Model uncertainty and forecast accuracy

In time-series analysis, a model is rarely pre-specified but rather is typically formulated in an iterative, interactive way using the given time-series data. Unfortunately the properties of the fitted model, and the forecasts from it, are generally calculated as if the model were known in the first place. This is theoretically incorrect, as least squares theory, for example, does not apply when the same data are used to formulates and fit a model. Ignoring prior model selection leads to biases, not only in estimates of model parameters but also in the subsequent construction of prediction intervals. The latter are typically too narrow, partly because they do not allow for model uncertainty. Empirical results also suggest that more complicated models tend to give a better fit but poorer ex-ante forecasts. The reasons behind these phenomena are reviewed. When comparing different forecasting models, the BIC is preferred to the AIC for identifying a model on the basis of within-sample fit, but out-of-sample forecasting accuracy provides the real test. Alternative approaches to forecasting, which avoid conditioning on a single model, include Bayesian model averaging and using a forecasting method which is not model-based but which is designed to be adaptable and robust.     

Bicorrelations and cross‐bicorrelations as non‐linearity tests and tools for exchange rate forecasting

This paper proposes and implements a new methodology for forecasting time series, based on bicorrelations and cross‐bicorrelations. It is shown that the forecasting technique arises as a natural extension of, and as a complement to, existing univariate and multivariate non‐linearity tests. The formulations are essentially modified autoregressive or vector autoregressive models respectively, which can be estimated using ordinary least squares. The techniques are applied to a set of high‐frequency exchange rate returns, and their out‐of‐sample forecasting performance is compared to that of other time series models. Copyright © 2001 John Wiley & Sons, Ltd.     

Common variants near ATM are associated with glycemic response to metformin in type 2 diabetes

Metformin is the most commonly used pharmacological therapy for type 2 diabetes. We report a genome-wide association study for glycemic response to metformin in 1,024 Scottish individuals with type 2 diabetes with replication in two cohorts including 1,783 Scottish individuals and 1,113 individuals from the UK Prospective Diabetes Study. In a combined meta-analysis, we identified a SNP, rs11212617, associated with treatment success (n = 3,920, P = 2.9 × 10(-9), odds ratio = 1.35, 95% CI 1.22-1.49) at a locus containing ATM, the ataxia telangiectasia mutated gene. In a rat hepatoma cell line, inhibition of ATM with KU-55933 attenuated the phosphorylation and activation of AMP-activated protein kinase in response to metformin. We conclude that ATM, a gene known to be involved in DNA repair and cell cycle control, plays a role in the effect of metformin upstream of AMP-activated protein kinase, and variation in this gene alters glycemic response to metformin.     

Including known covariates can reduce power to detect genetic effects in case-control studies

Genome-wide association studies (GWAS) search for associations between genetic variants and disease status, typically via logistic regression. Often there are covariates, such as sex or well-established major genetic factors, that are known to affect disease susceptibility and are independent of tested genotypes at the population level. We show theoretically and with data from recent GWAS on multiple sclerosis, psoriasis and ankylosing spondylitis that inclusion of known covariates can substantially reduce power for the identification of associated variants when the disease prevalence is lower than a few percent. Whether the inclusion of such covariates reduces or increases power to detect genetic effects depends on various factors, including the prevalence of the disease studied. When the disease is common (prevalence of >20%), the inclusion of covariates typically increases power, whereas, for rarer diseases, it can often decrease power to detect new genetic associations.     

Detectable clonal mosaicism from birth to old age and its relationship to cancer

We detected clonal mosaicism for large chromosomal anomalies (duplications, deletions and uniparental disomy) using SNP microarray data from over 50,000 subjects recruited for genome-wide association studies. This detection method requires a relatively high frequency of cells with the same abnormal karyotype (>5-10%; presumably of clonal origin) in the presence of normal cells. The frequency of detectable clonal mosaicism in peripheral blood is low (<0.5%) from birth until 50 years of age, after which it rapidly rises to 2-3% in the elderly. Many of the mosaic anomalies are characteristic of those found in hematological cancers and identify common deleted regions with genes previously associated with these cancers. Although only 3% of subjects with detectable clonal mosaicism had any record of hematological cancer before DNA sampling, those without a previous diagnosis have an estimated tenfold higher risk of a subsequent hematological cancer (95% confidence interval = 6-18).     

Overexpression of Fto leads to increased food intake and results in obesity

Genome-wide association studies have identified SNPs within FTO, the human fat mass and obesity-associated gene, that are strongly associated with obesity. Individuals homozygous for the at-risk rs9939609 A allele weigh, on average, ~3 kg more than individuals with the low-risk T allele. Mice that lack FTO function and/or Fto expression display increased energy expenditure and a lean phenotype. We show here that ubiquitous overexpression of Fto leads to a dose-dependent increase in body and fat mass, irrespective of whether mice are fed a standard or a high-fat diet. Our results suggest that increased body mass results primarily from increased food intake. Mice with increased Fto expression on a high-fat diet develop glucose intolerance. This study provides the first direct evidence that increased Fto expression causes obesity in mice.     

A genome-wide association study identifies new psoriasis susceptibility loci and an interaction between HLA-C and ERAP1

To identify new susceptibility loci for psoriasis, we undertook a genome-wide association study of 594,224 SNPs in 2,622 individuals with psoriasis and 5,667 controls. We identified associations at eight previously unreported genomic loci. Seven loci harbored genes with recognized immune functions (IL28RA, REL, IFIH1, ERAP1, TRAF3IP2, NFKBIA and TYK2). These associations were replicated in 9,079 European samples (six loci with a combined P < 5 × 10?? and two loci with a combined P < 5 × 10??). We also report compelling evidence for an interaction between the HLA-C and ERAP1 loci (combined P = 6.95 × 10??). ERAP1 plays an important role in MHC class I peptide processing. ERAP1 variants only influenced psoriasis susceptibility in individuals carrying the HLA-C risk allele. Our findings implicate pathways that integrate epidermal barrier dysfunction with innate and adaptive immune dysregulation in psoriasis pathogenesis.     

Interaction between ERAP1 and HLA-B27 in ankylosing spondylitis implicates peptide handling in the mechanism for HLA-B27 in disease susceptibility

Ankylosing spondylitis is a common form of inflammatory arthritis predominantly affecting the spine and pelvis that occurs in approximately 5 out of 1,000 adults of European descent. Here we report the identification of three variants in the RUNX3, LTBR-TNFRSF1A and IL12B regions convincingly associated with ankylosing spondylitis (P < 5 × 10(-8) in the combined discovery and replication datasets) and a further four loci at PTGER4, TBKBP1, ANTXR2 and CARD9 that show strong association across all our datasets (P < 5 × 10(-6) overall, with support in each of the three datasets studied). We also show that polymorphisms of ERAP1, which encodes an endoplasmic reticulum aminopeptidase involved in peptide trimming before HLA class I presentation, only affect ankylosing spondylitis risk in HLA-B27-positive individuals. These findings provide strong evidence that HLA-B27 operates in ankylosing spondylitis through a mechanism involving aberrant processing of antigenic peptides.