A Novel Approach for Prediction of Vitamin D Status Using Support Vector Regression

Background

Epidemiological evidence suggests that vitamin D deficiency is linked to various chronic diseases. However direct measurement of serum 25-hydroxyvitamin D (25(OH)D) concentration, the accepted biomarker of vitamin D status, may not be feasible in large epidemiological studies. An alternative approach is to estimate vitamin D status using a predictive model based on parameters derived from questionnaire data. In previous studies, models developed using Multiple Linear Regression (MLR) have explained a limited proportion of the variance and predicted values have correlated only modestly with measured values. Here, a new modelling approach, nonlinear radial basis function support vector regression (RBF SVR), was used in prediction of serum 25(OH)D concentration. Predicted scores were compared with those from a MLR model.

Methods

Determinants of serum 25(OH)D in Caucasian adults (n = 494) that had been previously identified were modelled using MLR and RBF SVR to develop a 25(OH)D prediction score and then validated in an independent dataset. The correlation between actual and predicted serum 25(OH)D concentrations was analysed with a Pearson correlation coefficient.

Results

Better correlation was observed between predicted scores and measured 25(OH)D concentrations using the RBF SVR model in comparison with MLR (Pearson correlation coefficient: 0.74 for RBF SVR; 0.51 for MLR). The RBF SVR model was more accurately able to identify individuals with lower 25(OH)D levels (<75 nmol/L).

Conclusion

Using identical determinants, the RBF SVR model provided improved prediction of serum 25(OH)D concentrations and vitamin D deficiency compared with a MLR model, in this dataset.     

Mapping the Genetic Variation of Regional Brain Volumes as Explained by All Common SNPs from the ADNI Study

Typically twin studies are used to investigate the aggregate effects of genetic and environmental influences on brain phenotypic measures. Although some phenotypic measures are highly heritable in twin studies, SNPs (single nucleotide polymorphisms) identified by genome-wide association studies (GWAS) account for only a small fraction of the heritability of these measures. We mapped the genetic variation (the proportion of phenotypic variance explained by variation among SNPs) of volumes of pre-defined regions across the whole brain, as explained by 512,905 SNPs genotyped on 747 adult participants from the Alzheimer''s Disease Neuroimaging Initiative (ADNI). We found that 85% of the variance of intracranial volume (ICV) (p = 0.04) was explained by considering all SNPs simultaneously, and after adjusting for ICV, total grey matter (GM) and white matter (WM) volumes had genetic variation estimates near zero (p = 0.5). We found varying estimates of genetic variation across 93 non-overlapping regions, with asymmetry in estimates between the left and right cerebral hemispheres. Several regions reported in previous studies to be related to Alzheimer''s disease progression were estimated to have a large proportion of volumetric variance explained by the SNPs.     

Using Manifold Learning for Atlas Selection in Multi-Atlas Segmentation

Multi-atlas segmentation has been widely used to segment various anatomical structures. The success of this technique partly relies on the selection of atlases that are best mapped to a new target image after registration. Recently, manifold learning has been proposed as a method for atlas selection. Each manifold learning technique seeks to optimize a unique objective function. Therefore, different techniques produce different embeddings even when applied to the same data set. Previous studies used a single technique in their method and gave no reason for the choice of the manifold learning technique employed nor the theoretical grounds for the choice of the manifold parameters. In this study, we compare side-by-side the results given by 3 manifold learning techniques (Isomap, Laplacian Eigenmaps and Locally Linear Embedding) on the same data set. We assess the ability of those 3 different techniques to select the best atlases to combine in the framework of multi-atlas segmentation. First, a leave-one-out experiment is used to optimize our method on a set of 110 manually segmented atlases of hippocampi and find the manifold learning technique and associated manifold parameters that give the best segmentation accuracy. Then, the optimal parameters are used to automatically segment 30 subjects from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). For our dataset, the selection of atlases with Locally Linear Embedding gives the best results. Our findings show that selection of atlases with manifold learning leads to segmentation accuracy close to or significantly higher than the state-of-the-art method and that accuracy can be increased by fine tuning the manifold learning process.     

