A method for detecting outliers in linear-circular non-parametric regression.
Ontology highlight
ABSTRACT:
SUBMITTER: Sert S
PROVIDER: S-EPMC10259788 | biostudies-literature | 2023
REPOSITORIES: biostudies-literature
Publications
A method for detecting outliers in linear-circular non-parametric regression.
Sert Sümeyra S Kardiyen Filiz F
PloS one 20230612 6
This study proposes a robust outlier detection method based on the circular median for non-parametric linear-circular regression in case the response variable includes outlier(s) and the residuals are Wrapped-Cauchy distributed. Nadaraya-Watson and local linear regression methods were employed to obtain non-parametric regression fits. The proposed method's performance was investigated by using a real dataset and a comprehensive simulation study with different sample sizes, contamination, and het ...[more]
Similar Datasets
Project description:BackgroundDysbiosis of the human gut microbiome is defined as a maladaptive or clinically relevant deviation of the community profile from the healthy or normal state. Dysbiosis has been implicated in an extensive set of metabolic, auto-immune, and infectious diseases, and yet there is substantial inter-individual variation in microbiome composition even within body sites of healthy humans. An individual's microbiome varies over time in a high-dimensional space to form their personal microbiome cloud. This cloud may or may not be similar to that of other people, both in terms of the average microbiome profile (conformity) and the diameter of the cloud (stability). However, there is currently no robust non-parametric test that determines whether a patient's microbiome cloud is an outlier with respect to a reference group of healthy individuals with widely varying microbiome profiles.MethodsHere, we propose a test for outliers' detection in the human gut microbiome that accounts for the wide range of microbiome phenotypes observed in a typical set of healthy individuals and for intra-individual temporal variation. Our robust nonparametric outlier detection test, the CLOUD test, performs two assessments of a patient's microbiome health: conformity, the extent to which the patient's microbiome cloud is ecologically similar to a subset of healthy subjects; and stability, which compares the cloud diameter of a patient to those of healthy subjects. The CLOUD test is based on locally linear embedded ecological distances, allowing it to account for widely varying microbiome compositions among reference individuals. It also leverages temporal variability within patients and reference individuals to increase the robustness of the test.ResultsWe describe the CLOUD test, and we apply it to one novel and two previously published cohorts of patients receiving fecal microbiota transplantation for recurrent Clostridium difficile colitis, as well as to two known healthy cohorts, demonstrating high concordance of the CLOUD conformity and stability indices with clinical outcomes.ConclusionsAlthough the CLOUD test is not, on its own, a test for clinical dysbiosis, it nonetheless provides a framework for outlier testing that could be incorporated into evaluation of suspected dysbiosis, which may play a role in diagnosis and prognosis of numerous pediatric and adult diseases.
| S-EPMC6080375 | biostudies-literature
Project description:In genome-enabled prediction, parametric, semi-parametric, and non-parametric regression models have been used. This study assessed the predictive ability of linear and non-linear models using dense molecular markers. The linear models were linear on marker effects and included the Bayesian LASSO, Bayesian ridge regression, Bayes A, and Bayes B. The non-linear models (this refers to non-linearity on markers) were reproducing kernel Hilbert space (RKHS) regression, Bayesian regularized neural networks (BRNN), and radial basis function neural networks (RBFNN). These statistical models were compared using 306 elite wheat lines from CIMMYT genotyped with 1717 diversity array technology (DArT) markers and two traits, days to heading (DTH) and grain yield (GY), measured in each of 12 environments. It was found that the three non-linear models had better overall prediction accuracy than the linear regression specification. Results showed a consistent superiority of RKHS and RBFNN over the Bayesian LASSO, Bayesian ridge regression, Bayes A, and Bayes B models.
| S-EPMC3516481 | biostudies-literature
Project description:The ordinary linear regression method is limited to bivariate data because it is based on the Cartesian representation y = f(x). Using the chain rule, we transform the method to the parametric representation (x(t), y(t)) and obtain a linear regression framework in which the weighted average is used as a parameter for a multivariate linear relation for a set of linearly related variable vectors (LRVVs). We confirm the proposed approach by a Monte Carlo simulation, where the minimum coefficient of variation for error (CVE) provides the optimal weights when forming a weighted average of LRVVs. Then, we describe a parametric linear regression (PLR) algorithm in which the Moore-Penrose pseudoinverse is used to estimate measurement error regression (MER) parameters individually for the given variable vectors. We demonstrate that MER parameters from the PLR and nonlinear ODRPACK methods are quite similar for a wide range of reliability ratios, but ODRPACK is formulated only for bivariate data. We identify scale invariant quantities for the PLR and weighted orthogonal regression (WOR) methods and their correspondences with the partitioned residual effects between the variable vectors. Thus, the specification of an error model for the data is essential for MER and we discuss the use of Monte Carlo methods for estimating the distributions and confidence intervals for MER slope and correlation coefficient. We distinguish between elementary covariance for the y = f(x) representation and covariance vector for the (x(t), y(t)) representation. We also discuss the multivariate generalization of the Pearson correlation as the contraction between Cartesian polyad alignment tensors for the LRVVs and weighted average. Finally, we demonstrate the use of multidimensional PLR in estimating the MER parameters for replicate RNA-Seq data and quadratic regression for estimating the parameters of the conical dispersion of read count data about the MER line.
