Computation of Carlson's Multiple Hypergeometric Function R for Bayesian Applications

Abstract

Carlson's multiple hypergeometric functions arise in Bayesian inference, including methods for multinomial data with missing category distinctions and for local smoothing of histograms. To use these methods one needs to calculate Carlson functions and their ratios. We discuss properties of the functions and explore computational methods for them, including closed form methods, expansion methods, Laplace approximations, and Monte Carlo methods. Examples are given to illustrate and compare methods.     

On the Orderings and Groupings of Conditional Maximizations within ECM-Type Algorithms

Abstract

The ECM and ECME algorithms are generalizations of the EM algorithm in which the maximization (M) step is replaced by several conditional maximization (CM) steps. The order that the CM-steps are performed is trivial to change and generally affects how fast the algorithm converges. Moreover, the same order of CM-steps need not be used at each iteration and in some applications it is feasible to group two or more CM-steps into one larger CM-step. These issues also arise when implementing the Gibbs sampler, and in this article we study them in the context of fitting log-linear and random-effects models with ECM-type algorithms. We find that some standard theoretical measures of the rate of convergence can be of little use in comparing the computational time required, and that common strategies such as using a random ordering may not provide the desired effects. We also develop two algorithms for fitting random-effects models to illustrate that with careful selection of CM-steps, ECM-type algorithms can be substantially faster than the standard EM algorithm.     

The effect of the biological variability of samples on Coomassie blue dye based fast staining for SDS‐PAGE in nonfixed gels

Fast‐staining protocols based on the use of Coomassie blue dye for SDS‐PAGE separated proteins, represent a quick and simple solution for protein visualization. It has been shown however, that in some cases a phenomenon of missing spots or spot discoloration may be observed in the proteome pattern when the standard fast‐staining protocol is used. In this work, it is demonstrated that this occurrence is affected by the biological variability of samples, and therefore, cannot be observed in all samples. Moreover, it is demonstrated that the phenomenon is manifested exclusively in nonfixed gels, and that including a fixation step into the fast‐staining protocol prevented this phenomenon. In conclusion, it has been demonstrated that standard Coomassie blue dye based fast staining for SDS‐PAGE resolved proteins is affected by the biological variability of samples in nonfixed gels.     

第五讲生物医学信息处理--DNA微阵列数据在医学中的应用

飞速发展的生物信息技术为现代医学提供了更为有效的工具．特别是随着人类基因组计划的基本完成和逐步细化，人们已经试图从基因水平上来认识生命现象，特别是一些重要疾病的机理．由于生物特性一般都涉及到多个基因的共同表达，这便出现了同时衡量成千上万个基因的表现水平的所谓DNA微阵列技术与数据．DNA微阵列数据也被称为大规模基因表达谱．根据这些微阵列数据，人们不仅能够对一些疾病进行分析，并且还能够发现一些新的生物特性与规律．另外，利用微阵列数据能够选取出疾病的相关基因并进行疾病的分类与诊断．这项研究无疑将推动医学的发展．最近，人们还进一步通过基因表达水平值来发现基因之间的调控方式，这将为疾病病理的研究与治疗提供更科学的依据．     

New in situ hybridization towards single-RNA detection

Fluorescence in situ hybridization (ISH) could be a powerful tool for analyzing the dynamics of memory phases if a semi-quantitative measurement of mRNA levels could be performed with a cellular resolution. While such a method does not yet exist, we show preliminary results in that direction where direct labeling of RNA probes with fluorescent dyes is combined with single-molecule-like detection.     

The role of physicochemical interactions and FLO genes expression in the immobilization of industrially important yeasts by adhesion

One of the industrially important qualities of yeast is their ability to provide the cell-cell and cell-support interactions. This feature of yeast is responsible for technologically significant phenomena such as flocculation (brewing) and yeast biofilm formation (immobilization to supports), whereas these phenomena are time, environment, and strain dependent. Therefore, the goal of this work was to verify the possibility to predict and subsequently select yeast strains capable to colonize solid supports by using physicochemical adhesion models. Three different industrial yeast strains (Saccharomyces cerevisiae) were tested for their adhesion onto spent grain particles in the continuous gas-lift reactor. The cell adhesion energies were calculated, based on physicochemical characteristics of surfaces involved, according to three adhesion models (DLVO theory, thermodynamic approach, and extended DLVO theory). The role of physicochemical surface properties in the cell-cell and cell-support interactions was evaluated by comparing the computed predictions with experimental results. The best agreement between forecast and observation of the yeast adhesion to spent grains was achieved with the extended DLVO (XDLVO) theory, the most complex adhesion model applied in this study. Despite its relative comprehensiveness, the XDLVO theory does not take into account specific biochemical interactions. Consequently, additional understanding of the yeast adhesion mechanism was obtained by means of quantifying the expression of selected FLO genes. The presented approach provides tools to select the appropriately adhesive yeast strains and match them with solid supports of convenient surface properties in order to design immobilized biocatalysts exploitable in biotechnological processes.     

