peter müller - Academia.edu (original) (raw)

Papers by peter müller

Bayesian Inference for Gene Expression and Proteomics

Journal of the Royal Statistical Society Series C: Applied Statistics, 2005

Summary We propose model-based inference for differential gene expression, using a nonparametric ... more Summary We propose model-based inference for differential gene expression, using a nonparametric Bayesian probability model for the distribution of gene intensities under various conditions. The probability model is a mixture of normal distributions. The resulting inference is similar to a popular empirical Bayes approach that is used for the same inference problem. The use of fully model-based inference mitigates some of the necessary limitations of the empirical Bayes method. We argue that inference is no more difficult than posterior simulation in traditional nonparametric mixture-of-normal models. The approach proposed is motivated by a microarray experiment that was carried out to identify genes that are differentially expressed between normal tissue and colon cancer tissue samples. Additionally, we carried out a small simulation study to verify the methods proposed. In the motivating case-studies we show how the nonparametric Bayes approach facilitates the evaluation of poster...

Journal of the Royal Statistical Society Series B: Statistical Methodology, 2009

SummaryWe discuss a Bayesian discovery procedure for multiple-comparison problems. We show that, ... more SummaryWe discuss a Bayesian discovery procedure for multiple-comparison problems. We show that, under a coherent decision theoretic framework, a loss function combining true positive and false positive counts leads to a decision rule that is based on a threshold of the posterior probability of the alternative. Under a semiparametric model for the data, we show that the Bayes rule can be approximated by the optimal discovery procedure, which was recently introduced by Storey. Improving the approximation leads us to a Bayesian discovery procedure, which exploits the multiple shrinkage in clusters that are implied by the assumed non-parametric model. We compare the Bayesian discovery procedure and the optimal discovery procedure estimates in a simple simulation study and in an assessment of differential gene expression based on microarray data from tumour samples. We extend the setting of the optimal discovery procedure by discussing modifications of the loss function that lead to dif...

Statistical Science, 2004

We review the current state of nonparametric Bayesian inference. The discussion follows a list of... more We review the current state of nonparametric Bayesian inference. The discussion follows a list of important statistical inference problems, including density estimation, regression, survival analysis, hierarchical models and model validation. For each inference problem we review relevant nonparametric Bayesian models and approaches including Dirichlet process (DP) models and variations, Pólya trees, wavelet based models, neural network models, spline regression, CART, dependent DP models and model validation with DP and Pólya tree extensions of parametric models.