Total Serum IgE in a Population-Based Study of Asian Children in Taiwan: Reference Value and Significance in the Diagnosis of Allergy

Background

Total serum immunoglobulin (IgE) test is usually performed to aid in the diagnosis of allergic diseases, but its reference values may vary among people of different ethnic backgrounds.

Objectives

To establish reference values of total IgE in Asian children and to assess their significance in the diagnosis of atopy and allergic diseases.

Study design

1321 Asian children aged 5-18 years in the Prediction of Allergies in Taiwanese CHildren (PATCH) study, a population-based cohort, were evaluated for total and specific IgE by ImmunoCAP and Phadiatop Infant, respectively.

Results

Male, atopy, allergic diseases, recent symptoms of upper respiratory infection, and lower FEV₁/FVC, were associated with higher total IgE levels in univariate analyses. Multivariate analysis revealed that atopy was the single most important determinant explaining 66.1% of the variability of total IgE levels in this population. The area under the receiver-operator characteristic (ROC) curve of total IgE for diagnosing atopy, asthma, rhinitis, and eczema were 0.92, 0.72, 0.70, and 0.70, respectively. The sensitivity, specificity, and positive and negative predictive values of total IgE at the optimal cutoff of 77.7 kU/L on the ROC curve for diagnosing atopy were 82.3%, 87.1%, 89.5%, and 78.6%, respectively. The corresponding values using the upper 95% CI of total IgE (164.3 kU/L) in non-atopic children were 61.2%, 95.0%, 94.3%, and 64.6%, respectively; whereas a customary cutoff (100 kU/L) provided accuracy between that of the aforementioned two cutoffs. Total IgE at the cutoff of 77.7 kU/L provided modest sensitivity and specificity (49.0%-78.3%) for diagnosing allergic diseases, but had high negative predictive values (84.2%-97.9%).

Conclusions

Total serum IgE discriminates Asian children with and without atopy independent of allergic symptoms, with an optimal cutoff of 77.7 kU/L. The study confirms the insufficient diagnostic accuracy of total IgE alone to detect allergic diseases, but low total IgE levels may help exclude allergic diseases.     

Association of TERC and OBFC1 Haplotypes with Mean Leukocyte Telomere Length and Risk for Coronary Heart Disease

Objective

To replicate the associations of leukocyte telomere length (LTL) with variants at four loci and to investigate their associations with coronary heart disease (CHD) and type II diabetes (T2D), in order to examine possible causal effects of telomere maintenance machinery on disease aetiology.

Methods

Four SNPs at three loci BICD1 (rs2630578 GγC), 18q12.2 (rs2162440 GγT), and OBFC1 (rs10786775 CγG, rs11591710 AγC) were genotyped in four studies comprised of 2353 subjects out of which 1148 had CHD and 566 T2D. Three SNPs (rs12696304 CγG, rs10936601G>T and rs16847897 GγC) at the TERC locus were genotyped in these four studies, in addition to an offspring study of 765 healthy students. For all samples, LTL had been measured using a real-time PCR-based method.

Results

Only one SNP was associated with a significant effect on LTL, with the minor allele G of OBFC1 rs10786775 SNP being associated with longer LTL (β=0.029, P=0.04). No SNPs were significantly associated with CHD or T2D. For OBFC1 the haplotype carrying both rare alleles (rs10786775G and rs11591710C, haplotype frequency 0.089) was associated with lower CHD prevalence (OR: 0.77; 95% CI: 0.61–0.97; P= 0.03). The TERC haplotype GTC (rs12696304G, rs10936601T and rs16847897C, haplotype frequency 0.210) was associated with lower risk for both CHD (OR: 0.86; 95% CI: 0.75-0.99; P=0.04) and T2D (OR: 0.74; 95% CI: 0.61–0.91; P= 0.004), with no effect on LTL. Only the last association remained after adjusting for multiple testing.

Conclusion

Of reported associations, only that between the OBFC1 rs10786775 SNP and LTL was confirmed, although our study has a limited power to detect modest effects. A 2-SNP OBFC1 haplotype was associated with higher risk of CHD, and a 3-SNP TERC haplotype was associated with both higher risk of CHD and T2D. Further work is required to confirm these results and explore the mechanisms of these effects.     