| S-EPMC8782541 | biostudies-literature
Project description:MotivationGene-environment (GxE) interactions are one of the least studied aspects of the genetic architecture of human traits and diseases. The environment of an individual is inherently high dimensional, evolves through time and can be expensive and time consuming to measure. The UK Biobank study, with all 500 000 participants having undergone an extensive baseline questionnaire, represents a unique opportunity to assess GxE heritability for many traits and diseases in a well powered setting.ResultsWe have developed a randomized Haseman-Elston non-linear regression method applicable when many environmental variables have been measured on each individual. The method (GPLEMMA) simultaneously estimates a linear environmental score (ES) and its GxE heritability. We compare the method via simulation to a whole-genome regression approach (LEMMA) for estimating GxE heritability. We show that GPLEMMA is more computationally efficient than LEMMA on large datasets, and produces results highly correlated with those from LEMMA when applied to simulated data and real data from the UK Biobank.Availability and implementationSoftware implementing the GPLEMMA method is available from https://jmarchini.org/gplemma/.Supplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC8023682 | biostudies-literature
Project description:Non-linear Mendelian randomization is an extension to standard Mendelian randomization to explore the shape of the causal relationship between an exposure and outcome using an instrumental variable. A stratification approach to non-linear Mendelian randomization divides the population into strata and calculates separate instrumental variable estimates in each stratum. However, the standard implementation of stratification, referred to as the residual method, relies on strong parametric assumptions of linearity and homogeneity between the instrument and the exposure to form the strata. If these stratification assumptions are violated, the instrumental variable assumptions may be violated in the strata even if they are satisfied in the population, resulting in misleading estimates. We propose a new stratification method, referred to as the doubly-ranked method, that does not require strict parametric assumptions to create strata with different average levels of the exposure such that the instrumental variable assumptions are satisfied within the strata. Our simulation study indicates that the doubly-ranked method can obtain unbiased stratum-specific estimates and appropriate coverage rates even when the effect of the instrument on the exposure is non-linear or heterogeneous. Moreover, it can also provide unbiased estimates when the exposure is coarsened (that is, rounded, binned into categories, or truncated), a scenario that is common in applied practice and leads to substantial bias in the residual method. We applied the proposed doubly-ranked method to investigate the effect of alcohol intake on systolic blood pressure, and found evidence of a positive effect of alcohol intake, particularly at higher levels of alcohol consumption.
| S-EPMC10343089 | biostudies-literature
Project description:We describe a novel approach for evaluating SNP genotypes of a genome-wide association scan to identify "ethnic outlier" subjects whose ethnicity is different or admixed compared to most other subjects in the genotyped sample set. Each ethnic outlier is detected by counting a genomic excess of "rare" heterozygotes and/or homozygotes whose frequencies are low (<1%) within genotypes of the sample set being evaluated. This method also enables simple and striking visualization of non-Caucasian chromosomal DNA segments interspersed within the chromosomes of ethnically admixed individuals. We show that this visualization of the mosaic structure of admixed human chromosomes gives results similar to another visualization method (SABER) but with much less computational time and burden. We also show that other methods for detecting ethnic outliers are enhanced by evaluating only genomic regions of visualized admixture rather than diluting outlier ancestry by evaluating the entire genome considered in aggregate. We have validated our method in the Wellcome Trust Case Control Consortium (WTCCC) study of 17,000 subjects as well as in HapMap subjects and simulated outliers of known ethnicity and admixture. The method's ability to precisely delineate chromosomal segments of non-Caucasian ethnicity has enabled us to demonstrate previously unreported non-Caucasian admixture in two HapMap Caucasian parents and in a number of WTCCC subjects. Its sensitive detection of ethnic outliers and simple visual discrimination of discrete chromosomal segments of different ethnicity implies that this method of rare heterozygotes and homozygotes (RHH) is likely to have diverse and important applications in humans and other species.