Efficiency of Monte Carlo EM and Simulated Maximum Likelihood in Two-Stage Hierarchical Models

Likelihood estimation in hierarchical models is often complicated by the fact that the likelihood function involves an analytically intractable integral. Numerical approximation to this integral is an option but it is generally not recommended when the integral dimension is high. An alternative approach is based on the ideas of Monte Carlo integration, which approximates the intractable integral by an empirical average based on simulations. This article investigates the efficiency of two Monte Carlo estimation methods, the Monte Carlo EM (MCEM) algorithm and simulated maximum likelihood (SML). We derive the asymptotic Monte Carlo errors of both methods and show that, even under the optimal SML importance sampling distribution, the efficiency of SML decreases rapidly (relative to that of MCEM) as the missing information about the unknown parameter increases. We illustrate our results in a simple mixed model example and perform a simulation study which shows that, compared to MCEM, SML can be extremely inefficient in practical applications.     

5’ and 3’ splicing signals evolution in vertebrates: Analysis in a conserved gene family

The mitochondrial solute carrier genes (SLC25) are highly conserved during vertebrate evolution. In most SLC25 genes of zebrafish, chicken, mouse, and human, the introns are located at exactly superimposable positions. In these topographically corresponding introns we studied the composition of the initial and terminal hexanucleotides (5’ss and 3’ss) which are instrumental in splicing signaling, focusing on the evolutionary conservation/mutation dynamics of these genetically related sequences. At each position, the per cent conservation of zebrafish individual nucleotides in chicken, mouse and human is proportional to their percent frequency in zebrafish; furthermore, nucleotide mutations are biased in favor of the more represented nucleotides, thus compensating for those highly represented zebrafish nucleotides which have not been conserved. As a result of these evolutionary dynamics, the general nucleotide composition at each position has remained relatively conserved throughout vertebrates. At 5’ss, following the canonical GT, A and G are largely prevailing at position +3, A at +4 and G at +5 (GT[A/G]AGx). At 3’ss, T and C are largely prevailing at positions −6, −5 and −3, preceding the canonical intron terminal AG ([C/T] [C/T]x[C/T]AG). However, the actual composition of the tetranucleotides at 5’ and 3’ often does not conform to the above scheme. At 5’ss the more canonical sequence is completely expressed in 63% of cases and partially (2 or 1 matches) in 37 % of cases. At 3’ss the more canonical sequence is completely expressed in 71 % of cases and partially (2 or 1 matches) in 29 % of cases. The nucleotide conservation loss (nucleotide mutation) is higher in the evolution from fish to the last common ancestor of birds and mammals (58 %), then diminishes in the successive evolution steps up to the mammalian common ancestor (10 %), and becomes still lower at the divergence of rodents and primates (5 %).     

A review of ARGs in WWTPs: Sources,stressors and elimination

Antibiotic resistance genes (ARGs) in aquatic environments, which seriously endanger human health and ecological safety, have become a worldwide concern due to their easy diffusion and proliferation. Wastewater treatment plants (WWTPs), which receive resistant bacteria and ARGs from a wide variety of sources (i.e., livestock farms, hospitals, antibiotic manufactures, and households), are regarded as important emission sources of aquatic ARGs. This review presents a quantitative profile of the majority sources of ARGs in the influent of WWTPs and discusses the potential factors that affect the concentration distribution of ARGs. Specifically, a noteworthy existence of ARGs, which ranged from 1E + 05 to 1E + 11 copies/mL, was detected in livestock breeding wastewater, and household wastewater (caused by the unlimited utilization of antibiotics) was determined to be the predominant contributor of ARGs in WWTPs. We summarized the selective pressure on ARGs and determined the positive correlation of the concentration of ARGs and the existence of many containments, including antibiotics, heavy metals (Zn and Cu were frequently reported), quaternary ammonium compounds, etc. In the last section, physical, chemical, and biological treatments for the removal of ARGs and their effluent in WWTPs are discussed and prospective future studies are summarized.