Microarray Analysis of Microbiota of Gingival Lesions in Noma Patients

Noma (cancrum oris) is a gangrenous disease of unknown etiology affecting the maxillo-facial region of young children in extremely limited resource countries. In an attempt to better understand the microbiological events occurring during this disease, we used phylogenetic and low-density microarrays targeting the 16S rRNA gene to characterize the gingival flora of acute noma and acute necrotizing gingivitis (ANG) lesions, and compared them to healthy control subjects of the same geographical and social background. Our observations raise doubts about Fusobacterium necrophorum, a previously suspected causative agent of noma, as this species was not associated with noma lesions. Various oral pathogens were more abundant in noma lesions, notably Atopobium spp., Prevotella intermedia, Peptostreptococcus spp., Streptococcus pyogenes and Streptococcus anginosus. On the other hand, pathogens associated with periodontal diseases such as Aggregatibacter actinomycetemcomitans, Capnocytophaga spp., Porphyromonas spp. and Fusobacteriales were more abundant in healthy controls. Importantly, the overall loss of bacterial diversity observed in noma samples as well as its homology to that of ANG microbiota supports the hypothesis that ANG might be the immediate step preceding noma.     

Life Expectancies of South African Adults Starting Antiretroviral Treatment: Collaborative Analysis of Cohort Studies

Background

Few estimates exist of the life expectancy of HIV-positive adults receiving antiretroviral treatment (ART) in low- and middle-income countries. We aimed to estimate the life expectancy of patients starting ART in South Africa and compare it with that of HIV-negative adults.

Methods and Findings

Data were collected from six South African ART cohorts. Analysis was restricted to 37,740 HIV-positive adults starting ART for the first time. Estimates of mortality were obtained by linking patient records to the national population register. Relative survival models were used to estimate the excess mortality attributable to HIV by age, for different baseline CD4 categories and different durations. Non-HIV mortality was estimated using a South African demographic model. The average life expectancy of men starting ART varied between 27.6 y (95% CI: 25.2–30.2) at age 20 y and 10.1 y (95% CI: 9.3–10.8) at age 60 y, while estimates for women at the same ages were substantially higher, at 36.8 y (95% CI: 34.0–39.7) and 14.4 y (95% CI: 13.3–15.3), respectively. The life expectancy of a 20-y-old woman was 43.1 y (95% CI: 40.1–46.0) if her baseline CD4 count was ≥200 cells/µl, compared to 29.5 y (95% CI: 26.2–33.0) if her baseline CD4 count was <50 cells/µl. Life expectancies of patients with baseline CD4 counts ≥200 cells/µl were between 70% and 86% of those in HIV-negative adults of the same age and sex, and life expectancies were increased by 15%–20% in patients who had survived 2 y after starting ART. However, the analysis was limited by a lack of mortality data at longer durations.

Conclusions

South African HIV-positive adults can have a near-normal life expectancy, provided that they start ART before their CD4 count drops below 200 cells/µl. These findings demonstrate that the near-normal life expectancies of HIV-positive individuals receiving ART in high-income countries can apply to low- and middle-income countries as well. Please see later in the article for the Editors'' Summary     

Associations between Active Travel to Work and Overweight,Hypertension, and Diabetes in India: A Cross-Sectional Study

Background

Increasing active travel (walking, bicycling, and public transport) is promoted as a key strategy to increase physical activity and reduce the growing burden of noncommunicable diseases (NCDs) globally. Little is known about patterns of active travel or associated cardiovascular health benefits in low- and middle-income countries. This study examines mode and duration of travel to work in rural and urban India and associations between active travel and overweight, hypertension, and diabetes.

Methods and Findings

Cross-sectional study of 3,902 participants (1,366 rural, 2,536 urban) in the Indian Migration Study. Associations between mode and duration of active travel and cardiovascular risk factors were assessed using random-effect logistic regression models adjusting for age, sex, caste, standard of living, occupation, factory location, leisure time physical activity, daily fat intake, smoking status, and alcohol use. Rural dwellers were significantly more likely to bicycle (68.3% versus 15.9%; p<0.001) to work than urban dwellers. The prevalence of overweight or obesity was 50.0%, 37.6%, 24.2%, 24.9%; hypertension was 17.7%, 11.8%, 6.5%, 9.8%; and diabetes was 10.8%, 7.4%, 3.8%, 7.3% in participants who travelled to work by private transport, public transport, bicycling, and walking, respectively. In the adjusted analysis, those walking (adjusted risk ratio [ARR] 0.72; 95% CI 0.58–0.88) or bicycling to work (ARR 0.66; 95% CI 0.55–0.77) were significantly less likely to be overweight or obese than those travelling by private transport. Those bicycling to work were significantly less likely to have hypertension (ARR 0.51; 95% CI 0.36–0.71) or diabetes (ARR 0.65; 95% CI 0.44–0.95). There was evidence of a dose-response relationship between duration of bicycling to work and being overweight, having hypertension or diabetes. The main limitation of the study is the cross-sectional design, which limits causal inference for the associations found.

