Correcting for non-participation bias in health surveys using record-linkage, synthetic observations and pattern mixture modelling.
Ontology highlight
ABSTRACT:
SUBMITTER: Gray L
PROVIDER: S-EPMC7188518 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:MotivationThe Illumina Infinium 450 k DNA Methylation Beadchip is a prime candidate technology for Epigenome-Wide Association Studies (EWAS). However, a difficulty associated with these beadarrays is that probes come in two different designs, characterized by widely different DNA methylation distributions and dynamic range, which may bias downstream analyses. A key statistical issue is therefore how best to adjust for the two different probe designs.ResultsHere we propose a novel model-based intra-array normalization strategy for 450 k data, called BMIQ (Beta MIxture Quantile dilation), to adjust the beta-values of type2 design probes into a statistical distribution characteristic of type1 probes. The strategy involves application of a three-state beta-mixture model to assign probes to methylation states, subsequent transformation of probabilities into quantiles and finally a methylation-dependent dilation transformation to preserve the monotonicity and continuity of the data. We validate our method on cell-line data, fresh frozen and paraffin-embedded tumour tissue samples and demonstrate that BMIQ compares favourably with two competing methods. Specifically, we show that BMIQ improves the robustness of the normalization procedure, reduces the technical variation and bias of type2 probe values and successfully eliminates the type1 enrichment bias caused by the lower dynamic range of type2 probes. BMIQ will be useful as a preprocessing step for any study using the Illumina Infinium 450 k platform.AvailabilityBMIQ is freely available from http://code.google.com/p/bmiq/.Contacta.teschendorff@ucl.ac.ukSupplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC3546795 | biostudies-literature
Project description:Studies involving the use of probabilistic record linkage are becoming increasingly common. However, the methods underpinning probabilistic record linkage are not widely taught or understood, and therefore these studies can appear to be a 'black box' research tool. In this article, we aim to describe the process of probabilistic record linkage through a simple exemplar. We first introduce the concept of deterministic linkage and contrast this with probabilistic linkage. We illustrate each step of the process using a simple exemplar and describe the data structure required to perform a probabilistic linkage. We describe the process of calculating and interpreting matched weights and how to convert matched weights into posterior probabilities of a match using Bayes theorem. We conclude this article with a brief discussion of some of the computational demands of record linkage, how you might assess the quality of your linkage algorithm, and how epidemiologists can maximize the value of their record-linked research using robust record linkage methods.
| S-EPMC5005943 | biostudies-literature
Project description:For the past 80 years, social scientists have been linking historical censuses across time to study economic and geographic mobility. In recent decades, the quantity of historical census record linkage has exploded, owing largely to the advent of new machine-readable data created by genealogical organizations. Investigators are examining economic and geographic mobility across multiple generations, but also engaging many new topics. Several analysts are exploring the effects of early-life socioeconomic conditions, environmental exposures, or natural disasters on family, health and economic outcomes in later life. Other studies exploit natural experiments to gauge the impact of policy interventions such as social welfare programs and educational reforms. The new data sources have led to a proliferation of record linkage methodology, and some widespread approaches inadvertently introduce errors that can lead to false inferences. A new generation of large-scale shared data infrastructure now in preparation will ameliorate weaknesses of current linkage methods.
| S-EPMC6200350 | biostudies-literature
Project description:MOTIVATION:Chromatin Interaction Analysis by Paired-End Tag sequencing (ChIA-PET) is an established method for detecting genome-wide looping interactions at high resolution. Current ChIA-PET analysis software packages either fail to correct for non-specific interactions due to genomic proximity or only address a fraction of the steps required for data processing. We present Mango, a complete ChIA-PET data analysis pipeline that provides statistical confidence estimates for interactions and corrects for major sources of bias including differential peak enrichment and genomic proximity. RESULTS:Comparison to the existing software packages, ChIA-PET Tool and ChiaSig revealed that Mango interactions exhibit much better agreement with high-resolution Hi-C data. Importantly, Mango executes all steps required for processing ChIA-PET datasets, whereas ChiaSig only completes 20% of the required steps. Application of Mango to multiple available ChIA-PET datasets permitted the independent rediscovery of known trends in chromatin loops including enrichment of CTCF, RAD21, SMC3 and ZNF143 at the anchor regions of interactions and strong bias for convergent CTCF motifs. AVAILABILITY AND IMPLEMENTATION:Mango is open source and distributed through github at https://github.com/dphansti/mango. CONTACT:mpsnyder@standford.edu SUPPLEMENTARY INFORMATION:Supplementary data are available at Bioinformatics online.
