older cows were associated with lower emissions per minute and per kg milk. Rank correlation for CH4 emissions among weeks of lactation was generally high. We conclude that CH4 emissions appear to change across and within lactations, but ranking of a herd remains consistent, which is useful for obtaining CH4 spot measurements.

Detection of Methane Eructation Peaks in Dairy Cows at a Robotic Milking Station Using Signal Processing.

The aim of this study was to investigate the use of signal processing to detect eructation peaks in CH4 released by cows during robotic milking, and to compare recordings from three gas analysers (Guardian SP and NG, and IRMAX) differing in volume of air sampled and response time. To allow comparison of gas analysers using the signal processing approach, CH4 in air (parts per million) was measured by each analyser at the same time and continuously every second from the feed bin of a robotic milking station. Peak analysis software was used to extract maximum CH4 amplitude (ppm) from the concentration signal during each milking. A total of 5512 CH4 spot measurements were recorded from 65 cows during three consecutive sampling periods. Data were analysed with a linear mixed model including analyser × period, parity, and days in milk as fixed effects, and cow ID as a random effect. In period one, air sampling volume and recorded CH4 concentration were the same for all analysers. In periods two and three, air sampling volume was increased for IRMAX, resulting in higher CH4 concentrations recorded by IRMAX and lower concentrations recorded by Guardian SP (p < 0.001), particularly in period three, but no change in average concentrations measured by Guardian NG across periods. Measurements by Guardian SP and IRMAX had the highest correlation; Guardian SP and NG produced similar repeatability and detected more variation among cows compared with IRMAX. The findings show that signal processing can provide a reliable and accurate means to detect CH4 eructations from animals when using different gas analysers.

The community structure and microbial linkage of rumen protozoa and methanogens in response to the addition of tea seed saponins in the diet of beef cattle.

BACKGROUND: This study investigated changes in rumen protozoal and methanogenic communities, along with the correlations among microbial taxa and methane (CH4) production of six Belmont Red Composite beef steers fed tea seed saponins (TSS). Animals were fed in three consecutive feeding periods, a high-grain basal diet for 14 d (BD period) then a period of progressive addition of TSS to the basal diet up to 30 g/d for 20 d (TSS period), followed by the basal diet for 13 d without TSS (BDP post-control period). RESULTS: The study found that TSS supplementation decreased the amount of the protozoal genus Entodinium and increased Polyplastron and Eudiplodinium genera. During BDP period, the protozoa community of steers did not return to the protozoal profiles observed in BD period, with higher proportions of Metadinium and Eudiplodinium and lower Isotricha. The addition of TSS was found to change the structure of methanogen community at the sub-genus level by decreasing the abundance of methanogens in the SGMT clade and increasing the abundance of methanogens in the RO clade. The correlation analysis indicated that the abundance of SGMT clade methanogens were positively correlated with Isotricha, and Isotricha genus and SGMT clade methanogens were positively correlated with CH4 production. While RO clade were positively correlated with the proportion of Metadinium genus, which was negatively correlated with CH4 emission. CONCLUSIONS: These results suggest that different genera of rumen protozoa ciliates appear to be selectively inhibited by TSS, and the change in methanogen community at the subgenus level may be due to the mutualistic relationships between methanogens and rumen ciliates.

Machine-Learning Techniques Can Enhance Dairy Cow Estrus Detection Using Location and Acceleration Data.

The aim of this study was to assess combining location, acceleration and machine learning technologies to detect estrus in dairy cows. Data were obtained from 12 cows, which were monitored continuously for 12 days. A neck mounted device collected 25,684 records for location and acceleration. Four machine-learning approaches were tested (K-nearest neighbor (KNN), back-propagation neural network (BPNN), linear discriminant analysis (LDA), and classification and regression tree (CART)) to automatically identify cows in estrus from estrus indicators determined by principal component analysis (PCA) of twelve behavioral metrics, which were: duration of standing, duration of lying, duration of walking, duration of feeding, duration of drinking, switching times between activity and lying, steps, displacement, average velocity, walking times, feeding times, and drinking times. The study showed that the neck tag had a static and dynamic positioning accuracy of 0.25 ± 0.06 m and 0.45 ± 0.15 m, respectively. In the 0.5-h, 1-h, and 1.5-h time windows, the machine learning approaches ranged from 73.3 to 99.4% for sensitivity, from 50 to 85.7% for specificity, from 77.8 to 95.8% for precision, from 55.6 to 93.7% for negative predictive value (NPV), from 72.7 to 95.4% for accuracy, and from 78.6 to 97.5% for F1 score. We found that the BPNN algorithm with 0.5-h time window was the best predictor of estrus in dairy cows. Based on these results, the integration of location, acceleration, and machine learning methods can improve dairy cow estrus detection.