| S-EPMC2883336 | biostudies-literature
Project description:Causal prediction has become a popular tool for neuroscience applications, as it allows the study of relationships between different brain areas during rest, cognitive tasks or brain disorders. We propose a nonparametric approach for the estimation of nonlinear causal prediction for multivariate time series. In the proposed estimator, C NPMR , Autoregressive modeling is replaced by Nonparametric Multiplicative Regression (NPMR). NPMR quantifies interactions between a response variable (effect) and a set of predictor variables (cause); here, we modified NPMR for model prediction. We also demonstrate how a particular measure, the sensitivity Q, could be used to reveal the structure of the underlying causal relationships. We apply C NPMR on artificial data with known ground truth (5 datasets), as well as physiological data (2 datasets). C NPMR correctly identifies both linear and nonlinear causal connections that are present in the artificial data, as well as physiologically relevant connectivity in the real data, and does not seem to be affected by filtering. The Sensitivity measure also provides useful information about the latent connectivity.The proposed estimator addresses many of the limitations of linear Granger causality and other nonlinear causality estimators. C NPMR is compared with pairwise and conditional Granger causality (linear) and Kernel-Granger causality (nonlinear). The proposed estimator can be applied to pairwise or multivariate estimations without any modifications to the main method. Its nonpametric nature, its ability to capture nonlinear relationships and its robustness to filtering make it appealing for a number of applications.
| S-EPMC4905976 | biostudies-literature
Project description:In the current study we introduce statistical methods based on trigonometry, to infer the shape of a (non)-linear bivariate genetic relationship. We do this based on a series of piecemeal GWASs of segments of a target (continuous) trait distribution, and the genetic correlations between those GWASs and a second trait. Simulations confirm that we are able to retrieve the shape of the relationship given certain assumptions about the nature of the relationship between the traits. We applied the method to the genetic relationship between BMI, sleep duration, and height, and psychiatric disorders (ADHD, anorexia nervosa, and depression) using data from approximately 450K individuals from UK Biobank. In the relationship between BMI and psychiatric traits, we found that the expected value of depression is a nonlinear function of BMI i.e. there is a nonlinear genetic relationship between both traits. We observed similar findings for the genetic relationship between BMI and anorexia, sleep duration and depression, and sleep duration and ADHD. We observed no underlying nonlinearity in the genetic relationship between height and psychiatric traits. Using a novel statistical approach, we show that nonlinear genetic relationships between traits are detectable and genetic associations as quantified using global estimators like genetic correlations are not informative about underlying complexities in these relationships. Our findings challenge assumptions of linearity in genetic epidemiology and suggest that bivariate genetic associations are not uniform across the phenotypic spectrum, which may have implications for the development of targeted interventions.
| S-EPMC11844478 | biostudies-literature
Project description:PurposeTo present a novel non-parametric algorithm for detecting the position of the human eye limbus in three dimensions and a new dynamic method for measuring the full 360° visible iris boundary known as white-to-white distance along the eye horizontal line.MethodsThe study included 88 participants aged 23 to 65 years (37.7±9.7), 47 females and 41 males. Clinical characteristics, height data and the apex coordinates and 1024×1280 pixel digital images of the eyes were taken by an Eye Surface Profiler and processed by custom-built MATLAB codes. A dynamic light intensity frequency based white-to-white detection process and a novel three-dimensional method for limbus detection is presented.ResultsEvidence of significant differences (p<0.001) between nasal-temporal and superior-inferior white-to-white distances in both right and left eyes were found (nasal-temporal direction; 11.74±0.42 mm in right eyes and 11.82±0.47 mm in left eyes & superior-inferior direction; 11.52±0.45 mm in right eyes and 11.55±0.46 mm in left eyes). Average limbus nasal-temporal diameters were 13.64±0.55 mm for right eyes, and 13.74±0.40 mm for left eyes, however the superior-inferior diameters were 13.65±0.54 mm, 13.75±0.38 mm for right and left eyes, respectively. No significant difference in limbus contours has been observed either between the nasal-temporal direction (p = 0.91) and the superior-inferior direction (p = 0.83) or between the right (p = 0.18) and left eyes (p = 0.16). Evidence of tilt towards the nasal-temporal side in the three-dimensional shape of the limbus was found. The right eyes mean limbus contour tilt around the X-axis was -0.3±1.35° however, their mean limbus contour tilt around the Y-axis was 1.76±0.9°. Likewise, the left eyes mean limbus contour tilt around the X-axis was 0.77±1.25° and the mean limbus contour tilt around the Y-axis was -1.54±0.89°.ConclusionsThe white-to-white distance in the human eye is significantly larger in the nasal-temporal direction than in the superior-inferior direction. The human limbus diameter was found not to vary significantly in these directions. The 3D measures show that the limbus contour does not lay in one plane and tends to be higher on the nasal-inferior side of the eye.
| S-EPMC6261009 | biostudies-literature
Project description:Variable selection is fundamental to high dimensional statistical modeling, and many approaches have been proposed. However, existing variable selection methods do not perform well in presence of outliers in response variable or/and covariates. In order to ensure a high probability of correct selection and efficient parameter estimation, we investigate a robust variable selection method based on a modified Huber's function with an exponential squared loss tail. We also prove that the proposed method has oracle properties. Furthermore, we carry out simulation studies to evaluate the performance of the proposed method for both p<n and p>n. Our simulation results indicate that the proposed method is efficient and robust against outliers and heavy-tailed distributions. Finally, a real dataset from an air pollution mortality study is used to illustrate the proposed method.
| S-EPMC9559330 | biostudies-literature