Abstract Epidemiological studies have suggested the joint effect of simultaneous exposures to multiple environments on disease risk. However, how environmental mixtures as a whole jointly modify genetic effect on disease risk is still largely unknown. Given the importance of gene-environment (G×E) interactions on many complex diseases, rigorously assessing the interaction effect between genes and environmental mixtures as a whole could shed novel insights into the etiology of complex diseases. For this purpose, we propose a generalized partial linear varying multi-index coefficient model (GPLVMICM) to capture the genetic effect on disease risk modulated by multiple environments as a whole. GPLVMICM is semiparametric in nature which allows different index loading parameters in different index functions. We estimate the parametric parameters by a profile procedure, and the nonparametric index functions by a B-spline backfitted kernel method. Under some regularity conditions, the proposed parametric and nonparametric estimators are shown to be consistent and asymptotically normal. We propose a generalized likelihood ratio (GLR) test to rigorously assess the linearity of the interaction effect between multiple environments and a gene, while apply a parametric likelihood test to detect linear G×E interaction effect. The finite sample performance of the proposed method is examined through simulation studies and is further illustrated through a real data analysis.
{"title":"Generalized partial linear varying multi-index coefficient model for gene-environment interactions","authors":"Xu Liu, Bin Gao, Yuehua Cui","doi":"10.1515/sagmb-2016-0045","DOIUrl":"https://doi.org/10.1515/sagmb-2016-0045","url":null,"abstract":"Abstract Epidemiological studies have suggested the joint effect of simultaneous exposures to multiple environments on disease risk. However, how environmental mixtures as a whole jointly modify genetic effect on disease risk is still largely unknown. Given the importance of gene-environment (G×E) interactions on many complex diseases, rigorously assessing the interaction effect between genes and environmental mixtures as a whole could shed novel insights into the etiology of complex diseases. For this purpose, we propose a generalized partial linear varying multi-index coefficient model (GPLVMICM) to capture the genetic effect on disease risk modulated by multiple environments as a whole. GPLVMICM is semiparametric in nature which allows different index loading parameters in different index functions. We estimate the parametric parameters by a profile procedure, and the nonparametric index functions by a B-spline backfitted kernel method. Under some regularity conditions, the proposed parametric and nonparametric estimators are shown to be consistent and asymptotically normal. We propose a generalized likelihood ratio (GLR) test to rigorously assess the linearity of the interaction effect between multiple environments and a gene, while apply a parametric likelihood test to detect linear G×E interaction effect. The finite sample performance of the proposed method is examined through simulation studies and is further illustrated through a real data analysis.","PeriodicalId":49477,"journal":{"name":"Statistical Applications in Genetics and Molecular Biology","volume":"16 1","pages":"59 - 74"},"PeriodicalIF":0.9,"publicationDate":"2017-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1515/sagmb-2016-0045","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44875724","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract Microbiomes, populations of microscopic organisms, have been found to be related to human health and it is expected further investigations will lead to novel perspectives of disease. The data used to analyze microbiomes is one of the newest types (the result of high-throughput technology) and the means to analyze these data is still rapidly evolving. One of the distributions that have been introduced into the microbiome literature, the Dirichlet-Multinomial, has received considerable attention. We extend this distribution’s use uncover compositional relationships between organisms at a taxonomic level. We apply our new method in two real microbiome data sets: one from human nasal passages and another from human stool samples.