Conclusions

Walking and bicycling to work was associated with reduced cardiovascular risk in the Indian population. Efforts to increase active travel in urban areas and halt declines in rural areas should be integral to strategies to maintain healthy weight and prevent NCDs in India. Please see later in the article for the Editors'' Summary     

Generative FDG-PET and MRI Model of Aging and Disease Progression in Alzheimer's Disease

The failure of current strategies to provide an explanation for controversial findings on the pattern of pathophysiological changes in Alzheimer''s Disease (AD) motivates the necessity to develop new integrative approaches based on multi-modal neuroimaging data that captures various aspects of disease pathology. Previous studies using [18F]fluorodeoxyglucose positron emission tomography (FDG-PET) and structural magnetic resonance imaging (sMRI) report controversial results about time-line, spatial extent and magnitude of glucose hypometabolism and atrophy in AD that depend on clinical and demographic characteristics of the studied populations. Here, we provide and validate at a group level a generative anatomical model of glucose hypo-metabolism and atrophy progression in AD based on FDG-PET and sMRI data of 80 patients and 79 healthy controls to describe expected age and symptom severity related changes in AD relative to a baseline provided by healthy aging. We demonstrate a high level of anatomical accuracy for both modalities yielding strongly age- and symptom-severity- dependant glucose hypometabolism in temporal, parietal and precuneal regions and a more extensive network of atrophy in hippocampal, temporal, parietal, occipital and posterior caudate regions. The model suggests greater and more consistent changes in FDG-PET compared to sMRI at earlier and the inversion of this pattern at more advanced AD stages. Our model describes, integrates and predicts characteristic patterns of AD related pathology, uncontaminated by normal age effects, derived from multi-modal data. It further provides an integrative explanation for findings suggesting a dissociation between early- and late-onset AD. The generative model offers a basis for further development of individualized biomarkers allowing accurate early diagnosis and treatment evaluation.     

Reconstructing Native American Migrations from Whole-Genome and Whole-Exome Data

There is great scientific and popular interest in understanding the genetic history of populations in the Americas. We wish to understand when different regions of the continent were inhabited, where settlers came from, and how current inhabitants relate genetically to earlier populations. Recent studies unraveled parts of the genetic history of the continent using genotyping arrays and uniparental markers. The 1000 Genomes Project provides a unique opportunity for improving our understanding of population genetic history by providing over a hundred sequenced low coverage genomes and exomes from Colombian (CLM), Mexican-American (MXL), and Puerto Rican (PUR) populations. Here, we explore the genomic contributions of African, European, and especially Native American ancestry to these populations. Estimated Native American ancestry is in MXL, in CLM, and in PUR. Native American ancestry in PUR is most closely related to populations surrounding the Orinoco River basin, confirming the Southern America ancestry of the Taíno people of the Caribbean. We present new methods to estimate the allele frequencies in the Native American fraction of the populations, and model their distribution using a demographic model for three ancestral Native American populations. These ancestral populations likely split in close succession: the most likely scenario, based on a peopling of the Americas thousand years ago (kya), supports that the MXL Ancestors split kya, with a subsequent split of the ancestors to CLM and PUR kya. The model also features effective populations of in Mexico, in Colombia, and in Puerto Rico. Modeling Identity-by-descent (IBD) and ancestry tract length, we show that post-contact populations also differ markedly in their effective sizes and migration patterns, with Puerto Rico showing the smallest effective size and the earlier migration from Europe. Finally, we compare IBD and ancestry assignments to find evidence for relatedness among European founders to the three populations.