Empirical Bayesian analysis of paired high-throughput sequencing data with a beta-binomial distribution.
Ontology highlight
ABSTRACT: Background
Pairing of samples arises naturally in many genomic experiments; for example, gene expression in tumour and normal tissue from the same patients. Methods for analysing high-throughput sequencing data from such experiments are required to identify differential expression, both within paired samples and between pairs under different experimental conditions.Results
We develop an empirical Bayesian method based on the beta-binomial distribution to model paired data from high-throughput sequencing experiments. We examine the performance of this method on simulated and real data in a variety of scenarios. Our methods are implemented as part of the RbaySeq package (versions 1.11.6 and greater) available from Bioconductor (http://www.bioconductor.org).Conclusions
We compare our approach to alternatives based on generalised linear modelling approaches and show that our method offers significant gains in performance on simulated data. In testing on real data from oral squamous cell carcinoma patients, we discover greater enrichment of previously identified head and neck squamous cell carcinoma associated gene sets than has previously been achieved through a generalised linear modelling approach, suggesting that similar gains in performance may be found in real data. Our methods thus show real and substantial improvements in analyses of high-throughput sequencing data from paired samples.
SUBMITTER: Hardcastle TJ
PROVIDER: S-EPMC3658937 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:BACKGROUND: Sequencing-by-synthesis technologies significantly improve over the Sanger method in terms of speed and cost per base. However, they still usually fail to compete in terms of read length and quality. Current high-throughput implementations of the pyrosequencing technique yield reads whose length approach those of the capillary electrophoresis method. A less obvious question is whether their quality is affected by platform-specific sequencing errors. RESULTS: We present an empirical study aimed at assessing the quality and characterising sequencing errors for high throughput pyrosequencing data. We have developed a procedure for extracting sequencing error data from genome assemblies and study their characteristics, in particular the length distribution of indel gaps and their relation to the sequence contexts where they occur. We used this procedure to analyse data from three prokaryotic genomes sequenced with the GS FLX technology. We also compared two models previously employed with success for peptide sequence alignment. CONCLUSIONS: We observed an overall very low error rate in the analysed data, with indel errors being much more abundant than substitutions. We also observed a dependence between the length of the gaps and that of the homopolymer context where they occur. As with protein alignments, a power-law model seems to approximate the indel errors more accurately, although the results are not so conclusive as to justify a depart from the commonly used affine gap penalty scheme. In whichever case, however, our procedure can be used to estimate more realistic error model parameters.
| S-EPMC3852801 | biostudies-literature
Project description:BackgroundSingle-cell RNA sequencing (scRNA-seq) is a powerful profiling technique at the single-cell resolution. Appropriate analysis of scRNA-seq data can characterize molecular heterogeneity and shed light into the underlying cellular process to better understand development and disease mechanisms. The unique analytic challenge is to appropriately model highly over-dispersed scRNA-seq count data with prevalent dropouts (zero counts), making zero-inflated dimensionality reduction techniques popular for scRNA-seq data analyses. Employing zero-inflated distributions, however, may place extra emphasis on zero counts, leading to potential bias when identifying the latent structure of the data.ResultsIn this paper, we propose a fully generative hierarchical gamma-negative binomial (hGNB) model of scRNA-seq data, obviating the need for explicitly modeling zero inflation. At the same time, hGNB can naturally account for covariate effects at both the gene and cell levels to identify complex latent representations of scRNA-seq data, without the need for commonly adopted pre-processing steps such as normalization. Efficient Bayesian model inference is derived by exploiting conditional conjugacy via novel data augmentation techniques.ConclusionExperimental results on both simulated data and several real-world scRNA-seq datasets suggest that hGNB is a powerful tool for cell cluster discovery as well as cell lineage inference.
| S-EPMC7487589 | biostudies-literature
Project description:A main application for mRNA sequencing (mRNAseq) is determining lists of differentially-expressed genes (DEGs) between two or more conditions. Several software packages exist to produce DEGs from mRNAseq data, but they typically yield different DEGs, sometimes markedly so. The underlying probability model used to describe mRNAseq data is central to deriving DEGs, and not surprisingly most softwares use different models and assumptions to analyze mRNAseq data. Here, we propose a mechanistic justification to model mRNAseq as a binomial process, with data from technical replicates given by a binomial distribution, and data from biological replicates well-described by a beta-binomial distribution. We demonstrate good agreement of this model with two large datasets. We show that an emergent feature of the beta-binomial distribution, given parameter regimes typical for mRNAseq experiments, is the well-known quadratic polynomial scaling of variance with the mean. The so-called dispersion parameter controls this scaling, and our analysis suggests that the dispersion parameter is a continually decreasing function of the mean, as opposed to current approaches that impose an asymptotic value to the dispersion parameter at moderate mean read counts. We show how this leads to current approaches overestimating variance for moderately to highly expressed genes, which inflates false negative rates. Describing mRNAseq data with a beta-binomial distribution thus may be preferred since its parameters are relatable to the mechanistic underpinnings of the technique and may improve the consistency of DEG analysis across softwares, particularly for moderately to highly expressed genes.