{"title":"Statistical models and computational algorithms for discovering relationships in microbiome data","authors":"M. Shaikh, J. Beyene","doi":"10.1515/sagmb-2015-0096","DOIUrl":"https://doi.org/10.1515/sagmb-2015-0096","url":null,"abstract":"Abstract Microbiomes, populations of microscopic organisms, have been found to be related to human health and it is expected further investigations will lead to novel perspectives of disease. The data used to analyze microbiomes is one of the newest types (the result of high-throughput technology) and the means to analyze these data is still rapidly evolving. One of the distributions that have been introduced into the microbiome literature, the Dirichlet-Multinomial, has received considerable attention. We extend this distribution’s use uncover compositional relationships between organisms at a taxonomic level. We apply our new method in two real microbiome data sets: one from human nasal passages and another from human stool samples.","PeriodicalId":49477,"journal":{"name":"Statistical Applications in Genetics and Molecular Biology","volume":"16 1","pages":"1 - 12"},"PeriodicalIF":0.9,"publicationDate":"2017-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1515/sagmb-2015-0096","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67003111","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract As RNA-seq rapidly develops and costs continually decrease, the quantity and frequency of samples being sequenced will grow exponentially. With proteomic investigations becoming more multivariate and quantitative, determining a study’s optimal sample size is now a vital step in experimental design. Current methods for calculating a study’s required sample size are mostly based on the hypothesis testing framework, which assumes each gene count can be modeled through Poisson or negative binomial distributions; however, these methods are limited when it comes to accommodating covariates. To address this limitation, we propose an estimating procedure based on the generalized linear model. This easy-to-use method constructs a representative exemplary dataset and estimates the conditional power, all without requiring complicated mathematical approximations or formulas. Even more attractive, the downstream analysis can be performed with current R/Bioconductor packages. To demonstrate the practicability and efficiency of this method, we apply it to three real-world studies, and introduce our on-line calculator developed to determine the optimal sample size for a RNA-seq study.
{"title":"Sample size calculation based on generalized linear models for differential expression analysis in RNA-seq data","authors":"Chung-I Li, Y. Shyr","doi":"10.1515/sagmb-2016-0008","DOIUrl":"https://doi.org/10.1515/sagmb-2016-0008","url":null,"abstract":"Abstract As RNA-seq rapidly develops and costs continually decrease, the quantity and frequency of samples being sequenced will grow exponentially. With proteomic investigations becoming more multivariate and quantitative, determining a study’s optimal sample size is now a vital step in experimental design. Current methods for calculating a study’s required sample size are mostly based on the hypothesis testing framework, which assumes each gene count can be modeled through Poisson or negative binomial distributions; however, these methods are limited when it comes to accommodating covariates. To address this limitation, we propose an estimating procedure based on the generalized linear model. This easy-to-use method constructs a representative exemplary dataset and estimates the conditional power, all without requiring complicated mathematical approximations or formulas. Even more attractive, the downstream analysis can be performed with current R/Bioconductor packages. To demonstrate the practicability and efficiency of this method, we apply it to three real-world studies, and introduce our on-line calculator developed to determine the optimal sample size for a RNA-seq study.","PeriodicalId":49477,"journal":{"name":"Statistical Applications in Genetics and Molecular Biology","volume":"30 1","pages":"491 - 505"},"PeriodicalIF":0.9,"publicationDate":"2016-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1515/sagmb-2016-0008","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67003125","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract As part of the data processing of high-throughput-sequencing experiments count data are produced representing the amount of reads that map to specific genomic regions. Count data also arise in mass spectrometric experiments for the detection of protein-protein interactions. For evaluating new computational methods for the analysis of sequencing count data or spectral count data from proteomics experiments artificial count data is thus required. Although, some methods for the generation of artificial sequencing count data have been proposed, all of them simulate single sequencing runs, omitting thus the correlation structure between the individual genomic features, or they are limited to specific structures. We propose to draw correlated data from the multivariate normal distribution and round these continuous data in order to obtain discrete counts. In our approach, the required distribution parameters can either be constructed in different ways or estimated from real count data. Because rounding affects the correlation structure we evaluate the use of shrinkage estimators that have already been used in the context of artificial expression data from DNA microarrays. Our approach turned out to be useful for the simulation of counts for defined subsets of features such as individual pathways or GO categories.