Inorganic Chemistry, 2014

Angewandte Chemie International Edition, 2007

All manipulations were performed in a Vacuum Atmospheres model MO-40M glove box under an atmosphe... more All manipulations were performed in a Vacuum Atmospheres model MO-40M glove box under an atmosphere of purified dinitrogen. Anhydrous-grade, oxygen-free tetrahydrofuran (THF) was purchased from Aldrich and further dried by passing through a column of molecular sieves and stirring with sodium for at least 12 hours prior to filtration through Celite to remove sodium. Hexamethyldisiloxane was distilled from sodium benzophenone and stored over molecular sieves. All other solvents were obtained anhydrous and oxygen-free by bubble degassing (N 2) and purification through columns of alumina and Q5. [1] Diethyl ether (Et 2 O) was additionally dried by stirring over sodium for 12 h, followed by filtration through Celite. Deuterated solvents were purchased from Cambridge Isotope Labs. Benzene-d 6 and toluene-d 8 were degassed and stored over 4 Å molecular sieves for at least 2 days prior to use. THF-d 8 was stirred over sodium and subsequently vacuum transferred. Celite 435 (EM Science), molecular sieves (Aldrich), and alumina were dried by heating at 200 ºC under dynamic vacuum for at least 24 hours prior to use. Sodium azide (Aldrich) was dried under vacuum and washed thoroughly with anhydrous THF. Ph 2 S 2 and Ph 2 Se 2 were dried under vacuum and recrystallized from anhydrous Et 2 O. p-Tetrachlorobenzoquinone was purchased from Aldrich and sublimed in vacuo prior to use. VCl 3 (THF) 3 , [2] LiN(Np)Ar•Et 2 O (Np = neopentyl, Ar = 3,5-C 6 H 3 Me 2), [3] and Ti(N[tBu]Ar) 3 [4] were prepared according to literature procedures. Anhydrous SnCl 2 was purchased from Strem and used without further purification. All other reagents were purchased from commercial sources and used without further purification. All glassware was oven-dried at >170 ºC prior to use. NMR spectra were obtained on Varian Mercury 300 or Varian Inova 500 instruments equipped with Oxford Instruments superconducting magnets and referenced to residual C 6 D 5 H (7.16 ppm), C 4 D 7 HO (1.73 ppm) or CD 3 C 6 D 4 H (7.00 ppm). 77 Se NMR spectra were referenced to Me 2 Se (0 ppm) by comparison to external Ph 2 Se 2 (460 ppm, CDCl 3). 51 V and 31 P NMR spectra were referenced externally to neat OVCl 3 and 85% H 3 PO 4 , respectively. Conversion yields for 1-SPh, 1-SePh, and 4 were taken using 8-scan, 1 H NMR spectra with a delay time of 128 s; the observed integral ratios did not change upon lengthening the delay time to 256 s. Elemental analyses were performed by Midwest Microlab, Indianapolis, Indiana. V(N[Np]Ar) 3 (2): Pink VCl 3 (THF) 3 powder (13.334 g, 35.689 mmol) was slowly added over the course of 1 min to a chilled (ca.-50 ºC), stirring solution of LiN(Np)Ar•Et 2 O (29.488 g, 108.66 mmol, 3.04 equiv) in 100 mL Et 2 O. The color of the resulting solution turned first pink, then brown, and finally dark green. The reaction mixture was allowed to stir while warming to 23 ºC for 5 h, after which time it was concentrated to dryness under dynamic vacuum. The products were dissolved in pentane (ca. 100 mL) and the solution was filtered through Celite to remove LiCl. The filtrate was concentrated in vacuo and solids were subsequently crystallized from this solution at-35 ºC in two crops. The crystallized solids were washed with cold pentane (-35 °C, 20 mL) and subsequently dried under vacuum to yield a dark green powder (

ECOOP 2007 – Object-Oriented Programming

Proceedings 27th EUROMICRO Conference. 2001: A Net Odyssey

During the last years the use of component-based software development (CBSE) gets more and more c... more During the last years the use of component-based software development (CBSE) gets more and more common for desktop applications to speed up time to market and reduce development costs. Especially in the area of embedded real-time systems the reuse of tested and robust parts of prior applications is of great desire, to fulfill the strong requirements on maturity, availability and cost. But further requirements like low power design and real-time execution of components makes it impossible to use component frameworks well known in the desktop area. This paper discusses the problems of componentbased software development for embedded real-time systems and derives requirements for a component framework for this domain. First insights for such a framework are presented.

Lecture Notes in Computer Science, 2006

Model fields are specification-only fields that encode abstractions of the concrete state of a da... more Model fields are specification-only fields that encode abstractions of the concrete state of a data structure. They allow specifications to describe the behavior of object-oriented programs without exposing implementation details. This paper presents a sound verification methodology for model fields that handles object-oriented features, supports data abstraction, and can be applied to a variety of realistic programs. The key innovation of the methodology is a novel encoding of model fields, where updates of the concrete state do not automatically change the values of model fields. Model fields are updated only by a special pack statement. The methodology guarantees that the specified relation between a model field and the concrete state of an object holds whenever the object is valid, that is, is known to satisfy its invariant. The methodology also improves on previous work in three significant ways: First, the formalization of model fields prevents unsoundness, even if an interface specification is inconsistent. Second, the methodology fully supports inheritance. Third, the methodology enables modular reasoning about frame properties without using explicit dependencies, which are not handled well by automatic theorem provers.

