Gaining Confidence in Software Inspection Using a Bayesian Belief Model

In this paper I describe how confidence in a software inspection can be obtained through calculating an estimate of its effectiveness. The method uses a Bayesian Belief Network to model the software inspection process and calculates the inference on how effective a particular inspection was. This technique was selected as it provides a means of initialising the model with inspectors' experience and has the ability to learn and optimise performance. This technique provides answers to some of the questions and limitations raised by current models used to predict inspection effectiveness. The application of the model to a major software project is discussed, covering the initial practitioner survey, model initialisation, model calibration and verification results obtained.     

A Method for Conceptualising Legal Domains. An Example from the Dutch Unemployment Benefits Act

There has been much talk of the need to build intermediate models of the expertise required preparatory to constructing a knowledge-based system in the legal domain. Such models offer advantages for verification, validation, maintenance and reuse. As yet, however, few such models have been reported at a useful level of detail. In this paper we describe a method for conceptualising legal domains as well as its application to a substantial fragment of the Dutch Unemployment Benefits Act (DUBA).We first discuss the intermediate models (called expertise models), then present a three-stage method for their construction, drawing on the CommonKADS work in knowledge acquisition, conceptual models of statute law, and the KANT method of knowledge analysis. Subsequently, we describe how these techniques were applied to the DUBA, and provide detailed examples of the resulting model. Finally, conclusions on the framework and guidelines are given as well as means of recording and presenting the various design choices.     

Calibration of AFM cantilever spring constants

In this paper we present two simple, reliable and readily applicable methods for calibrating cantilevers and measuring the thickness of thin gold films. The spring constant calibration requires knowledge of the Young's modulus, density of the cantilever and resonant frequency. The thickness of thin gold layers was determined by measuring changes in the resonant frequency and Q-factor of beam shaped AFM cantilevers before and after coating.The techniques for measuring the spring constant and thin film thickness provide accuracy on the order of 10-15%.     

Isolation and characterization of the microbial population of different South African kefir grains

Kefir, a slightly acidic fermented milk, is produced by adding lactic acid bacteria and yeasts, in the form of grains, to milk. The bacteria and yeasts present in the kefir grains are known to vary widely. Selective growth media and morphological and biochemical characteristics were used for the isolation and identification of the microbes present in the grains from eight different sources in South Africa. The kefir grains were activated in milk for only 24 h to prevent any changes in the microbial population of the grains. The microbial numbers varied between 6.4 × 10⁴ and 8.5 × 10 ⁸  cfu/g on the media selective for the bacterial species and between 1.5 × 10 ⁵ and 3.7 × 10 ⁸   cfu/g on the media selective for the yeast species. The bacterial genera that were identified included Lactobacillus , Leuconostoc and Lactococcus and the yeast genera included Zygosaccharomyces , Candida and Saccharomyces . The distribution frequencies of the microbes in the different grains were determined and most of the grains were dominated by two microbial species. No pediococci, acetic acid bacteria or propionibacteria were detected.     

Echium oil is better than rapeseed oil in improving the response of barramundi to a disease challenge

Pathogen infection stimulates the fatty acid (FA) metabolism and the production of pro-inflammatory derivatives of FA. Barramundi, Lates calcarifer, was fed on a diet rich in preformed long-chain (?C₂₀) polyunsaturated fatty acids (LC-PUFA) from fish oil (FO), to compare with diets containing high levels of C₁₈ precursors for LC-PUFA – stearidonic (SDA) and γ-linolenic acid (GLA) – from Echium plantagineum (EO), or rapeseed oil (RO) rich in α-linolenic acid (ALA), but a poor source of LC-PUFA and their precursors. After 6 weeks, when growth rates were similar amongst the dietary treatments, a sub-lethal dose of Streptococcus iniae was administered to half of the fish, while the other half were maintained unchallenged and were pair-fed with the infected fish. Under a disease challenge situation, the tissue FA depots depleted at 3 days post-infection (DPI) and were then restored to their previous concentrations at 7 DPI. During the infection period, EO fish had a higher content of n3 and n6 PUFA in their tissues, higher n3:n6 PUFA ratio and reduced levels of the eicosanoids, TXB₂ and 6-keto-PGF_1α, in their plasma compared with RO fish. Fish fed on FO and EO had a longer lasting and enduring response in their FA and eicosanoid concentrations, following a week of bacterial infection, compared with those fed on RO. EO, containing SDA and GLA and with a comparatively higher n3:n6 PUFA ratio, proved more effective than RO in compensating for immunity stress.     