{"title":"A simulation framework for correlated count data of features subsets in high-throughput sequencing or proteomics experiments","authors":"Jochen Kruppa, F. Kramer, T. Beissbarth, K. Jung","doi":"10.1515/sagmb-2015-0082","DOIUrl":"https://doi.org/10.1515/sagmb-2015-0082","url":null,"abstract":"Abstract As part of the data processing of high-throughput-sequencing experiments count data are produced representing the amount of reads that map to specific genomic regions. Count data also arise in mass spectrometric experiments for the detection of protein-protein interactions. For evaluating new computational methods for the analysis of sequencing count data or spectral count data from proteomics experiments artificial count data is thus required. Although, some methods for the generation of artificial sequencing count data have been proposed, all of them simulate single sequencing runs, omitting thus the correlation structure between the individual genomic features, or they are limited to specific structures. We propose to draw correlated data from the multivariate normal distribution and round these continuous data in order to obtain discrete counts. In our approach, the required distribution parameters can either be constructed in different ways or estimated from real count data. Because rounding affects the correlation structure we evaluate the use of shrinkage estimators that have already been used in the context of artificial expression data from DNA microarrays. Our approach turned out to be useful for the simulation of counts for defined subsets of features such as individual pathways or GO categories.","PeriodicalId":49477,"journal":{"name":"Statistical Applications in Genetics and Molecular Biology","volume":"15 1","pages":"401 - 414"},"PeriodicalIF":0.9,"publicationDate":"2016-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1515/sagmb-2015-0082","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67003073","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A. Kakourou, W. Vach, S. Nicolardi, Y. V. D. van der Burgt, B. Mertens
Abstract Mass spectrometry based clinical proteomics has emerged as a powerful tool for high-throughput protein profiling and biomarker discovery. Recent improvements in mass spectrometry technology have boosted the potential of proteomic studies in biomedical research. However, the complexity of the proteomic expression introduces new statistical challenges in summarizing and analyzing the acquired data. Statistical methods for optimally processing proteomic data are currently a growing field of research. In this paper we present simple, yet appropriate methods to preprocess, summarize and analyze high-throughput MALDI-FTICR mass spectrometry data, collected in a case-control fashion, while dealing with the statistical challenges that accompany such data. The known statistical properties of the isotopic distribution of the peptide molecules are used to preprocess the spectra and translate the proteomic expression into a condensed data set. Information on either the intensity level or the shape of the identified isotopic clusters is used to derive summary measures on which diagnostic rules for disease status allocation will be based. Results indicate that both the shape of the identified isotopic clusters and the overall intensity level carry information on the class outcome and can be used to predict the presence or absence of the disease.
{"title":"Accounting for isotopic clustering in Fourier transform mass spectrometry data analysis for clinical diagnostic studies","authors":"A. Kakourou, W. Vach, S. Nicolardi, Y. V. D. van der Burgt, B. Mertens","doi":"10.1515/sagmb-2016-0005","DOIUrl":"https://doi.org/10.1515/sagmb-2016-0005","url":null,"abstract":"Abstract Mass spectrometry based clinical proteomics has emerged as a powerful tool for high-throughput protein profiling and biomarker discovery. Recent improvements in mass spectrometry technology have boosted the potential of proteomic studies in biomedical research. However, the complexity of the proteomic expression introduces new statistical challenges in summarizing and analyzing the acquired data. Statistical methods for optimally processing proteomic data are currently a growing field of research. In this paper we present simple, yet appropriate methods to preprocess, summarize and analyze high-throughput MALDI-FTICR mass spectrometry data, collected in a case-control fashion, while dealing with the statistical challenges that accompany such data. The known statistical properties of the isotopic distribution of the peptide molecules are used to preprocess the spectra and translate the proteomic expression into a condensed data set. Information on either the intensity level or the shape of the identified isotopic clusters is used to derive summary measures on which diagnostic rules for disease status allocation will be based. Results indicate that both the shape of the identified isotopic clusters and the overall intensity level carry information on the class outcome and can be used to predict the presence or absence of the disease.","PeriodicalId":49477,"journal":{"name":"Statistical Applications in Genetics and Molecular Biology","volume":"15 1","pages":"415 - 430"},"PeriodicalIF":0.9,"publicationDate":"2016-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1515/sagmb-2016-0005","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67003117","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract Factorial Gaussian graphical Models (fGGMs) have recently been proposed for inferring dynamic gene regulatory networks from genomic high-throughput data. In the search for true regulatory relationships amongst the vast space of possible networks, these models allow the imposition of certain restrictions on the dynamic nature of these relationships, such as Markov dependencies of low order – some entries of the precision matrix are a priori zeros – or equal dependency strengths across time lags – some entries of the precision matrix are assumed to be equal. The precision matrix is then estimated by l1-penalized maximum likelihood, imposing a further constraint on the absolute value of its entries, which results in sparse networks. Selecting the optimal sparsity level is a major challenge for this type of approaches. In this paper, we evaluate the performance of a number of model selection criteria for fGGMs by means of two simulated regulatory networks from realistic biological processes. The analysis reveals a good performance of fGGMs in comparison with other methods for inferring dynamic networks and of the KLCV criterion in particular for model selection. Finally, we present an application on a high-resolution time-course microarray data from the Neisseria meningitidis bacterium, a causative agent of life-threatening infections such as meningitis. The methodology described in this paper is implemented in the R package sglasso, freely available at CRAN, http://CRAN.R-project.org/package=sglasso.
