Fast and Accurate Approaches for Large-Scale, Automated Mapping of Food Diaries on Food Composition Tables.

Aim of Study: The use of weighed food diaries in nutritional studies provides a powerful method to quantify food and nutrient intakes. Yet, mapping these records onto food composition tables (FCTs) is a challenging, time-consuming and error-prone process. Experts make this effort manually and no automation has been previously proposed. Our study aimed to assess automated approaches to map food items onto FCTs. Methods: We used food diaries (~170,000 records pertaining to 4,200 unique food items) from the DiOGenes randomized clinical trial. We attempted to map these items onto six FCTs available from the EuroFIR resource. Two approaches were tested: the first was based solely on food name similarity (fuzzy matching). The second used a machine learning approach (C5.0 classifier) combining both fuzzy matching and food energy. We tested mapping food items using their original names and also an English-translation. Top matching pairs were reviewed manually to derive performance metrics: precision (the percentage of correctly mapped items) and recall (percentage of mapped items). Results: The simpler approach: fuzzy matching, provided very good performance. Under a relaxed threshold (score > 50%), this approach enabled to remap 99.49% of the items with a precision of 88.75%. With a slightly more stringent threshold (score > 63%), the precision could be significantly improved to 96.81% while keeping a recall rate > 95% (i.e., only 5% of the queried items would not be mapped). The machine learning approach did not lead to any improvements compared to the fuzzy matching. However, it could increase substantially the recall rate for food items without any clear equivalent in the FCTs (+7 and +20% when mapping items using their original or English-translated names). Our approaches have been implemented as R packages and are freely available from GitHub. Conclusion: This study is the first to provide automated approaches for large-scale food item mapping onto FCTs. We demonstrate that both high precision and recall can be achieved. Our solutions can be used with any FCT and do not require any programming background. These methodologies and findings are useful to any small or large nutritional study (observational as well as interventional).

Atacicept in multiple sclerosis (ATAMS): a randomised, placebo-controlled, double-blind, phase 2 trial.

BACKGROUND: Depletion of B lymphocytes is associated with suppression of inflammatory activity in multiple sclerosis. We aimed to assess the safety and efficacy of atacicept, a recombinant fusion protein that suppresses B-cell function and antibody production. METHODS: In this placebo-controlled, double-blind, 36-week, phase 2 trial (ATAMS) in Australia, Canada, Europe, and the USA, patients aged 18-60 years with relapsing multiple sclerosis were randomly assigned via an interactive voice response system in a 1:1:1:1 ratio, stratified by geographical region, to receive weekly subcutaneous injections with atacicept (25, 75, or 150 mg) or placebo. Both patients and study personnel were masked to treatment assignment. The primary endpoint was the change in mean number of gadolinium-enhancing lesions on T1-weighted MRI per patient per scan between weeks 12 and 36. Efficacy endpoints were analysed in the intention-to-treat population. Patients who completed week 36 were eligible to participate in a long-term extension study (ATAMS EXT), consisting of a double-blind phase followed by an open-label phase, for a total study time of up to 5 years. The study was terminated early after the independent data and safety monitoring board noted an increased annualised relapse rate with atacicept. The protocol was subsequently amended to include a 60-week safety follow-up, to allow treatment with approved multiple sclerosis drugs, and to change the primary endpoint to gadolinium-enhancing T1 lesions per scan during the entire double-blind period of ATAMS. Both the trial and the extension are registered with ClinicalTrials.gov, numbers NCT00642902 (ATAMS) and NCT00853762 (ATAMS EXT). FINDINGS: Between April 23, 2008, and early study termination on Sept 11, 2009, 255 patients were randomly assigned: 63 to placebo, 63 to atacicept 25 mg, 64 to 75 mg, and 65 to 150 mg. 90 (35%) patients completed the week 36 treatment visit, 26 (10%) discontinued before study termination (including one who dropped out before receiving study treatment), and 139 (55%) discontinued because of study termination. During the double-blind period of ATAMS, annualised relapse rates were higher in the atacicept groups than in the placebo group (atacicept 25 mg, 0·86, 95% CI 0·43-1·74; 75 mg, 0·79, 0·40-1·58; 150 mg, 0·98, 0·52-1·81; placebo, 0·38, 0·17-0·87). Mean numbers of gadolinium-enhancing T1 lesions per scan were similar in all groups (25 mg, 2·26, 0·97-5·27; 75 mg, 2·30, 1·08-4·92; 150 mg, 2·49, 1·18-5·27; placebo, 3·07, 1·40-6·77). Seven patients (one taking placebo and six atacicept) discontinued treatment because of adverse events. One death occurred in the placebo group. During the safety follow-up, immunoglobulin concentrations and B-cell counts returned towards predose values and annualised relapse rates in the atacicept groups decreased until they were similar to that of the placebo group INTERPRETATION: Increased clinical disease activity associated with atacicept suggests that the role of B cells and humoral immunity in multiple sclerosis is complex. For studies that explore therapeutic immunomodulation in multiple sclerosis, rigorous monitoring for negative effects on clinical and MRI outcomes is warranted. FUNDING: Merck Serono (Merck KGaA) and EMD Serono (Merck KGaA).

