Approximation of Bayesian predictive p-values with regression ABC

In the Bayesian framework a standard approach to model criticism is to compare some function of the observed data to a reference predictive distribution. The result of the comparison can be summarized in the form of a p-value, and it's well known that computation of some kinds of Bayesian predictive p-values can be challenging. The use of regression adjustment approximate Bayesian computation (ABC) methods is explored for this task. Two problems are considered. The first is the calibration of posterior predictive p-values so that they are uniformly distributed under some reference distribution for the data. Computation is difficult because the calibration process requires repeated approximation of the posterior for different data sets under the reference distribution. The second problem considered is approximation of distributions of prior predictive p-values for the purpose of choosing weakly informative priors in the case where the model checking statistic is expensive to compute. Here the computation is difficult because of the need to repeatedly sample from a prior predictive distribution for different values of a prior hyperparameter. In both these problems we argue that high accuracy in the computations is not required, which makes fast approximations such as regression adjustment ABC very useful. We illustrate our methods with several samples.

Bayesian P-splines with a multiplicative term in EMG trace data

A method is proposed to describe force or compound muscle action potential (CMAP) trace data collected in an electromyography study for motor unit number estimation (MUNE). Experimental data was collected using incre- mental stimulation at multiple durations. However, stimulus information, vital for alternate MUNE methods, is not comparable for multiple duration data and therefore previous methods of MUNE (Ridall et al., 2006, 2007) cannot be used with any reliability. Hypothesised ring combinations of motor units are mod- elled using a multiplicative factor and Bayesian P-spline formulation. The model describes the process for force and CMAP in a meaningful way.

Bayesian versus frequentist statistical inference for investi gating a one-off cancer cluster reported to a health department

Background The problem of silent multiple comparisons is one of the most difficult statistical problems faced by scientists. It is a particular problem for investigating a one-off cancer cluster reported to a health department because any one of hundreds, or possibly thousands, of neighbourhoods, schools, or workplaces could have reported a cluster, which could have been for any one of several types of cancer or any one of several time periods. Methods This paper contrasts the frequentist approach with a Bayesian approach for dealing with silent multiple comparisons in the context of a one-off cluster reported to a health department. Two published cluster investigations were re-analysed using the Dunn-Sidak method to adjust frequentist p-values and confidence intervals for silent multiple comparisons. Bayesian methods were based on the Gamma distribution. Results Bayesian analysis with non-informative priors produced results similar to the frequentist analysis, and suggested that both clusters represented a statistical excess. In the frequentist framework, the statistical significance of both clusters was extremely sensitive to the number of silent multiple comparisons, which can only ever be a subjective "guesstimate". The Bayesian approach is also subjective: whether there is an apparent statistical excess depends on the specified prior. Conclusion In cluster investigations, the frequentist approach is just as subjective as the Bayesian approach, but the Bayesian approach is less ambitious in that it treats the analysis as a synthesis of data and personal judgements (possibly poor ones), rather than objective reality. Bayesian analysis is (arguably) a useful tool to support complicated decision-making, because it makes the uncertainty associated with silent multiple comparisons explicit.

Forecasting crashes at the planning level: Simultaneous negative binomial crash model applied in Tucson, Arizona

At least two important transportation planning activities rely on planning-level crash prediction models. One is motivated by the Transportation Equity Act for the 21st Century, which requires departments of transportation and metropolitan planning organizations to consider safety explicitly in the transportation planning process. The second could arise from a need for state agencies to establish incentive programs to reduce injuries and save lives. Both applications require a forecast of safety for a future period. Planning-level crash prediction models for the Tucson, Arizona, metropolitan region are presented to demonstrate the feasibility of such models. Data were separated into fatal, injury, and property-damage crashes. To accommodate overdispersion in the data, negative binomial regression models were applied. To accommodate the simultaneity of fatality and injury crash outcomes, simultaneous estimation of the models was conducted. All models produce crash forecasts at the traffic analysis zone level. Statistically significant (p-values < 0.05) and theoretically meaningful variables for the fatal crash model included population density, persons 17 years old or younger as a percentage of the total population, and intersection density. Significant variables for the injury and property-damage crash models were population density, number of employees, intersections density, percentage of miles of principal arterial, percentage of miles of minor arterials, and percentage of miles of urban collectors. Among several conclusions it is suggested that planning-level safety models are feasible and may play a role in future planning activities. However, caution must be exercised with such models.