Col4a3-/- Mice on Balb/C Background Have Less Severe Cardiorespiratory Phenotype and SGLT2 Over-Expression Compared to 129x1/SvJ and C57Bl/6 Backgrounds

Alport syndrome (AS) is a hereditary renal disorder with no etiological therapy. In the preclinical Col4a3^-/- model of AS, disease progression and severity vary depending on mouse strain. The sodium-glucose cotransporter 2 (SGLT2) is emerging as an attractive therapeutic target in cardiac/renal pathologies, but its application to AS remains untested. This study investigates cardiorespiratory function and SGLT2 renal expression in Col4a3^-/- mice from three different genetic backgrounds, 129x1/SvJ, C57Bl/6 and Balb/C. male Col4a3^-/- 129x1/SvJ mice displayed alterations consistent with heart failure with preserved ejection fraction (HFpEF). Female, but not male, C57Bl/6 and Balb/C Col4a3^-/- mice exhibited mild changes in systolic and diastolic function of the heart by echocardiography. Male C57Bl/6 Col4a3^-/- mice presented systolic dysfunction by invasive hemodynamic analysis. All strains except Balb/C males demonstrated alterations in respiratory function. SGLT2 expression was significantly increased in AS compared to WT mice from all strains. However, cardiorespiratory abnormalities and SGLT2 over-expression were significantly less in AS Balb/C mice compared to the other two strains. Systolic blood pressure was significantly elevated only in mutant 129x1/SvJ mice. The results provide further evidence for strain-dependent cardiorespiratory and hypertensive phenotype variations in mouse AS models, corroborated by renal SGLT2 expression, and support ongoing initiatives to develop SGLT2 inhibitors for the treatment of AS.     

Consistent flavor naming predicts recognition memory in children and young adults

The relationship between flavor naming and recognition memory was evaluated in groups of children and young adults who either self-generated flavor names or chose flavor names from a list of four alternatives. The adults were more successful at naming and remembering the flavors, with the age differences in naming performance tending to be larger than differences in memory performance. Differences in recognition memory among the age groups were modest, ranging from 3% to 12%. Providing flavor labels improved naming accuracy, but not naming consistency or memory for younger children (4-7 years old). Labels improved naming accuracy and consistency for the older children (8-11 years old), and showed a trend toward improving memory. Labels improved naming accuracy, consistency and recognition memory in adults. Consistent naming was a powerful predictor of recognition memory for each age group and test condition. Consistently named flavors were remembered correctly on 88% of the trials while inconsistently named flavors showed no evidence of memory. It was concluded that flavor recognition memory and naming consistency were both rooted in successfully matching flavor percepts to a multi-dimensional flavor representation that includes semantic information.     

Visualising Errors in Animal Pedigree Genotype Data

Genetic analysis of a breeding animal population involves determining the inheritance pattern of genotypes for multiple genetic markers across the individuals in the population pedigree structure. However, experimental pedigree genotype data invariably contains errors in both the pedigree structure and in the associated individual genotypes, introducing inconsistencies into the dataset, rendering them useless for further analysis. The resolution of these errors requires consideration of genotype inheritance patterns in the context of the pedigree structure. Existing pedigree visualisations are typically more suited to human pedigrees and are less suitable for large complex animal pedigrees which may exhibit cross generational inbreeding. Similarly, table‐based viewers of genotype marker data can highlight where errors become apparent but lack the functionality and interactive visual feedback to allow users to locate the origin of errors within the pedigree. In this paper, we detail a design study steered by biologists who work with pedigree data, and describe successive iterations through approaches and prototypes for viewing genotyping errors in the context of a displayed pedigree. We describe how each approach performs with real pedigree genotype data and why eventually we deemed them unsuitable. Finally, a novel prototype visualisation for pedigrees, which we term the ‘sandwich view’, is detailed and we demonstrate how the approach effectively communicates errors in the pedigree context, supporting the biologist in the error identification task.