European biophysics journal : EBJ, 2003

We have determined the average location and dynamic reorientation of the fluorophore 7-nitrobenz-... more We have determined the average location and dynamic reorientation of the fluorophore 7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD) attached to a C12 sn-2 chain of a phosphatidylserine (PS) analogue (C12-NBD-PS) in zwitterionic phosphatidylcholine (PC) and negatively charged phosphatidylserine (PS) host membranes. (1)H magic angle spinning nuclear Overhauser enhancement spectroscopy indicates a highly dynamic reorientation of the aromatic molecule in the membrane. The average location of NBD is characterized by a broad distribution function along the membrane director with a maximum indicating the location of the probe in the lipid/water interface of the lipid membrane. This behavior can be explained by a backfolding of the sn-2 chain towards the aqueous phase. Small differences in the distribution profiles of the NBD group along the membrane normal between PC and PS host membranes were found: in a PC host membrane, the NBD distribution has its maximum in the glycerol region; in a PS host ...

Physical Review E, 2004

We present event-driven simulations of a granular gas of inelastic hard disks with incomplete nor... more We present event-driven simulations of a granular gas of inelastic hard disks with incomplete normal restitution in two dimensions between vibrating walls (without gravity). We measure hydrodynamic quantities such as the stress tensor, density and temperature profiles, as well as velocity distributions. Relating the local pressure to the local temperature and local density, we construct a local constitutive equation. For strong inelasticities the local constitutive relation depends on global system parameters, like the volume fraction and the aspect ratio. For moderate inelasticities the constitutive relation is approximately independent of the system parameters and can hence be regarded as a local equation of state, even though the system is highly inhomogeneous with heterogeneous temperature and density profiles arising as a consequence of the energy injection. Concerning the local velocity distributions we find that they do not scale with the square root of the local granular temperature. Moreover the high-velocity tails are different for the distribution of the xand the y-component of the velocity, and even depend on the position in the sample, the global volume fraction, and the coefficient of restitution.

Journal of the Royal Statistical Society: Series C (Applied Statistics), 2013

Data obtained using modern sequencing technologies are often summarized by recording the frequenc... more Data obtained using modern sequencing technologies are often summarized by recording the frequencies of observed sequences. Examples include the analysis of T cell counts in immunological research and studies of gene expression based on counts of RNA fragments. In both cases the items being counted are sequences, of proteins and base pairs, respectively. The resulting sequence-abundance distribution is usually characterized by overdispersion. We propose a Bayesian semi-parametric approach to implement inference for such data. Besides modeling the overdispersion, the approach takes also into account two related sources of bias that are usually associated with sequence counts data: some sequence types may not be recorded during the experiment and the total count may differ from one experiment to another. We illustrate our methodology with two data sets, one regarding the analysis of CD4+ T cell counts in healthy and diabetic mice and another data set concerning the comparison of mRNA fragments recorded in a Serial Analysis of Gene Expression (SAGE) experiment with gastrointestinal tissue of healthy and cancer patients.

Electronic Notes in Theoretical Computer Science, 2005

Program logics for bytecode languages such as Java bytecode or the .NET CIL can be used to apply ... more Program logics for bytecode languages such as Java bytecode or the .NET CIL can be used to apply Proof-Carrying Code concepts to bytecode programs and to verify correctness properties of bytecode programs. This paper presents a Hoare-style logic for a sequential bytecode kernel language similar to Java bytecode and CIL. The logic handles object-oriented features such as inheritance, dynamic method binding, and object structures with destructive updates, as well as unstructured control flow with jumps. It is sound and complete.

Biometrics, 2013

We discuss inference for a human phage display experiment with three stages. The data are tripept... more We discuss inference for a human phage display experiment with three stages. The data are tripeptide counts by tissue and stage. The primary aim of the experiment is to identify ligands that bind with high affinity to a given tissue. We formalize the research question as inference about the monotonicity of mean counts over stages. The inference goal is then to identify a list of peptidetissue pairs with significant increase over stages. We use a semi-parametric Dirichlet process mixture Biometrics XX, 1-?? DOI: XX.XXXX/j.XXXX-XXXX.20XX.XXXXX.x Month 200X of Poisson model. The posterior distribution under this model allows the desired inference about the monotonicity of mean counts. However, the desired inference summary as a list of peptide-tissue pairs with significant increase involves a massive multiplicity problem. We consider two alternative approaches to address this multiplicity issue. First we propose an approach based on the control of the posterior expected false discovery rate. We notice that the implied solution ignores the relative size of the increase. This motivates a second approach based on a utility function that includes explicit weights for the size of the increase.

