A New Statistic to Evaluate Imputation Reliability

Background

As the amount of data from genome wide association studies grows dramatically, many interesting scientific questions require imputation to combine or expand datasets. However, there are two situations for which imputation has been problematic: (1) polymorphisms with low minor allele frequency (MAF), and (2) datasets where subjects are genotyped on different platforms. Traditional measures of imputation cannot effectively address these problems.

Methodology/Principal Findings

We introduce a new statistic, the imputation quality score (IQS). In order to differentiate between well-imputed and poorly-imputed single nucleotide polymorphisms (SNPs), IQS adjusts the concordance between imputed and genotyped SNPs for chance. We first evaluated IQS in relation to minor allele frequency. Using a sample of subjects genotyped on the Illumina 1 M array, we extracted those SNPs that were also on the Illumina 550 K array and imputed them to the full set of the 1 M SNPs. As expected, the average IQS value drops dramatically with a decrease in minor allele frequency, indicating that IQS appropriately adjusts for minor allele frequency. We then evaluated whether IQS can filter poorly-imputed SNPs in situations where cases and controls are genotyped on different platforms. Randomly dividing the data into “cases” and “controls”, we extracted the Illumina 550 K SNPs from the cases and imputed the remaining Illumina 1 M SNPs. The initial Q-Q plot for the test of association between cases and controls was grossly distorted (λ = 1.15) and had 4016 false positives, reflecting imputation error. After filtering out SNPs with IQS<0.9, the Q-Q plot was acceptable and there were no longer false positives. We then evaluated the robustness of IQS computed independently on the two halves of the data. In both European Americans and African Americans the correlation was >0.99 demonstrating that a database of IQS values from common imputations could be used as an effective filter to combine data genotyped on different platforms.

Conclusions/Significance

IQS effectively differentiates well-imputed and poorly-imputed SNPs. It is particularly useful for SNPs with low minor allele frequency and when datasets are genotyped on different platforms.     

Associations between retinol-binding protein 4 and cardiometabolic risk factors and subclinical atherosclerosis in recently postmenopausal women: Cross-sectional analyses from the KEEPS Study

ABSTRACT: BACKGROUND: The published literature regarding the relationships between retinol-binding protein 4 (RBP4) and cardiometabolic risk factors and subclinical atherosclerosis is conflicting, likely due, in part, to limitations of frequently used RBP4 assays. Prior large studies have not utilized the gold-standard Western Blot analysis of RBP4 levels. METHODS: Full-length serum RBP4 levels were measured by Western Blot in 709 postmenopausal women screened for the Kronos Early Estrogen Prevention Study. Cross-sectional analyses related RBP4 levels to cardiometabolic risk factors, carotid artery intima-media thickness (CIMT), and coronary artery calcification (CAC). RESULTS: The mean age of women was 52.9 (+/- 2.6) years, and the median RBP4 level was 49.0 (interquartile range 36.9-61.5) ug/mL. Higher RBP4 levels were weakly associated with higher triglycerides (age, race, and smoking-adjusted partial Spearman correlation coefficient= 0.10; P=0.01), but were unrelated to blood pressure, cholesterol, C-reactive protein, glucose, insulin, and CIMT levels (all partial Spearman correlation coefficients [less than or equal to]0.06, P>0.05). Results suggested a curvilinear association between RBP4 levels and CAC, with women in the bottom and upper quartiles of RBP4 having higher odds of CAC (odds ratio [95% confidence interval] 2.10 [1.07-4.09], 2.00 [1.02-3.92], 1.64 [0.82-3.27] for the 1st, 3rd, and 4th RBP4 quartiles vs. the 2nd quartile). However, a squared RBP4 term in regression modeling was non-significant (P=0.10). CONCLUSIONS: In these healthy, recently postmenopausal women, higher RBP4 levels were weakly associated with elevations in triglycerides and with CAC, but not with other risk factors or CIMT. These data using the gold standard of RBP4 methodology only weakly support the possibility that perturbations in RBP4 homeostasis may be an additional risk factor for subclinical coronary atherosclerosis. Trial Registration: ClinicalTrials.gov number NCT00154180.     

Sodium phenylbutyrate controls neuroinflammatory and antioxidant activities and protects dopaminergic neurons in mouse models of Parkinson's disease

Neuroinflammation and oxidative stress underlie the pathogenesis of various neurodegenerative disorders. Here we demonstrate that sodium phenylbutyrate (NaPB), an FDA-approved therapy for reducing plasma ammonia and glutamine in urea cycle disorders, can suppress both proinflammatory molecules and reactive oxygen species (ROS) in activated glial cells. Interestingly, NaPB also decreased the level of cholesterol but involved only intermediates, not the end product of cholesterol biosynthesis pathway for these functions. While inhibitors of both geranylgeranyl transferase (GGTI) and farnesyl transferase (FTI) inhibited the activation of NF-κB, inhibitor of GGTI, but not FTI, suppressed the production of ROS. Accordingly, a dominant-negative mutant of p21(rac), but not p21(ras), attenuated the production of ROS from activated microglia. Inhibition of both p21(ras) and p21(rac) activation by NaPB in microglial cells suggests that NaPB exerts anti-inflammatory and antioxidative effects via inhibition of these small G proteins. Consistently, we found activation of both p21(ras) and p21(rac)in vivo in the substantia nigra of acute 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mouse model of Parkinson's disease. Oral administration of NaPB reduced nigral activation of p21(ras) and p21(rac), protected nigral reduced glutathione, attenuated nigral activation of NF-κB, inhibited nigral expression of proinflammatory molecules, and suppressed nigral activation of glial cells. These findings paralleled dopaminergic neuronal protection, normalized striatal neurotransmitters, and improved motor functions in MPTP-intoxicated mice. Consistently, FTI and GGTI also protected nigrostriata in MPTP-intoxicated mice. Furthermore, NaPB also halted the disease progression in a chronic MPTP mouse model. These results identify novel mode of action of NaPB and suggest that NaPB may be of therapeutic benefit for neurodegenerative disorders.     