{"title":"Model selection for factorial Gaussian graphical models with an application to dynamic regulatory networks","authors":"V. Vinciotti, L. Augugliaro, A. Abbruzzo, E. Wit","doi":"10.1515/sagmb-2014-0075","DOIUrl":"https://doi.org/10.1515/sagmb-2014-0075","url":null,"abstract":"Abstract Factorial Gaussian graphical Models (fGGMs) have recently been proposed for inferring dynamic gene regulatory networks from genomic high-throughput data. In the search for true regulatory relationships amongst the vast space of possible networks, these models allow the imposition of certain restrictions on the dynamic nature of these relationships, such as Markov dependencies of low order – some entries of the precision matrix are a priori zeros – or equal dependency strengths across time lags – some entries of the precision matrix are assumed to be equal. The precision matrix is then estimated by l1-penalized maximum likelihood, imposing a further constraint on the absolute value of its entries, which results in sparse networks. Selecting the optimal sparsity level is a major challenge for this type of approaches. In this paper, we evaluate the performance of a number of model selection criteria for fGGMs by means of two simulated regulatory networks from realistic biological processes. The analysis reveals a good performance of fGGMs in comparison with other methods for inferring dynamic networks and of the KLCV criterion in particular for model selection. Finally, we present an application on a high-resolution time-course microarray data from the Neisseria meningitidis bacterium, a causative agent of life-threatening infections such as meningitis. The methodology described in this paper is implemented in the R package sglasso, freely available at CRAN, http://CRAN.R-project.org/package=sglasso.","PeriodicalId":49477,"journal":{"name":"Statistical Applications in Genetics and Molecular Biology","volume":"15 1","pages":"193 - 212"},"PeriodicalIF":0.9,"publicationDate":"2016-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1515/sagmb-2014-0075","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67002224","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract When testing for differentially expressed genes between more than two groups, the groups are often defined by dose levels in dose-response experiments or ordinal phenotypes, such as disease stages. We discuss the potential of a new approach that uses the levels’ ordering without making any structural assumptions, such as monotonicity, by testing for zero variance components in a mixed models framework. Since the mixed effects model approach borrows strength across doses/levels, the test proposed can also be applied when the number of dose levels/phenotypes is large and/or the number of subjects per group is small. We illustrate the new test in simulation studies and on several publicly available datasets and compare it to alternative testing procedures. All tests considered are implemented in R and are publicly available. The new approach offers a very fast and powerful way to test for differentially expressed genes between ordered groups without making restrictive assumptions with respect to the true relationship between factor levels and response.