Modelling the Effect of Diet Composition on Enteric Methane Emissions across Sheep, Beef Cattle and Dairy Cows.

Enteric methane (CH 4 ) is a by-product from fermentation of feed consumed by ruminants, which represents a nutritional loss and is also considered a contributor to climate change. The aim of this research was to use individual animal data from 17 published experiments that included sheep ( n = 288), beef cattle ( n = 71) and dairy cows ( n = 284) to develop an empirical model to describe enteric CH 4 emissions from both cattle and sheep, and then evaluate the model alongside equations from the literature. Data were obtained from studies in the United Kingdom (UK) and Australia, which measured enteric CH 4 emissions from individual animals in calorimeters. Animals were either fed solely forage or a mixed ration of forage with a compound feed. The feed intake of sheep was restricted to a maintenance amount of 875 g of DM per day (maintenance level), whereas beef cattle and dairy cows were fed to meet their metabolizable energy (ME) requirement (i.e., production level). A linear mixed model approach was used to develop a multiple linear regression model to predict an individual animal's CH 4 yield (g CH 4 /kg dry matter intake) from the composition of its diet. The diet components that had significant effects on CH 4 yield were digestible organic matter (DOMD), ether extract (EE) (both g/kg DM) and feeding level above maintenance intake: CH 4 (g/kg DM intake) = 0.046 (±0.001) × DOMD - 0.113 (±0.023) × EE - 2.47 (±0.29) × (feeding level - 1), with concordance correlation coefficient ( CCC ) = 0.655 and RMSPE = 14.0%. The predictive ability of the model developed was as reliable as other models assessed from the literature. These components can be used to predict effects of diet composition on enteric CH 4 yield from sheep, beef and dairy cattle from feed analysis information.

Competitive growth in a cooperative mammal.

In many animal societies where hierarchies govern access to reproduction, the social rank of individuals is related to their age and weight and slow-growing animals may lose their place in breeding queues to younger 'challengers' that grow faster. The threat of being displaced might be expected to favour the evolution of competitive growth strategies, where individuals increase their own rate of growth in response to increases in the growth of potential rivals. Although growth rates have been shown to vary in relation to changes in the social environment in several vertebrates including fish and mammals, it is not yet known whether individuals increase their growth rates in response to increases in the growth of particular reproductive rivals. Here we show that, in wild Kalahari meerkats (Suricata suricatta), subordinates of both sexes respond to experimentally induced increases in the growth of same-sex rivals by raising their own growth rate and food intake. In addition, when individuals acquire dominant status, they show a secondary period of accelerated growth whose magnitude increases if the difference between their own weight and that of the heaviest subordinate of the same sex in their group is small. Our results show that individuals adjust their growth to the size of their closest competitor and raise the possibility that similar plastic responses to the risk of competition may occur in other social mammals, including domestic animals and primates.

Protein- and DNA-based anthrax toxin vaccines confer protection in guinea pigs against inhalational challenge with Bacillus cereus G9241.

In the past decade, several Bacillus cereus strains have been isolated from otherwise healthy individuals who succumbed to bacterial pneumonia presenting symptoms resembling inhalational anthrax. One strain was indistinguishable from B. cereus G9241, previously cultured from an individual who survived a similar pneumonia-like illness and which was shown to possess a complete set of plasmid-borne anthrax toxin-encoding homologs. The finding that B. cereus G9241 pathogenesis in mice is dependent on pagA1-derived protective antigen (PA) synthesis suggests that an anthrax toxin-based vaccine may be effective against this toxin-encoding B. cereus strain. Dunkin Hartley guinea pigs were immunized with protein- and DNA-based anthrax toxin-based vaccines, immune responses were evaluated and survival rates were calculated after lethal aerosol exposure with B. cereus G9241 spores. Each vaccine induced seroconversion with the protein immunization regimen eliciting significantly higher serum levels of antigen-specific antibodies at the prechallenge time-point compared with the DNA-protein prime-boost immunization schedule. Complete protection against lethal challenge was observed in all groups with a detectable prechallenge serum titer of toxin neutralizing antibodies. For the first time, we demonstrated that the efficacy of fully defined anthrax toxin-based vaccines was protective against lethal B. cereus G9241 aerosol challenge in the guinea pig animal model.