| S-EPMC4915702 | biostudies-literature
Project description:High-throughput bisulfite sequencing technologies have provided a comprehensive and well-fitted way to investigate DNA methylation at single-base resolution. However, there are substantial bioinformatic challenges to distinguish precisely methylcytosines from unconverted cytosines based on bisulfite sequencing data. The challenges arise, at least in part, from cell heterozygosis caused by multicellular sequencing and the still limited number of statistical methods that are available for methylcytosine calling based on bisulfite sequencing data. Here, we present an algorithm, termed Bycom, a new Bayesian model that can perform methylcytosine calling with high accuracy. Bycom considers cell heterozygosis along with sequencing errors and bisulfite conversion efficiency to improve calling accuracy. Bycom performance was compared with the performance of Lister, the method most widely used to identify methylcytosines from bisulfite sequencing data. The results showed that the performance of Bycom was better than that of Lister for data with high methylation levels. Bycom also showed higher sensitivity and specificity for low methylation level samples (<1%) than Lister. A validation experiment based on reduced representation bisulfite sequencing data suggested that Bycom had a false positive rate of about 4% while maintaining an accuracy of close to 94%. This study demonstrated that Bycom had a low false calling rate at any methylation level and accurate methylcytosine calling at high methylation levels. Bycom will contribute significantly to studies aimed at recalibrating the methylation level of genomic regions based on the presence of methylcytosines.
| S-EPMC4177668 | biostudies-literature
Project description:A common goal of microarray and related high-throughput genomic experiments is to identify genes that vary across biological condition. Most often this is accomplished by identifying genes with changes in mean expression level, so called differentially expressed (DE) genes, and a number of effective methods for identifying DE genes have been developed. Although useful, these approaches do not accommodate other types of differential regulation. An important example concerns differential coexpression (DC). Investigations of this class of genes are hampered by the large cardinality of the space to be interrogated as well as by influential outliers. As a result, existing DC approaches are often underpowered, exceedingly prone to false discoveries, and/or computationally intractable for even a moderately large number of pairs. To address this, an empirical Bayesian approach for identifying DC gene pairs is developed. The approach provides a false discovery rate controlled list of significant DC gene pairs without sacrificing power. It is applicable within a single study as well as across multiple studies. Computations are greatly facilitated by a modification to the expectation-maximization algorithm and a procedural heuristic. Simulations suggest that the proposed approach outperforms existing methods in far less computational time; and case study results suggest that the approach will likely prove to be a useful complement to current DE methods in high-throughput genomic studies.
| S-EPMC3482945 | biostudies-literature
Project description:The beta-binomial model has been widely used as an analytically tractable alternative that captures the overdispersion of an intra-correlated, binomial random variable, X. However, the model validation for X has been rarely investigated. As a beta-binomial mass function takes on a few different shapes, the model validation is examined for each of the classified shapes in this paper. Further, the mean square error (MSE) is illustrated for each shape by the maximum likelihood estimator (MLE) based on a beta-binomial model approach and the method of moments estimator (MME) in order to gauge when and how much the MLE is biased.
| S-EPMC5736152 | biostudies-literature
Project description:It is important to correctly understand the associations among addiction to multiple drugs and between co-occurring substance use and psychiatric disorders. Substance-specific outcomes (e.g. number of days used cannabis) have distributional characteristics which range widely depending on the substance and the sample being evaluated.We recommend a four-part strategy for determining the appropriate distribution for modeling substance use data. We demonstrate this strategy by comparing the model fit and resulting inferences from applying four different distributions to model use of substances that range greatly in the prevalence and frequency of their use.Using Timeline Followback (TLFB) data from a previously-published study, we used negative binomial, beta-binomial and their zero-inflated counterparts to model proportion of days during treatment of cannabis, cigarettes, alcohol, and opioid use. The fit for each distribution was evaluated with statistical model selection criteria, visual plots and a comparison of the resulting inferences.We demonstrate the feasibility and utility of modeling each substance individually and show that no single distribution provides the best fit for all substances. Inferences regarding use of each substance and associations with important clinical variables were not consistent across models and differed by substance.Thus, the distribution chosen for modeling substance use must be carefully selected and evaluated because it may impact the resulting conclusions. Furthermore, the common procedure of aggregating use across different substances may not be ideal.