{"title":"Testing differentially expressed genes in dose-response studies and with ordinal phenotypes","authors":"E. Sweeney, C. Crainiceanu, J. Gertheiss","doi":"10.1515/sagmb-2015-0091","DOIUrl":"https://doi.org/10.1515/sagmb-2015-0091","url":null,"abstract":"Abstract When testing for differentially expressed genes between more than two groups, the groups are often defined by dose levels in dose-response experiments or ordinal phenotypes, such as disease stages. We discuss the potential of a new approach that uses the levels’ ordering without making any structural assumptions, such as monotonicity, by testing for zero variance components in a mixed models framework. Since the mixed effects model approach borrows strength across doses/levels, the test proposed can also be applied when the number of dose levels/phenotypes is large and/or the number of subjects per group is small. We illustrate the new test in simulation studies and on several publicly available datasets and compare it to alternative testing procedures. All tests considered are implemented in R and are publicly available. The new approach offers a very fast and powerful way to test for differentially expressed genes between ordered groups without making restrictive assumptions with respect to the true relationship between factor levels and response.","PeriodicalId":49477,"journal":{"name":"Statistical Applications in Genetics and Molecular Biology","volume":"55 1","pages":"213 - 235"},"PeriodicalIF":0.9,"publicationDate":"2016-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1515/sagmb-2015-0091","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67003076","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
D. Ryu, Hongyang Xu, Varghese George, S. Su, Xiaoling Wang, Huidong Shi, R. Podolsky
Abstract Differential methylation of regulatory elements is critical in epigenetic researches and can be statistically tested. We developed a new statistical test, the generalized integrated functional test (GIFT), that tests for regional differences in methylation based on the methylation percent at each CpG site within a genomic region. The GIFT uses estimated subject-specific profiles with smoothing methods, specifically wavelet smoothing, and calculates an ANOVA-like test to compare the average profile of groups. In this way, possibly correlated CpG sites within the regulatory region are compared all together. Simulations and analyses of data obtained from patients with chronic lymphocytic leukemia indicate that GIFT has good statistical properties and is able to identify promising genomic regions. Further, GIFT is likely to work with multiple different types of experiments since different smoothing methods can be used to estimate the profiles of data without noise. Matlab code for GIFT and sample data are available at http://www.augusta.edu/mcg/biostatepi/people/software/gift.html.
{"title":"Differential methylation tests of regulatory regions","authors":"D. Ryu, Hongyang Xu, Varghese George, S. Su, Xiaoling Wang, Huidong Shi, R. Podolsky","doi":"10.1515/sagmb-2015-0037","DOIUrl":"https://doi.org/10.1515/sagmb-2015-0037","url":null,"abstract":"Abstract Differential methylation of regulatory elements is critical in epigenetic researches and can be statistically tested. We developed a new statistical test, the generalized integrated functional test (GIFT), that tests for regional differences in methylation based on the methylation percent at each CpG site within a genomic region. The GIFT uses estimated subject-specific profiles with smoothing methods, specifically wavelet smoothing, and calculates an ANOVA-like test to compare the average profile of groups. In this way, possibly correlated CpG sites within the regulatory region are compared all together. Simulations and analyses of data obtained from patients with chronic lymphocytic leukemia indicate that GIFT has good statistical properties and is able to identify promising genomic regions. Further, GIFT is likely to work with multiple different types of experiments since different smoothing methods can be used to estimate the profiles of data without noise. Matlab code for GIFT and sample data are available at http://www.augusta.edu/mcg/biostatepi/people/software/gift.html.","PeriodicalId":49477,"journal":{"name":"Statistical Applications in Genetics and Molecular Biology","volume":"15 1","pages":"237 - 251"},"PeriodicalIF":0.9,"publicationDate":"2016-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1515/sagmb-2015-0037","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67002915","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract The instability in the selection of models is a major concern with data sets containing a large number of covariates. We focus on stability selection which is used as a technique to improve variable selection performance for a range of selection methods, based on aggregating the results of applying a selection procedure to sub-samples of the data where the observations are subject to right censoring. The accelerated failure time (AFT) models have proved useful in many contexts including the heavy censoring (as for example in cancer survival) and the high dimensionality (as for example in micro-array data). We implement the stability selection approach using three variable selection techniques—Lasso, ridge regression, and elastic net applied to censored data using AFT models. We compare the performances of these regularized techniques with and without stability selection approaches with simulation studies and two real data examples–a breast cancer data and a diffuse large B-cell lymphoma data. The results suggest that stability selection gives always stable scenario about the selection of variables and that as the dimension of data increases the performance of methods with stability selection also improves compared to methods without stability selection irrespective of the collinearity between the covariates.