| S-EPMC4592333 | biostudies-literature
Project description:This paper identifies trends within and relationships between the amount of participation and the quality of contributions in three crowdsourced surveys. Participants were asked to perform a collective problem solving task that lacked any explicit incentive: they were instructed not only to respond to survey questions but also to pose new questions that they thought might-if responded to by others-predict an outcome variable of interest to them. While the three surveys had very different outcome variables, target audiences, methods of advertisement, and lengths of deployment, we found very similar patterns of collective behavior. In particular, we found that: the rate at which participants submitted new survey questions followed a heavy-tailed distribution; the distribution in the types of questions posed was similar; and many users posed non-obvious yet predictive questions. By analyzing responses to questions that contained a built-in range of valid response we found that less than 0.2% of responses lay outside of those ranges, indicating that most participants tend to respond honestly to surveys of this form, even without explicit incentives for honesty. While we did not find a significant relationship between the quantity of participation and the quality of contribution for both response submissions and question submissions, we did find several other more nuanced participant behavior patterns, which did correlate with contribution in one of the three surveys. We conclude that there exists an optimal time for users to pose questions early on in their participation, but only after they have submitted a few responses to other questions. This suggests that future crowdsourced surveys may attract more predictive questions by prompting users to pose new questions at specific times during their participation and limiting question submission at non-optimal times.
| S-EPMC4383627 | biostudies-literature
Project description:INTRODUCTION:Decreasing participation levels in health surveys pose a threat to the validity of estimates intended to be representative of their target population. If participants and non-participants differ systematically, the results may be biased. The application of traditional non-response adjustment methods, such as weighting, can fail to correct for such biases, as estimates are typically based on the sociodemographic information available. Therefore, a dedicated methodology to infer on non-participants offers advancement by employing survey data linked to administrative health records, with reference to data on the general population. We aim to validate such a methodology in a register-based setting, where individual-level data on participants and non-participants are available, taking alcohol consumption estimation as the exemplar focus. METHODS AND ANALYSIS:We made use of the selected sample of the Health 2000 survey conducted in Finland and a separate register-based sample of the contemporaneous population, with follow-up until 2012. Finland has nationally representative administrative and health registers available for individual-level record linkage to the Health 2000 survey participants and invited non-participants, and the population sample. By comparing the population sample and the participants, synthetic observations representing the non-participants may be generated, as per the developed methodology. We can compare the distribution of the synthetic non-participants with the true distribution from the register data. Multiple imputation was then used to estimate alcohol consumption based on both the actual and synthetic data for non-participants, and the estimates can be compared to evaluate the methodology's performance. ETHICS AND DISSEMINATION:Ethical approval and access to the Health 2000 survey data and data from administrative and health registers have been given by the Health 2000 Scientific Advisory Board, Statistics Finland and the National Institute for Health and Welfare. The outputs will include two publications in public health and statistical methodology journals and conference presentations.