Convergent functional genomics of oligodendrocyte differentiation identifies multiple autoinhibitory signaling circuits.

Inadequate remyelination of brain white matter lesions has been associated with a failure of oligodendrocyte precursors to differentiate into mature, myelin-producing cells. In order to better understand which genes play a critical role in oligodendrocyte differentiation, we performed time-dependent, genome-wide gene expression studies of mouse Oli-neu cells as they differentiate into process-forming and myelin basic protein-producing cells, following treatment with three different agents. Our data indicate that different inducers activate distinct pathways that ultimately converge into the completely differentiated state, where regulated gene sets overlap maximally. In order to also gain insight into the functional role of genes that are regulated in this process, we silenced 88 of these genes using small interfering RNA and identified multiple repressors of spontaneous differentiation of Oli-neu, most of which were confirmed in rat primary oligodendrocyte precursors cells. Among these repressors were CNP, a well-known myelin constituent, and three phosphatases, each known to negatively control mitogen-activated protein kinase cascades. We show that a novel inhibitor for one of the identified genes, dual-specificity phosphatase DUSP10/MKP5, was also capable of inducing oligodendrocyte differentiation in primary oligodendrocyte precursors. Oligodendrocytic differentiation feedback loops may therefore yield pharmacological targets to treat disease related to dysfunctional myelin deposition.

Gene-based bin analysis of genome-wide association studies.

BACKGROUND: With the improvement of genotyping technologies and the exponentially growing number of available markers, case-control genome-wide association studies promise to be a key tool for investigation of complex diseases. However new analytical methods have to be developed to face the problems induced by this data scale-up, such as statistical multiple testing, data quality control and computational tractability. RESULTS: We present a novel method to analyze genome-wide association studies results. The algorithm is based on a Bayesian model that integrates genotyping errors and genomic structure dependencies. p-values are assigned to genomic regions termed bins, which are defined from a gene-biased partitioning of the genome, and the false-discovery rate is estimated. We have applied this algorithm to data coming from three genome-wide association studies of Multiple Sclerosis. CONCLUSION: The method practically overcomes the scale-up problems and permits to identify new putative regions statistically associated with the disease.

Universal false discovery rate estimation methodology for genome-wide association studies.

Genome-wide case-control association studies aim at identifying significant differential markers between sick and healthy populations. With the development of large-scale technologies allowing the genotyping of thousands of single nucleotide polymorphisms (SNPs) comes the multiple testing problem and the practical issue of selecting the most probable set of associated markers. Several False Discovery Rate (FDR) estimation methods have been developed and tuned mainly for differential gene expression studies. However they are based on hypotheses and designs that are not necessarily relevant in genetic association studies. In this article we present a universal methodology to estimate the FDR of genome-wide association results. It uses a single global probability value per SNP and is applicable in practice for any study design, using any statistic. We have benchmarked this algorithm on simulated data and shown that it outperforms previous methods in cases requiring non-parametric estimation. We exemplified the usefulness of the method by applying it to the analysis of experimental genotyping data of three Multiple Sclerosis case-control association studies.

Detecting local high-scoring segments: a first-stage approach for genome-wide association studies.

Genetic epidemiology aims at identifying biological mechanisms responsible for human diseases. Genome-wide association studies, made possible by recent improvements in genotyping technologies, are now promisingly investigated. In these studies, common first-stage strategies focus on marginal effects but lead to multiple-testing and are unable to capture the possibly complex interplay between genetic factors. We have adapted the use of the local score statistic, already successfully applied to analyse long molecular sequences. Via sum statistics, this method captures local and possible distant dependences between markers. Dedicated to genome-wide association studies, it is fast to compute, able to handle large datasets, circumvents the the multiple-testing problem and outlines a set of genomic regions (segments) for further analyses. Applied to simulated and real data, our approach outperforms classical Bonferroni and FDR corrections for multiple-testing. It is implemented in a software termed LHiSA for Local High-scoring Segments for Association and available at: http://stat.genopole.cnrs.fr/software/lhisa.

Analysis of fragments induced by simulated lattice protein folding.

The folding process of a set of 42 proteins, representative of the various folds, has been simulated by means of a Monte Carlo method on a discrete lattice, using two different potentials of mean force. Multiple compact fragments of contiguous residues are formed in the simulation, stable in composition, but not in geometry. During time, the number of fragments decreases until one final compact globular state is reached. We focused on the early steps of the folding in order to evidence the maximum number of fragments, provided they are sufficiently stable in sequence. A correlation has been established between these proto fragments and regular secondary-structure elements, whatever their nature, alpha helices or beta strands. Quantitatively, this is revealed by an overall mean one-residue quality factor of nearly 60%, which is better for proteins mainly composed of alpha helices. The correspondence between the number of fragments and the number of secondary-structure elements is of 77% and the regions separating successive fragments are mainly located in loops. Besides, hydrophobic clusters deduced from HCA correspond to fragments with an equivalent accuracy. These results suggest that folding pathways do not contain structurally static intermediate. However, since the beginning of folding, most residues that will later form one given secondary structure are kept close in space by being involved in the same fragment. This aggregation may be a way to accelerate the formation of the native state and enforces the key role played by hydrophobic residues in the formation of the fragments, thus in the folding process itself.