Psychometric properties of the parents' fever management scale in a Turkish population

Aim: This paper reports a study designed to assess the psychometric properties (validity and reliability) of a Turkish version of the Australian Parents’ Fever Management Scale (PFMS). Background: Little is known about childhood fever management among Turkish parents. No scales to measure parents’ fever management practices in Turkey are available. Design: This is a methodological study. Methods: Eighty parents, of febrile children aged six months to five years, were randomly selected from the paedaitric hospital and two community family health centers in Sakarya, Turkey. The PFMS was back translated; language equivalence and content validity were validated. PFMS and socio-demographic data were collected in 2009. Means and standard deviations were calculated for interval level data and p values greater than 0.05 were considered statistically significant. Unrotated principal component analysis was used to determine construct validity and Cronbach’s coefficient alpha determined the internal consistency reliability. Results: The PFMS was psychometrically sound in this population. Construct validity, confirmed by confirmatory factor analysis [KMO 0.812, Bartlett’s Specificity (χ² = 182.799, df=28, P < 0·001)] revealed the Turkish version to be comprised of the eight original PFMS items. Internal consistency reliability coefficient was 0.80 and the scale’s total-item correlation coefficients ranged from 0.15 to 0.66 and were significant (p<0.001). Interestingly parents reported high scores on the PFMS 34.52±4.60 (range 8-40 with 40 indicating a high burden of care for febrile children). Conclusion: The PFMS was as psychometrically robust in a Turkish population as in an Australian population and is, therefore, a useful tool for health professionals to identify parents’ practices, provide targeted education thereby in reducing the unnecessary burden of care they place on themselves when caring for a febrile child. Relevance to clinical practice. Testing in different populations, cultures and healthcare systems will further assist in reporting the PFMS usefulness in clinical practice and research.

Development and validation of novel methods for microbial source tracking based on Escherichia coli as an indicator of water quality

Microbial pollution in water periodically affects human health in Australia, particularly in times of drought and flood. There is an increasing need for the control of waterborn microbial pathogens. Methods, allowing the determination of the origin of faecal contamination in water, are generally referred to as Microbial Source Tracking (MST). Various approaches have been evaluated as indicatorsof microbial pathogens in water samples, including detection of different microorganisms and various host-specific markers. However, until today there have been no universal MST methods that could reliably determine the source (human or animal) of faecal contamination. Therefore, the use of multiple approaches is frequently advised. MST is currently recognised as a research tool, rather than something to be included in routine practices. The main focus of this research was to develop novel and universally applicable methods to meet the demands for MST methods in routine testing of water samples. Escherichia coli was chosen initially as the object organism for our studies as, historically and globally, it is the standard indicator of microbial contamination in water. In this thesis, three approaches are described: single nucleotide polymorphism (SNP) genotyping, clustered regularly interspaced short palindromic repeats (CRISPR) screening using high resolution melt analysis (HRMA) methods and phage detection development based on CRISPR types. The advantage of the combination SNP genotyping and CRISPR genes has been discussed in this study. For the first time, a highly discriminatory single nucleotide polymorphism interrogation of E. coli population was applied to identify the host-specific cluster. Six human and one animal-specific SNP profile were revealed. SNP genotyping was successfully applied in the field investigations of the Coomera watershed, South-East Queensland, Australia. Four human profiles [11], [29], [32] and [45] and animal specific SNP profile [7] were detected in water. Two human-specific profiles [29] and [11] were found to be prevalent in the samples over a time period of years. The rainfall (24 and 72 hours), tide height and time, general land use (rural, suburban), seasons, distance from the river mouth and salinity show a lack of relashionship with the diversity of SNP profiles present in the Coomera watershed (p values > 0.05). Nevertheless, SNP genotyping method is able to identify and distinquish between human- and non-human specific E. coli isolates in water sources within one day. In some samples, only mixed profiles were detected. To further investigate host-specificity in these mixed profiles CRISPR screening protocol was developed, to be used on the set of E. coli, previously analysed for SNP profiles. CRISPR loci, which are the pattern of previous DNA coliphages attacks, were considered to be a promising tool for detecting host-specific markers in E. coli. Spacers in CRISPR loci could also reveal the dynamics of virulence in E. coli as well in other pathogens in water. Despite the fact that host-specificity was not observed in the set of E. coli analysed, CRISPR alleles were shown to be useful in detection of the geographical site of sources. HRMA allows determination of ‘different’ and ‘same’ CRISPR alleles and can be introduced in water monitoring as a cost-effective and rapid method. Overall, we show that the identified human specific SNP profiles [11], [29], [32] and [45] can be useful as marker genotypes globally for identification of human faecal contamination in water. Developed in the current study, the SNP typing approach can be used in water monitoring laboratories as an inexpensive, high-throughput and easy adapted protocol. The unique approach based on E. coli spacers for the search for unknown phage was developed to examine the host-specifity in phage sequences. Preliminary experiments on the recombinant plasmids showed the possibility of using this method for recovering phage sequences. Future studies will determine the host-specificity of DNA phage genotyping as soon as first reliable sequences can be acquired. No doubt, only implication of multiple approaches in MST will allow identification of the character of microbial contamination with higher confidence and readability.