Evolution of electroporated DNA vaccines.

Vaccines have evolved for hundreds of years, but all utilize the premise that safely pre-exposing the host to some component of a pathogen allows for enhanced immune recognition, and potential protection from disease, upon encountering the pathogen at a later date. Early vaccination strategies used inactivated or attenuated vaccines, many of which contained toxins and other components that resulted in reactogenicity or risk of reversion to virulence. DNA vaccines supplant many of the issues associated with inactivated or attenuated vaccines, but these vaccines tend to provide weak immunological responses, particularly in primates. DNA Electroporation may prove to be the "missing link" in the evolution of DNA vaccines allowing for enhanced immune responses from DNA vaccination in humans thereby resulting in protection from disease post-pathogen exposure.

DNA electroporation of multi-agent vaccines conferring protection against select agent challenge: TriGrid delivery system.

Effective multi-agent/multivalent vaccines that confer protection against more than one disease are highly desirable to the patient and to health-care professionals. Electroporation of DNA vaccines, whereby tissues injected with DNA are subjected to localized electrical currents, is an ideal platform technology that achieves protective immune responses to multivalent vaccination. Here, we describe an electroporation-based immunization technique capable of administering a cocktail of DNA vaccinations in vivo. Immune response measurements, including protection from pathogen challenge and induction of antigen-specific antibody responses and cell-mediated immune responses, are also discussed.

Targeting interleukin-13 with tralokinumab attenuates lung fibrosis and epithelial damage in a humanized SCID idiopathic pulmonary fibrosis model.

The aberrant fibrotic and repair responses in the lung are major hallmarks of idiopathic pulmonary fibrosis (IPF). Numerous antifibrotic strategies have been used in the clinic with limited success, raising the possibility that an effective therapeutic strategy in this disease must inhibit fibrosis and promote appropriate lung repair mechanisms. IL-13 represents an attractive target in IPF, but its disease association and mechanism of action remains unknown. In the present study, an overexpression of IL-13 and IL-13 pathway markers was associated with IPF, particularly a rapidly progressive form of this disease. Targeting IL-13 in a humanized experimental model of pulmonary fibrosis using tralokinumab (CAT354) was found to therapeutically block aberrant lung remodeling in this model. However, targeting IL-13 was also found to promote lung repair and to restore epithelial integrity. Thus, targeting IL-13 inhibits fibrotic processes and enhances repair processes in the lung.

Electroporation of a multivalent DNA vaccine cocktail elicits a protective immune response against anthrax and plague.

Electroporation of DNA vaccines represents a platform technology well positioned for the development of multivalent biodefense vaccines. To evaluate this hypothesis, three vaccine constructs were produced using codon-optimized genes encoding Bacillus anthracis Protective Antigen (PA), and the Yersinia pestis genes LcrV and F1, cloned into pVAX1. A/J mice were immunized on a prime-boost schedule with these constructs using the electroporation-based TriGrid Delivery System. Immunization with the individual pDNA vaccines elicited higher levels of antigen-specific IgG than when used in combination. DNA vaccine effectiveness was proven, the pVAX-PA titers were toxin neutralizing and fully protective against a lethal B. anthracis spore challenge when administered alone or co-formulated with the plague pDNA vaccines. LcrV and F1 pVAX vaccines against plague were synergistic, resulting in 100% survival, but less protective individually and when co-formulated with pVAX-PA. These DNA vaccine responses were Th1/Th2 balanced with high levels of IFN-γ and IL-4 in splenocyte recall assays, contrary to complimentary protein Alum vaccinations displaying a Th2 bias with increased IL-4 and low levels of IFN-γ. These results demonstrate the feasibility of electroporation to deliver and maintain the overall efficacy of an anthrax-plague DNA vaccine cocktail whose individual components have qualitative immunological differences when combined.

Luminescence of colloidal CdSe/ZnS nanoparticles: high sensitivity to solvent phase transitions.

We investigate nanosecond photoluminescence processes in colloidal core/shell CdSe/ZnS nanoparticles dissolved in water and found strong sensitivity of luminescence to the solvent state. Several pronounced changes have been observed in the narrow temperature interval near the water melting point. First of all, the luminescence intensity substantially (approximately 50%) increases near the transition. In a large temperature scale, the energy peak of the photoluminescence decreases with temperature due to temperature dependence of the energy gap. Near the melting point, the peak shows N-type dependence with the maximal changes of approximately 30 meV. The line width increases with temperature and also shows N-type dependence near the melting point. The observed effects are associated with the reconstruction of ligands near the ice/water phase transition.