| S-EPMC6500270 | biostudies-other
Project description:MotivationSystematic bias in mass measurement adversely affects data quality and negates the advantages of high precision instruments.ResultsWe introduce the mzRefinery tool for calibration of mass spectrometry data files. Using confident peptide spectrum matches, three different calibration methods are explored and the optimal transform function is chosen. After calibration, systematic bias is removed and the mass measurement errors are centered at 0?ppm. Because it is part of the ProteoWizard package, mzRefinery can read and write a wide variety of file formats.Availability and implementationThe mzRefinery tool is part of msConvert, available with the ProteoWizard open source package at http://proteowizard.sourceforge.net/Contactsamuel.payne@pnnl.gov.Supplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC4653383 | biostudies-literature
Project description:The arrival of submillimeter ultra high-field fMRI makes it possible to compare activation profiles across cortical layers. However, the blood oxygenation level dependent (BOLD) signal measured by gradient echo (GE) fMRI is biased toward superficial layers of the cortex, which is a serious confound for laminar analysis. Several univariate and multivariate analysis methods have been proposed to correct this bias. We compare these methods using computational simulations of 7T fMRI data from regions of interest (ROI) during a visual attention paradigm. We also tested the methods on a pilot dataset of human 7T fMRI data. The simulations show that two methods-the ratio of ROI means across conditions and a novel application of Deming regression-offer the most robust correction for superficial bias. Deming regression has the additional advantage that it does not require that the conditions differ in their mean activation over voxels within an ROI. When applied to the pilot dataset, we observed strikingly different layer profiles when different attention metrics were used, but were unable to discern any differences in laminar attention across layers when Deming regression or ROI ratio was applied. Our simulations demonstrates that accurate correction of superficial bias is crucial to avoid drawing erroneous conclusions from laminar analyses of GE fMRI data, and this is affirmed by the results from our pilot 7T fMRI data.
| S-EPMC8494131 | biostudies-literature
Project description:Errors in botanical surveying are a common problem. The presence of a species is easily overlooked, leading to false-absences; while misidentifications and other mistakes lead to false-positive observations. While it is common knowledge that these errors occur, there are few data that can be used to quantify and describe these errors. Here we characterise false-positive errors for a controlled set of surveys conducted as part of a field identification test of botanical skill. Surveys were conducted at sites with a verified list of vascular plant species. The candidates were asked to list all the species they could identify in a defined botanically rich area. They were told beforehand that their final score would be the sum of the correct species they listed, but false-positive errors counted against their overall grade. The number of errors varied considerably between people, some people create a high proportion of false-positive errors, but these are scattered across all skill levels. Therefore, a person's ability to correctly identify a large number of species is not a safeguard against the generation of false-positive errors. There was no phylogenetic pattern to falsely observed species; however, rare species are more likely to be false-positive as are species from species rich genera. Raising the threshold for the acceptance of an observation reduced false-positive observations dramatically, but at the expense of more false negative errors. False-positive errors are higher in field surveying of plants than many people may appreciate. Greater stringency is required before accepting species as present at a site, particularly for rare species. Combining multiple surveys resolves the problem, but requires a considerable increase in effort to achieve the same sensitivity as a single survey. Therefore, other methods should be used to raise the threshold for the acceptance of a species. For example, digital data input systems that can verify, feedback and inform the user are likely to reduce false-positive errors significantly.
| S-EPMC5437866 | biostudies-literature
Project description:Although efficient methods exist to assemble synthetic oligonucleotides into genes and genomes, these suffer from the presence of 1-3 random errors/kb of DNA. Here, we introduce a new method termed consensus shuffling and demonstrate its use to significantly reduce random errors in synthetic DNA. In this method, errors are revealed as mismatches by re-hybridization of the population. The DNA is fragmented, and mismatched fragments are removed upon binding to an immobilized mismatch binding protein (MutS). PCR assembly of the remaining fragments yields a new population of full-length sequences enriched for the consensus sequence of the input population. We show that two iterations of consensus shuffling improved a population of synthetic green fluorescent protein (GFPuv) clones from approximately 60 to >90% fluorescent, and decreased errors 3.5- to 4.3-fold to final values of approximately 1 error per 3500 bp. In addition, two iterations of consensus shuffling corrected a population of GFPuv clones where all members were non-functional, to a population where 82% of clones were fluorescent. Consensus shuffling should facilitate the rapid and accurate synthesis of long DNA sequences.
| S-EPMC1072806 | biostudies-literature