Body composition and appetite : fat-free mass (but not fat mass or BMI) is positively associated with self-determined meal size and daily energy intake in humans

The idea of body weight regulation implies that a biological mechanism exerts control over energy expenditure and food intake. This is a central tenet of energy homeostasis. However, the source and identity of the controlling mechanism have not been identified, although it is often presumed to be some long-acting signal related to body fat, such as leptin. Using a comprehensive experimental platform, we have investigated the relationship between biological and behavioural variables in two separate studies over a 12-week intervention period in obese adults (total n 92). All variables have been measured objectively and with a similar degree of scientific control and precision, including anthropometric factors, body composition, RMR and accumulative energy consumed at individual meals across the whole day. Results showed that meal size and daily energy intake (EI) were significantly correlated with fat-free mass (FFM, P values ,0·02–0·05) but not with fat mass (FM) or BMI (P values 0·11–0·45) (study 1, n 58). In study 2 (n 34), FFM (but not FM or BMI) predicted meal size and daily EI under two distinct dietary conditions (high-fat and low-fat). These data appear to indicate that, under these circumstances, some signal associated with lean mass (but not FM) exerts a determining effect over self-selected food consumption. This signal may be postulated to interact with a separate class of signals generated by FM. This finding may have implications for investigations of the molecular control of food intake and body weight and for the management of obesity.

Investigation of lymphotoxin α genetic variants in migraine

Migraine is a common neurological disease with a genetic basis affecting approximately 12% of the population. Pain during a migraine attack is associated with activation of the trigeminal nerve system, which carries pain signals from the meninges and the blood vessels infusing the meninges to the trigeminal nucleus in the brain stem. The release of inflammatory mediators following cortical spreading depression (CSD) may further promote and sustain the activation and sensitization of meningeal nociceptors, inducing the persistent throbbing headache characterised in migraine. Lymphotoxin α (LTA) is a cytokine secreted by lymphocytes and is a member of the tumour necrosis factor (TNF) family. Genetic variation with the TNF and LTA genes may contribute to threshold brain excitability, propagation of neuronal hyperexcitability and thus initiation and maintenance of a migraine attack. Three LTA variants rs2009658, rs2844482 and rs2229094 were identified in a recent pGWAS study conducted in the Norfolk Island population as being potentially implicated in migraine with nominally significant p values of p = 0.0093, p = 0.0088 and p = 0.033 respectively. To determine whether these SNPs played a role in migraine in a general outbred population these SNPs were gentoyped in a large case control Australian Caucasian population and tested for association with migraine. All three SNPs showed no association in our cohort (p > 0.05). Validation of GWAS data in independent case-controls cohorts is essential to establish risk validity within specific population groups. The importance of cytokines in modulating neural inflammation and pain threshold in addition to other studies showing associations between TNF-α and SNPs in the LTA gene with migraine, suggests that LTA could be an important factor contributing to migraine. Although the present study did not support a role for the tested LTA variants in migraine, investigation of other variants within the LTA gene is still warranted.

Investigation of the NOTCH3 and TNFSF7 genes on C19p13 as candidates for migraine

To investigate the migraine locus around the C19p13 region through analysis of the NOTCH3 gene (C19p13.2-p13.1), previously shown to be a gene involved in CADASIL and the TNFSF7 gene (C19p13), homologous to the ligands of TNF-alpha and TNF-beta, genes that have previously been associated with migraine. The NOTCH3 gene was analysed by sequencing all exons with known CADASIL mutations in a typical (non-familial hemiplegic) migraine family (MF1) that has previously been shown to be linked to C19p13. The TNFSF7 gene was investigated through SNP association analysis using a matched case-control migraine population. NOTCH3 gene sequencing results for affected members of MF1 proved to be negative for all known sequence variants giving rise to mutations for CADASIL. TNFSF7 gene chi-square results showed non-significant P values across all populations tested against controls, except for the MO subgroup which displayed a possible association with the TNFSF7 SNP (genotype, allele analysis P = 0.036, P = 0.017 respectively). Our results suggest that common migraine is not caused by any known CADASIL mutations in the NOTCH3 gene of interest. However, the TNFSF7 gene displayed signs of involvement in a MO affected population and indicates that further independent studies of this marker are warranted.

