University of St Andrews Research Portal
Not a member yet
81329 research outputs found
Sort by
Proofs with 'star' and 'perp'
No full textIn this paper, I show how to incorporate insights from the model-theoretic semantics for negation (insights due the late J. Michael Dunn among others), into a proof-first understanding of the semantics of negation. I then discuss the different ways a logical pluralist may understand the underlying accounts of proofs and their significance.</p
Coreset Spectral Clustering
No full textCoresets have become an invaluable tool for solving k-means and kernel k-means clustering problems on large datasets with small numbers of clusters. On the other hand, spectral clustering works well on sparse graphs and has recently been extended to scale efficiently to large numbers of clusters. We exploit the connection between kernel k-means and the normalised cut problem to combine the benefits of both. Our main result is a coreset spectral clustering algorithm for graphs that clusters a coreset graph to infer a good labelling of the original graph. We prove that an α-approximation for the normalised cut problem on the coreset graph is an O(α)- approximation on the original. We also improve the running time of the state-of-the-art coreset algorithm for kernel k-means on sparse kernels, from Õ(nk) to Õ(n · min{k, davg}), where davg is the average number of non-zero entries in each row of the n × n kernel matrix. Our experiments confirm our coreset algorithm is asymptotically faster on large real-world graphs with many clusters, and show that our clustering algorithm overcomes the main challenge faced by coreset kernel k-means on sparse kernels which is getting stuck in local optima
Elevated δ<sup>15</sup>N linked to inhibited nitrification coupled to ammonia volatilization in sediments of shallow alkaline-hypersaline lakes
No full textAlkaline lakes are among the most bioproductive aquatic ecosystems on Earth. The factors that ultimately limit productivity in these systems can vary, but nitrogen (N) cycling in particular has been shown to be adversely affected by high salinity, evidently due to the inhibition of nitrifying bacteria (i.e., those that convert ammonic species to nitrogen oxides). The coastal plain of Coorong National Park in South Australia, which hosts several alkaline lakes along 130 km of coastline, provides an ideal natural laboratory for examining how fine-scale differences in the geochemistry of such environments can lead to broad variations in nitrogen cycling through time, as manifest in sedimentary δ15N. Moreover, the lakes provide a gradient of aqueous conditions that allows us to assess the effects of pH, salinity, and carbonate chemistry on the sedimentary record. We report a wide range of δ15N values (3.8‰–18.6‰) measured in the sediments (0–35 cm depth) of five lakes of the Coorong region. Additional data include major element abundances, carbonate δ13C and δ18O values, and the results of principal component analyses. Stable nitrogen isotopes and wt% sodium (Na) display positive correlation (R2 = 0.59, p < 0.001) across all lake systems. Principal component analyses further support the notion that salinity has historically impacted nitrogen cycling. We propose that the inhibition of nitrification at elevated salinity may lead to the accumulation of ammonic species, which, when exposed to the water column, are prone to ammonia volatilization facilitated by intervals of elevated pH. This process is accompanied by a significant isotope fractionation effect, isotopically enriching the nitrogen that remains in the lake water. This nitrogen is eventually buried in the sediments, preserving a record of these combined processes. Analogous enrichments in the rock record may provide important constraints on past chemical conditions and their associated microbial ecologies. Specifically, ancient terrestrial aquatic systems with high δ15N values attributed to denitrification and thus oxygen deficiency may warrant re-evaluation within the framework of this alternative. Constraints on pH as provided by elevated δ15N via ammonia volatilization may also inform critical aspects of closed-basin paleoenvironments and their suitability for a de novo origin of life
Brief report of a pilot test of an audit tool for assurance of Declarations of Interest policy in NICE COVID-19 guidance production
No full textIntroduction: The National Institute for Health and Care Excellence (NICE) UK helps practitioners and commissioners to achieve high-quality, cost-effective patient care by publishing relevant evidence-based recommendations, including guidelines. These are developed by NICE advisory committees comprising health and care professionals, service users, and their carers. To ensure NICE's independence it has a comprehensive policy on declaring and managing interests for advisory committees. The application of the policy is not straightforward and the aim of this work was to pilot test an audit tool for external assurance of self-declared interests; and to produce methodology with wider applicability for other research and guideline-producing organisations. Methods: A pilot study of adherence to the NICE DOI policy was undertaken. The number and type of DOIs made by the 48 panel members developing the NG191 Managing COVID-19 guideline was assessed, using multiple external sources to check for undeclared interests. DOIs were reviewed from a 6-month period at the height of activity (1 March?31 August 2021). DOIs were checked by assessors (experienced NICE staff involved in application of the DOI policy) against a comprehensive search strategy including multiple public sources (PubMed, Google Scholar, Disclosure, etc.) of information for undisclosed DOIs over the 12 months preceding the member's contribution and the 12 months during which they contributed. Results: There was a total of 126 unique self-declared interests during the 6-month time period and we found a further 280 undeclared interests. Of these 280, 75 were deemed likely to be relevant (27%) but only 10 (4%), would have led to a request for more information from the panel member had it been declared (according to DOI policy version 1.4 in place at that time). The team considered that 7 of the undeclared interests would have led to exclusion from decision-making had they been known (3% of the undeclared interests). All related to academic output. Discussion: This pilot provided useful insight into NICE DOI policy compliance. The methodology needs further testing and development for wider, routine contexts. A digital version of the audit tool would enable routine, more efficient audit of declarations of interest. This study was an important quality assurance exercise from which we have been able to propose methodology that can be automated and widely adopted
Resistify: a novel NLR classifier that reveals helitron-associated NLR expansion in Solanaceae
No full textNucleotide-binding domain leucine-rich repeat (NLR) proteins are a key component of the plant innate immune system. In plant genomes, NLRs exhibit considerable presence/absence variation and sequence diversity. Recent advances in sequencing technologies have made the generation of high-quality novel plant genome assemblies considerably more straightforward. Accurately identifying NLRs from these genomes is a prerequisite for improving our understanding of NLRs and identifying novel sources of disease resistance. While several tools have been developed to predict NLRs, they are hampered by low accuracy, speed, and availability. Here, the NLR annotation tool Resistify is presented. Resistify is an easy-to-use, rapid, and accurate tool to identify and classify NLRs from protein sequences. Applying Resistify to the RefPlantNLR database demonstrates that it can correctly identify NLRs from a diverse range of species. Applying Resistify in combination with tools to identify transposable elements to a panel of Solanaceae genomes reveals a previously undescribed association between NLRs and Helitron transposable elements
Sigmatropic rearrangement enables access to a highly stable spirocyclic nitroxide for protein spin labelling
No full textSpin labelling enables the study of biomolecules using electron paramagnetic resonance (EPR) spectroscopy. Here, we describe the synthesis of a cysteine-reactive spin label based on a spirocyclic pyrrolidinyl nitroxide containing an iodoacetamide moiety. The spin label was shown to be highly persistent under reducing conditions while maintaining excellent EPR relaxation parameters up to a temperature of 180 K. After successful double spin labelling of a calmodulin variant, interspin distances were measured by the EPR experiment double electron–electron resonance (DEER) at 120 K