Endothelium-dependent vasodilation in human mesenteric artery is primarily mediated by myoendothelial gap junctions intermediate conductance calcium-activated K+ channel and nitric oxide.

Myoendothelial microdomain signaling via localized calcium-activated potassium channel (K(Ca)) and gap junction connexins (Cx) is critical for endothelium-dependent vasodilation in rat mesenteric artery. The present study determines the relative contribution of NO and gap junction-K(Ca) mediated microdomain signaling to endothelium-dependent vasodilation in human mesenteric artery. The hypothesis tested was that such activity is due to NO and localized K(Ca) and Cx activity. In mesenteric arteries from intestinal surgery patients, endothelium-dependent vasodilation was characterized using pressure myography with pharmacological intervention. Vessel morphology was examined using immunohistochemical and ultrastructural techniques. In vessel segments at 80 mm Hg, the intermediate (I)K(Ca) blocker 1-[(2-chlorophenyl)diphenyl-methyl]-1H-pyrazole (TRAM-34; 1 µM) inhibited bradykinin (0.1 nM-3 µM)-induced vasodilation, whereas the small (S) K(Ca) blocker apamin (50 and 100 nM) had no effect. Direct IK(Ca) activation with 1-ethyl-2-benzimidazolinone (1-EBIO; 10-300 µM) induced vasodilation, whereas cyclohexyl-[2-(3,5-dimethyl-pyrazol-1-yl)-6-methyl-pyrimidin-4-yl]-amine (1-30 µM), the SK(Ca) activator, failed to dilate arteries, whereas dilation induced by 1-EBIO (10-100 µM) was blocked by TRAM-34. Bradykinin-mediated vasodilation was attenuated by putative gap junction block with carbenoxolone (100 µM), with remaining dilation blocked by N-nitro l-arginine methyl ester (100 µM) and [1H-[1,2,4]oxadiazolo-[4,3-a]quinoxalin-1-one] (10 µM), NO synthase and soluble guanylate cyclase blockers, respectively. In human mesenteric artery, myoendothelial gap junction and IK(Ca) activity are consistent with Cx37 and IK(Ca) microdomain expression and distribution. Data suggest that endothelium-dependent vasodilation is primarily mediated by NO, IK(Ca), and gap junction Cx37 in this vessel. Myoendothelial microdomain signaling sites are present in human mesenteric artery and are likely to contribute to endothelium-dependent vasodilation via a mechanism that is conserved between species.

Functional and molecular characterization of metabotropic glutamate receptors expressed in rat striatal cholinergic interneurones.

In the present study we have used single-cell RT-PCR in conjunction with electrophysiology to examine the expression and functional properties of metabotropic glutamate receptors (mGluRs) expressed within biochemically identified cholinergic interneurones in the rat striatum. Using single-cell RT-PCR, it was possible to demonstrate the presence of mGluR1, mGluR2, mGluR3, mGluR5 and mGluR7 mRNAs within single cholinergic interneurones. Bath application of the non-selective mGluR agonist (1 S,3R)-1-aminocyclopentane-1,3-dicarboxylic acid (1 S,3R-ACPD) or the group-I mGluR agonist 3,5-dihydroxyphenylglycine (DHPG) depolarized all cholinergic neurones tested by activation of an inward current at -60 mV. The effects of DHPG were partially inhibited by the mGluR5 selective antagonist 6-methyl-2-(pherazo)-3-pyridinol and by the non-selective group-I antagonist alpha-methyl-4-carboxyphenylglycine but were not mimicked by the group-II and group-III selective mGluR agonists 2-(2,3-dicarboxycyclopropyl)glycine (DCG-IV) and L-2-amino-4-phosphonobutanoate (L-AP4), respectively. Intrastriatal stimulation evoked an excitatory postsynaptic current within cholinergic neurones that was reversibly inhibited by bath application of the group-II and group-III selective mGluR agonists DCG-IV and L-AP4, respectively, via presynaptic actions. In summary, we have identified the mGluRs expressed by striatal cholinergic interneurones and demonstrated that their activation produces modulatory effects via both pre- and postsynaptic mechanisms.