Association analysis of chromosome 1 migraine candidate genes

Migraine with aura (MA) is a subtype of typical migraine. Migraine with aura (MA) also encompasses a rare severe subtype Familial Hemiplegic Migraine (FHM) with several known genetic loci. The type 2 FHM (FHM-2) susceptibility locus maps to chromosome 1q23 and mutations in the ATP1A2 gene at this site have recently been implicated. We have previously provided evidence of linkage of typical migraine (predominantly MA) to microsatellite markers on chromosome 1, in the 1q31 and 1q23 regions. In this study, we have undertaken a large genomic investigation involving candidate genes that lie within the chromosome 1q23 and 1q31 regions using an association analysis approach. Methods We have genotyped a large population of case-controls (243 unrelated Caucasian migraineurs versus 243 controls) examining a set of 5 single nucleotide polymorphisms (SNPs) and the Fas Ligand dinucleotide repeat marker, located within the chromosome 1q23 and 1q31 regions. Results Several genes have been studied including membrane protein (ATP 1 subtype A4 and FasL), cytoplasmic glycoprotein (CASQ 1) genes and potassium (KCN J9 and KCN J10) and calcium (CACNA1E) channel genes in 243 migraineurs (including 85% MA and 15% of migraine without aura (MO)) and 243 matched controls. After correction for multiple testing, chi-square results showed non-significant P values (P > 0.008) across all SNPs (and a CA repeat) tested in these different genes, however results with the KCN J10 marker gave interesting results (P = 0.02) that may be worth exploring further in other populations. Conclusion These results do not show a significant role for the tested candidate gene variants and also do not support the hypothesis that a common chromosome 1 defective gene influences both FHM and the more common forms of migraine.

No role for estrogen receptor 1 gene intron 1 Pvu II and exon 4 C325G polymorphisms in migraine susceptibility

Background We have previously reported an association between the estrogen receptor 1 (ESR1) gene exon 8 G594A polymorphism and migraine susceptibility in two independent Australian cohorts. In this paper we report results of analysis of two further single nucleotide polymorphisms (SNPs) in the ESR1 gene in the same study group, the T/C Pvu II SNP in intron 1 and the C325G SNP in exon 4, as well as results of linkage disequilibrium (LD) analysis on these markers. Methods We investigated these variants by case-control association analysis in a cohort of 240 migraineurs and 240 matched controls. The SNPs were genotyped using specific restriction enzyme assays. Results were analysed using contingency table methods incorporating the chi-squared statistic. LD results are presented as D' statistics with associated P values. Results We found no evidence for association of the Pvu II T/C polymorphism and the C325G polymorphism and migraine susceptibility and no evidence for LD between these two SNPs and the previously implicated exon 8 G594A marker. Conclusion We have found no role for the polymorphisms in intron 1 and exon 4 with migraine susceptibility. To further investigate our previously implicated exon 8 marker, we suggest the need for studies with a high density of polymorphisms be undertaken, with particular focus on markers in LD with the exon 8 marker.

Genome-wide association study identifies a possible susceptibility locus for endometrial cancer

Background: Genome-wide association studies (GWAS) have identified more than 100 genetic loci for various cancers. However, only one is for endometrial cancer. Methods: We conducted a three-stage GWAS including 8,492 endometrial cancer cases and 16,596 controls. After analyzing 585,963 single-nucleotide polymorphisms (SNP) in 832 cases and 2,682 controls (stage I) from the Shanghai Endometrial Cancer Genetics Study, we selected the top 106 SNPs for in silico replication among 1,265 cases and 5,190 controls from the Australian/British Endometrial Cancer GWAS (stage II). Nine SNPs showed results consistent in direction with stage I with P < 0.1. These nine SNPs were investigated among 459 cases and 558 controls (stage IIIa) and six SNPs showed a direction of association consistent with stages I and II. These six SNPs, plus two additional SNPs selected on the basis of linkage disequilibrium and P values in stage II, were investigated among 5,936 cases and 8,166 controls from an additional 11 studies (stage IIIb). Results: SNP rs1202524, near the CAPN9 gene on chromosome 1q42.2, showed a consistent association with endometrial cancer risk across all three stages, with ORs of 1.09 [95% confidence interval (CI), 1.03–1.16] for the A/G genotype and 1.17 (95% CI, 1.05–1.30) for the G/G genotype (P = 1.6 × 10−4 in combined analyses of all samples). The association was stronger when limited to the endometrioid subtype, with ORs (95% CI) of 1.11 (1.04–1.18) and 1.21 (1.08–1.35), respectively (P = 2.4 × 10−5). Conclusions: Chromosome 1q42.2 may host an endometrial cancer susceptibility locus. Impact: This study identified a potential genetic locus for endometrial cancer risk