Biochimica et Biophysica Acta (BBA) - Biomembranes, 2011

Molecular rods are synthetical molecules consisting of a hydrophobic backbone which are functiona... more Molecular rods are synthetical molecules consisting of a hydrophobic backbone which are functionalized with varying terminal groups. Here, we report on the interaction of a recently described new class of molecular rods with lipid and biological membranes. In order to characterize this interaction, different fluorescently labeled rods were synthesized allowing for the application of fluorescence spectroscopy and microscopy based approaches. Our data show that the rods are incorporated into membranes with a perpendicular orientation to the membrane surface and enrich preferentially in liquid-disordered lipid domains. These characteristics underline that rods can be applied as stable membrane-associated anchors for functionalizing membrane surfaces.

Biophysical Journal, 2002

The transbilayer movement of fluorescent phospholipid analogs in liposomes was studied at the lip... more The transbilayer movement of fluorescent phospholipid analogs in liposomes was studied at the lipid phase transition of phospholipid membranes. Two NBD-labeled analogs were used, one bearing the fluorescent moiety at a short fatty acid chain in the sn-2 position (C 6-NBD-PC) and one headgroup-labeled analog having two long fatty acyl chains (N-NBD-PE). The transbilayer redistribution of the analogs was assessed by a dithionite-based assay. We observed a drastic increase of the transbilayer movement of both analogs at the lipid phase transition of DPPC (T c ϭ 41°C) and DMPC (T c ϭ 23°C). The flip-flop of analogs was fast at the T c of DPPC with a half-time (t 1/2) of ϳ6-10 min and even faster at the T c of DMPC with t 1/2 on the order of Ͻ2 min, as shown for C 6-NBD-PC. Suppressing the phase transition by the addition of cholesterol, the rapid transbilayer movement was abolished. Molecular packing defects at the phase transition are assumed to be responsible for the rapid transbilayer movement. The relevance of those defects for understanding of the activity of flippases is discussed.

Journal of the Royal Statistical Society: Series C (Applied Statistics), 2008

We discuss the analysis of data from single nucleotide polymorphism (SNP) arrays comparing tumor ... more We discuss the analysis of data from single nucleotide polymorphism (SNP) arrays comparing tumor and normal tissues. The data consist of sequences of indicators for loss of heterozygosity (LOH) and involve three nested levels of repetition: chromosomes for a given patient, regions within chromosomes, and SNPs nested within regions. We propose to analyze these data using a semiparametric model for multi-level repeated binary data. At the top level of the hierarchy we assume a sampling model for the observed binary LOH sequences that arises from a partial exchangeability argument. This implies a mixture of Markov chains model. The mixture is defined with respect to the Markov transition probabilities. We assume a nonparametric prior for the random mixing measure. The resulting model takes the form of a semiparametric random effects model with the matrix of transition probabilities being the random effects. The model includes appropriate dependence assumptions for the two remaining levels of the hierarchy, i.e., for regions within chromosomes and for chromosomes within patient.

Biometrics, 2007

We discuss inference for data with repeated measurements at multiple levels. The motivating examp... more We discuss inference for data with repeated measurements at multiple levels. The motivating example are data with blood counts from cancer patients undergoing multiple cycles of chemotheraphy, with days nested within cycles. Some inference questions relate to repeated measurements over days within cycle, while other questions are concerned with the dependence across cycles. When the desired inference relates to both levels of repetition, it becomes important to reflect the data structure in the model. We develop a semi-parametric Bayesian modeling approach, restricting attention to two levels of repeated measurements. For the top level longitudinal sampling model we use random effects to introduce the desired dependence across repeated measurements. We use a non-parametric prior for the random effects 1 distribution. Inference about dependence across second-level repetition is implemented by the clustering implied in the non-parametric random effects model. Practical use of the model requires that the posterior distribution on the latent random effects be reasonably precise.