| S-EPMC4692373 | biostudies-literature
Project description:NGS studies have uncovered an ever-growing catalog of human variation while leaving an enormous gap between observed variation and experimental characterization of variant function. High-throughput screens powered by NGS have greatly increased the rate of variant functionalization, but the development of comprehensive statistical methods to analyze screen data has lagged. In the massively parallel reporter assay (MPRA), short barcodes are counted by sequencing DNA libraries transfected into cells and the cell's output RNA in order to simultaneously measure the shifts in transcription induced by thousands of genetic variants. These counts present many statistical challenges, including overdispersion, depth dependence, and uncertain DNA concentrations. So far, the statistical methods used have been rudimentary, employing transformations on count level data and disregarding experimental and technical structure while failing to quantify uncertainty in the statistical model. We have developed an extensive framework for the analysis of NGS functionalization screens available as an R package called malacoda (available from github.com/andrewGhazi/malacoda). Our software implements a probabilistic, fully Bayesian model of screen data. The model uses the negative binomial distribution with gamma priors to model sequencing counts while accounting for effects from input library preparation and sequencing depth. The method leverages the high-throughput nature of the assay to estimate the priors empirically. External annotations such as ENCODE data or DeepSea predictions can also be incorporated to obtain more informative priors-a transformative capability for data integration. The package also includes quality control and utility functions, including automated barcode counting and visualization methods. To validate our method, we analyzed several datasets using malacoda and alternative MPRA analysis methods. These data include experiments from the literature, simulated assays, and primary MPRA data. We also used luciferase assays to experimentally validate several hits from our primary data, as well as variants for which the various methods disagree and variants detectable only with the aid of external annotations.
| S-EPMC7394446 | biostudies-literature
Project description:BackgroundLarge insert paired-end sequencing technologies are important tools for assembling genomes, delineating associated breakpoints and detecting structural rearrangements. To facilitate the comprehensive detection of inter- and intra-chromosomal structural rearrangements or variants (SVs) and complex genome assembly with long repeats and segmental duplications, we developed a new method based on single-molecule real-time synthesis sequencing technology for generating long paired-end sequences of large insert DNA libraries.ResultsA Fosmid vector, pHZAUFOS3, was developed with the following new features: (1) two 18-bp non-palindromic I-SceI sites flank the cloning site, and another two sites are present in the skeleton of the vector, allowing long DNA inserts (and the long paired-ends in this paper) to be recovered as single fragments and the vector (~?8 kb) to be fragmented into 2-3 kb fragments by I-SceI digestion and therefore was effectively removed from the long paired-ends (5-10 kb); (2) the chloramphenicol (Cm) resistance gene and replicon (oriV), necessary for colony growth, are located near the two sides of the cloning site, helping to increase the proportion of the paired-end fragments to single-end fragments in the paired-end libraries. Paired-end libraries were constructed by ligating the size-selected, mechanically sheared pooled Fosmid DNA fragments to the Ampicillin (Amp) resistance gene fragment and screening the colonies with Cm and Amp. We tested this method on yeast and Setaria italica Yugu1. Fosmid-size paired-ends with an average length longer than 2 kb for each end were generated. The N50 scaffold lengths of the de novo assemblies of the yeast and S. italica Yugu1 genomes were significantly improved. Five large and five small structural rearrangements or assembly errors spanning tens of bp to tens of kb were identified in S. italica Yugu1 including deletions, inversions, duplications and translocations.ConclusionsWe developed a new method for long paired-end sequencing of large insert libraries, which can efficiently improve the quality of de novo genome assembly and identify large and small structural rearrangements or assembly errors.
| S-EPMC6878638 | biostudies-literature
Project description:Recent advances in high-throughput sequencing technologies have enabled a comprehensive dissection of the cancer genome clarifying a large number of somatic mutations in a wide variety of cancer types. A number of methods have been proposed for mutation calling based on a large amount of sequencing data, which is accomplished in most cases by statistically evaluating the difference in the observed allele frequencies of possible single nucleotide variants between tumours and paired normal samples. However, an accurate detection of mutations remains a challenge under low sequencing depths or tumour contents. To overcome this problem, we propose a novel method, Empirical Bayesian mutation Calling (https://github.com/friend1ws/EBCall), for detecting somatic mutations. Unlike previous methods, the proposed method discriminates somatic mutations from sequencing errors based on an empirical Bayesian framework, where the model parameters are estimated using sequencing data from multiple non-paired normal samples. Using 13 whole-exome sequencing data with 87.5-206.3 mean sequencing depths, we demonstrate that our method not only outperforms several existing methods in the calling of mutations with moderate allele frequencies but also enables accurate calling of mutations with low allele frequencies (? 10%) harboured within a minor tumour subpopulation, thus allowing for the deciphering of fine substructures within a tumour specimen.
| S-EPMC3627598 | biostudies-literature