Identifying Appropriate Probabilistic Models for Sparse Discrete Omics Data

Aldirawi, Hani; Yang, Jie

doi:10.1109/bhi.2019.8834661

Cited by 14 publications

(16 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To identify the silvicultural, ecological and environmental variables that best explained variations in sapling density and occurrence of tolerant hardwood species (American beech and sugar maple saplings), we compared a series of zeroinflated negative binomial (ZINB) models (function "zeroinfl", package "pscl" in R) Zeileis et al 2008). Several authors recommend using ZINB models for over-dispersed count data (e.g., sapling density) with extra zeros (Thomas et al 2018;Aldirawi et al 2019;Wang et al 2020). In this case, the ZINB distribution is more appropriate for modeling count data compared to the commonly used Poisson distribution (Aldirawi et al 2019).…”

Section: Statistical Analysesmentioning

confidence: 99%

“…Several authors recommend using ZINB models for over-dispersed count data (e.g., sapling density) with extra zeros (Thomas et al 2018;Aldirawi et al 2019;Wang et al 2020). In this case, the ZINB distribution is more appropriate for modeling count data compared to the commonly used Poisson distribution (Aldirawi et al 2019). In addition, we used ZINB models because they can simultaneously predict the density and the occurrence probability of saplings, while modeling nonlinear effects of silvicultural, environmental, and ecological factors.…”

Section: Statistical Analysesmentioning

confidence: 99%

See 1 more Smart Citation

Modeling tolerant hardwood sapling density and occurrence probability in the Acadian forests of New Brunswick, Canada: Results 14 years after harvesting

HennebMohammed

PelletierGaetan²,

FortinMathieu

et al. 2021

The Forestry Chronicle

View full text Add to dashboard Cite

Natural forest regeneration after natural or anthropogenic disturbance is difficult to predict given its high variability. The process is poorly documented for commercial northern hardwood species in the Acadian forest of eastern Canada. Our objective was to identify the silvicultural, environmental, and ecological factors that best explain the variability in sapling density and occurrence of two commercial tolerant hardwood species in New Brunswick: American beech (Fagus grandifolia Ehrh.) and sugar maple (Acer saccharum Marsh.). Forty-three permanent sample plots were established in 2002 and measured before harvesting in 2004. Sapling density and occurrence were measured 14 years after harvesting. The results showed that the interactions between the species and the residual merchantable basal area and between the species and the percent of hardwoods in the original stand best explained the sapling density and occurrence variation of tolerant hardwoods. The sapling density of sugar maple increased with increasing merchantable residual basal area. However, the effect of this variable was not significant for the density of American beech saplings. The density and occurrence of tolerant hardwood saplings both increased along with the percent of hardwoods in the original stand. These results provide an improved understanding about tolerant hardwood regeneration dynamics in New Brunswick forests.

show abstract

Section: Statistical Analysesmentioning

confidence: 99%

Section: Statistical Analysesmentioning

confidence: 99%

Modeling tolerant hardwood sapling density and occurrence probability in the Acadian forests of New Brunswick, Canada: Results 14 years after harvesting

HennebMohammed

PelletierGaetan²,

FortinMathieu

et al. 2021

The Forestry Chronicle

View full text Add to dashboard Cite

show abstract

“…With appropriate probabilistic models identified successfully, we can improve the power of the statistical test significantly. Recently, Aldirawi, Yang, and Metwally (2019) proposed a statistical procedure for identifying the most appropriate discrete probabilistic models for zero-inflated or hurdle models based on the p-value of the discrete Kolmogorov-Smirnov (KS) test. The same procedure could be used for more general zero-inflated or hurdle models, including the ones with continuous baseline distributions.…”

Section: Introductionmentioning

confidence: 99%

“…More specifically, the goal is to test if the sample X = {X 1 , X 2 , .., X n } comes from a discrete or mixed distribution with cumulative distribution function (CDF) F θ (x) where the parameter(s) θ is unknown. Algorithm 1, which is regenerated from Aldirawi, Yang, and Metwally (2019), provides our procedures in details.…”

Section: Introductionmentioning

confidence: 99%

Modeling Sparse Data Using MLE with Applications to Microbiome Data

Aldirawi¹,

Yang²

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Modeling sparse data such as microbiome and transcriptomics (RNA-seq) data is very challenging due to the exceeded number of zeros and skewness of the distribution. Many probabilistic models have been used for modeling sparse data, including Poisson, negative binomial, zero-inflated Poisson, and zero-inflated negative binomial models. One way to identify the most appropriate probabilistic models for zero-inflated or hurdle models is based on the p-value of the Kolmogorov-Smirnov (KS) test. The main challenge for identifying the probabilistic model is that the model parameters are typically unknown in practice. This paper derives the maximum likelihood estimator (MLE) for a general class of zero-inflated and hurdle models. We also derive the corresponding Fisher information matrices for exploring the estimator's asymptotic properties. We include new probabilistic models such as zero-inflated beta binomial and zero-inflated beta negative binomial models. Our application to microbiome data shows that our new models are more appropriate for modeling microbiome data than commonly used models in the literature.

show abstract

“…For example, quite often people do not have dental caries (Preisser and others , 2016), and the majority of a population does not make a hospital visit during a given year (Gurmu, 1997). In omics data, either because of technological reasons related to sequencing or due to some biologoical reasons, counts are often very sparse (Risso and others , 2018; Van den Berge and others , 2018; Aldirawi and others , 2019). Such count data with excess number of zeros are frequently modeled using the zero-inflated Poisson (ZIP) or the zero-inflated negative binomial (ZINB) distributions (Lambert, 1992; Greene, 1994).…”

Section: Introductionmentioning

confidence: 99%

A bivariate zero-inflated negative binomial model and its applications to biomedical settings

Cho

Liu

Preisser

et al. 2020

Preprint

View full text Add to dashboard Cite

Measuring gene-gene dependence in single cell RNA sequencing (scRNA-seq) count data is often of interest and remains challenging, because an unidentified portion of the zero counts represent non-detected RNA due to technical reasons. Conventional statistical methods that fail to account for technical zeros incorrectly measure the dependence among genes. To address this problem, we propose a bivariate zero-inflated negative binomial (BZINB) model constructed using a bivariate Poisson-gamma mixture with dropout indicators for the technical (excess) zeros. Parameters are estimated based on the EM algorithm and are used to measure the underlying dependence by decomposing the two sources of zeros. Compared to existing models, the proposed BZINB model is specifically designed for estimating dependence and is more flexible, while preserving the marginal zero-inflated negative binomial distributions. Additionally, it has a simple latent variable framework, allowing parameters to have clear and intuitive interpretations, and its computation is feasible with large scale data. Using a recent scRNA-seq dataset, we illustrate model fitting and how the model-based measures can be different from naive measures. The inferential ability of the proposed model is evaluated in a simulation study. An R package 'bzinb' is available on CRAN.

show abstract

Identifying Appropriate Probabilistic Models for Sparse Discrete Omics Data

Cited by 14 publications

References 18 publications

Modeling tolerant hardwood sapling density and occurrence probability in the Acadian forests of New Brunswick, Canada: Results 14 years after harvesting

Modeling tolerant hardwood sapling density and occurrence probability in the Acadian forests of New Brunswick, Canada: Results 14 years after harvesting

Modeling Sparse Data Using MLE with Applications to Microbiome Data

A bivariate zero-inflated negative binomial model and its applications to biomedical settings

Contact Info

Product

Resources

About