Physcomitrella patens has lipoxygenases for both eicosanoid and octadecanoid pathways

Mosses have substantial amounts of long chain C20 polyunsaturated fatty acids, such as arachidonic and eicosapentaenoic acid, in addition to the shorter chain C18 α-linolenic and linoleic acids, which are typical substrates of lipoxygenases in flowering plants. To identify the fatty acid substrates used by moss lipoxygenases, eight lipoxygenase genes from Physcomitrella patens were heterologously expressed in Escherichia coli, and then analyzed for lipoxygenase activity using linoleic, α-linolenic and arachidonic acids as substrates. Among the eight moss lipoxygenases, only seven were found to be enzymatically active in vitro, two of which selectively used arachidonic acid as the substrate, while the other five preferred α-linolenic acid. Based on enzyme assays using a Clark-type oxygen electrode, all of the active lipoxygenases had an optimum pH at 7.0, except for one with highest activity at pH 5.0. HPLC analyses indicated that the two arachidonic acid lipoxygenases form (12S)-hydroperoxy eicosatetraenoic acid as the main product, while the other five lipoxygenases produce mainly (13S)-hydroperoxy octadecatrienoic acid from α-linolenic acid. These results suggest that mosses may have both C20 and C18 based oxylipin pathways.     

Adaptation of Saffold virus 2 for high-titer growth in mammalian cells

Saffold viruses (SAFV) are a recently discovered group of human Cardioviruses closely related to Theiler's murine encephalomyelitis viruses (TMEV). Unlike TMEV and encephalomyocarditis virus, each of which is monotypic, SAFV are genetically diverse and include at least eight genotypes. To date, only Saffold virus 3 (SAFV-3) has been grown efficiently in mammalian cells in vitro. Here, we report the successful adaptation of SAFV-2 for efficient growth in HeLa cells after 13 passages in the alpha/beta interferon-deficient human glial cell line U118 MG. Nine amino acid changes were found in the adapted virus, with single mutations in VP2, VP3, and 2B, while 6 mutations arose in VP1. Most capsid mutations were in surface loops. Analysis of SAFV-2 revealed virus growth and cytopathic effect only in human cell lines, with large plaques forming in HeLa cells, with minimal cell association, and without using sialic acid to enter cells. Despite the limited growth of SAFV-2 in rodent cells in vitro, BALB/c mice inoculated with SAFV-2 showed antibody titers of >1:10(6), and fluorescence-activated cell sorting (FACS) analysis revealed only minimal cross-reactivity with SFV-3. Intracerebral inoculation of 6-week-old FVB/n mice produced paralysis and acute neuropathological changes, including meningeal infiltrates, encephalitis, particularly of the limbic system, and spinal cord white matter inflammation.     

Surfactant protein D reduces alveolar macrophage apoptosis in vivo

Surfactant protein D (SP-D) is a molecule of the innate immune system that recognizes the patterns of surface carbohydrate on pathogens and targets them for phagocytosis and killing. SP-D-deficient mice show an increased number of macrophages in the alveolar space, excess surfactant phospholipid, overproduction of reactive oxygen species, and the development of emphysema. We report here that SP-D-deficient mice have a 5- to 10-fold increase in the number of apoptotic and necrotic alveolar macrophages, as defined by annexin V and propidium iodine staining, respectively. Intrapulmonary administration of a truncated 60-kDa fragment of human recombinant SP-D reduces the number of apoptotic and necrotic alveolar macrophages and partially corrects the lipid accumulation in SP-D-deficient mice. The same SP-D fragment binds preferentially to apoptotic and necrotic alveolar macrophages in vitro, suggesting that SP-D contributes to immune homeostasis in the lung by recognizing and promoting removal of necrotic and apoptotic cells.     

Plasmacytoid Dendritic Cells Sequester High Prion Titres at Early Stages of Prion Infection

In most transmissible spongiform encephalopathies prions accumulate in the lymphoreticular system (LRS) long before they are detectable in the central nervous system. While a considerable body of evidence showed that B lymphocytes and follicular dendritic cells play a major role in prion colonization of lymphoid organs, the contribution of various other cell types, including antigen-presenting cells, to the accumulation and the spread of prions in the LRS are not well understood. A comprehensive study to compare prion titers of candidate cell types has not been performed to date, mainly due to limitations in the scope of animal bioassays where prohibitively large numbers of mice would be required to obtain sufficiently accurate data. By taking advantage of quantitative in vitro prion determination and magnetic-activated cell sorting, we studied the kinetics of prion accumulation in various splenic cell types at early stages of prion infection. Robust estimates for infectious titers were obtained by statistical modelling using a generalized linear model. Whilst prions were detectable in B and T lymphocytes and in antigen-presenting cells like dendritic cells and macrophages, highest infectious titers were determined in two cell types that have previously not been associated with prion pathogenesis, plasmacytoid dendritic (pDC) and natural killer (NK) cells. At 30 days after infection, NK cells were more than twice, and pDCs about seven-fold, as infectious as lymphocytes respectively. This result was unexpected since, in accordance to previous reports prion protein, an obligate requirement for prion replication, was undetectable in pDCs. This underscores the importance of prion sequestration and dissemination by antigen-presenting cells which are among the first cells of the immune system to encounter pathogens. We furthermore report the first evidence for a release of prions from lymphocytes and DCs of scrapie-infected mice ex vivo, a process that is associated with a release of exosome-like membrane vesicles.