{"title":"Stability selection for lasso, ridge and elastic net implemented with AFT models","authors":"M. H. R. Khan, Anamika Bhadra, Tamanna Howlader","doi":"10.1515/sagmb-2017-0001","DOIUrl":"https://doi.org/10.1515/sagmb-2017-0001","url":null,"abstract":"Abstract The instability in the selection of models is a major concern with data sets containing a large number of covariates. We focus on stability selection which is used as a technique to improve variable selection performance for a range of selection methods, based on aggregating the results of applying a selection procedure to sub-samples of the data where the observations are subject to right censoring. The accelerated failure time (AFT) models have proved useful in many contexts including the heavy censoring (as for example in cancer survival) and the high dimensionality (as for example in micro-array data). We implement the stability selection approach using three variable selection techniques—Lasso, ridge regression, and elastic net applied to censored data using AFT models. We compare the performances of these regularized techniques with and without stability selection approaches with simulation studies and two real data examples–a breast cancer data and a diffuse large B-cell lymphoma data. The results suggest that stability selection gives always stable scenario about the selection of variables and that as the dimension of data increases the performance of methods with stability selection also improves compared to methods without stability selection irrespective of the collinearity between the covariates.","PeriodicalId":49477,"journal":{"name":"Statistical Applications in Genetics and Molecular Biology","volume":"18 1","pages":""},"PeriodicalIF":0.9,"publicationDate":"2016-04-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1515/sagmb-2017-0001","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67002675","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The statistical methodology developed in this study was motivated by our interest in studying neurodevelopment using the mouse brain RNA-Seq data set, where gene expression levels were measured in multiple layers in the somatosensory cortex across time in both female and male samples. We aim to identify differentially expressed genes between adjacent time points, which may provide insights on the dynamics of brain development. Because of the extremely small sample size (one male and female at each time point), simple marginal analysis may be underpowered. We propose a Markov random field (MRF)-based approach to capitalizing on the between layers similarity, temporal dependency and the similarity between sex. The model parameters are estimated by an efficient EM algorithm with mean field-like approximation. Simulation results and real data analysis suggest that the proposed model improves the power to detect differentially expressed genes than simple marginal analysis. Our method also reveals biologically interesting results in the mouse brain RNA-Seq data set.
本研究采用的统计方法源自我们对利用小鼠大脑 RNA-Seq 数据集研究神经发育的兴趣，在该数据集中，我们测量了雌性和雄性样本体感皮层多层基因在不同时间的表达水平。我们的目标是找出相邻时间点之间表达不同的基因，从而为了解大脑发育的动态提供依据。由于样本量极小（每个时间点只有一名男性和一名女性），简单的边际分析可能无法达到预期效果。我们提出了一种基于马尔可夫随机场（MRF）的方法，以利用层间相似性、时间依赖性和性别相似性。模型参数通过有效的 EM 算法进行估计，并采用类似均值场的近似方法。模拟结果和实际数据分析表明，与简单的边际分析相比，所提出的模型提高了检测差异表达基因的能力。我们的方法还揭示了小鼠大脑 RNA-Seq 数据集中有趣的生物学结果。
{"title":"A Markov random field-based approach for joint estimation of differentially expressed genes in mouse transcriptome data.","authors":"Zhixiang Lin, Mingfeng Li, Nenad Sestan, Hongyu Zhao","doi":"10.1515/sagmb-2015-0070","DOIUrl":"10.1515/sagmb-2015-0070","url":null,"abstract":"<p><p>The statistical methodology developed in this study was motivated by our interest in studying neurodevelopment using the mouse brain RNA-Seq data set, where gene expression levels were measured in multiple layers in the somatosensory cortex across time in both female and male samples. We aim to identify differentially expressed genes between adjacent time points, which may provide insights on the dynamics of brain development. Because of the extremely small sample size (one male and female at each time point), simple marginal analysis may be underpowered. We propose a Markov random field (MRF)-based approach to capitalizing on the between layers similarity, temporal dependency and the similarity between sex. The model parameters are estimated by an efficient EM algorithm with mean field-like approximation. Simulation results and real data analysis suggest that the proposed model improves the power to detect differentially expressed genes than simple marginal analysis. Our method also reveals biologically interesting results in the mouse brain RNA-Seq data set. </p>","PeriodicalId":49477,"journal":{"name":"Statistical Applications in Genetics and Molecular Biology","volume":"15 1","pages":"139-50"},"PeriodicalIF":0.9,"publicationDate":"2016-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5587217/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"67002968","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"数学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：