A pilot randomised trial of Medilixir cream shows some promising results, but control treatment may have been inappropriate

Aim To test the efficacy of Medilixir [cream] against the standard treatment of aqueous cream in the provision of relief from the symptoms of postburn itch. Design RCT with two parallel arms. Setting Professor Stuart Pegg Adult Burns Centre, Royal Brisbane and Women's Hospital, Brisbane, Australia. Participants Fifty-two patients aged between 18 and 80 years, admitted directly to the burns centre between 10 March and 22 July 2008, were able to provide informed consent, and had shown no allergic reaction to a patch test with the study medication, were randomised. Patients admitted from intensive care or high dependency were excluded. Main results Effect estimates and confidence intervals were not reported for any of the outcomes; only group means/proportions and P-values from hypothesis testing were provided. More patients in the intervention group reported itch reduction compared to comparison treatment (91 vs. 82%, P=0.001). Itch recurrence after cream application occurred later in the intervention group compared to the control group (P<0.001). Use of antipruritic medication was significantly greater in the control group (P=0.023). There was no difference in sleep disturbance between groups (not quantified). On average, Medilixir took longer to apply than aqueous cream (157s for Medilixir vs. 139s for aqueous cream; mean difference 17s), but authors noted that the groups did not differ significantly (CI for mean difference and P-values were not reported).

Genetic polymorphisms in miRNAs targeting the estrogen receptor and their effect on breast cancer risk

Breast cancer is the cancer that most commonly affects women worldwide. This type of cancer is genetically complex, but is strongly linked to steroid hormone signalling systems. Because microRNAs act as translational regulators of multiple genes, including the steroid nuclear receptors, single nucleotide polymorphisms (SNPs) in microRNAs genes can have potentially wide-ranging influences on breast cancer development. Thus, this study was conducted to investigate the relationships between six SNPs (rs6977848, rs199981120, rs185641358, rs113054794, rs66461782, and rs12940701) located in four miRNA genes predicted to target the estrogen receptor (miR-148a, miR-221, miR-186, and miR-152) and breast cancer risk in Caucasian Australian women. By using high resolution melt analysis (HRM) and polymerase chain reaction- restriction fragment length polymorphism (PCR-RFLP), 487 samples including 225 controls and 262 cases were genotyped. Analysis of their genotype and allele frequencies indicated that the differences between case and control populations was not significant for rs6977848, rs66461782, and rs12940701 because their p-values are 0.81, 0.93, 0.1 which are all above the threshold value (p=0.05). Our data thus suggests that these SNPs do not affect breast cancer risk in the tested population. In addition, rs199981120, rs185641358, and rs113054794 could not be found in this population, suggesting that these SNPs do not occur in Caucasian Australians.

Biological monitoring in workers in a nitrobenzene reduction plant : Haemoglobin versus serum albumin adducts

The high priority of monitoring workers exposed to nitrobenzene is a consequence of clear findings of experimental carcinogenicity of nitrobenzene and the associated evaluations by the International Agency for Research on Cancer. Eighty male employees of a nitrobenzene reduction plant, with potential skin contact with nitrobenzene and aniline, participated in a current medical surveillance programme. Blood samples were routinely taken and analysed for aniline, 4-aminodiphenyl (4-ADP) and benzidine adducts of haemoglobin (Hb) and human serum albumin (HSA). Also, levels of methaemoglobin (Met-Hb) and of carbon monoxide haemoglobin (CO-Hb) were monitored. Effects of smoking were straightforward. Using the rank sum test of Wilcoxon, we found that very clear-cut and statistically significant smoking effects (about 3-fold increases) were apparent on CO-Hb (P = 0.00085) and on the Hb adduct of 4-ADP (P = 0.0006). The mean aniline-Hb adduct level in smokers was 1.5 times higher than in non-smokers; the significance (P = 0.05375) was close to the 5% level. The strongest correlation was evident between the Hb and HSA adducts of aniline (rs = 0.846). Less pronounced correlations (but with P values < 0.02) appeared between aniline-Hb and 4-ADP-Hb adducts (rs = 0.388), between 4-ADP and 4-ADP-HSA adducts (rs = 0.373), and between 4-ADP-Hb and aniline-HSA adducts (rs = 0.275). In view of the proposal for additional use of the aniline-HSA adduct for biological monitoring, particularly in cases of acute overexposures or poisonings, the strong correlation of the Hb and HSA conjugates is noteworthy; the ratio aniline-HSA:aniline-Hb was 1:42 for the entire cohort.