Jörg Zimmermann

Adapters, shims, and glue---service interoperability for in silico experiments

MOTIVATION Computationally, in silico experiments in biology are workflows describing the collaboration of people, data and methods. The Grid and Web services are proposed to be the next generation infrastructure supporting the deployment of bioinformatics workflows. But the growing number of autonomous and heterogeneous services pose challenges to the used middleware w.r.t. composition, i.e. discovery and interoperability of services required within in silico experiments. In the IRIS project, we handle the problem of service interoperability by a semi-automatic procedure for identifying and placing customizable adapters into workflows built by service composition. RESULTS We show the effectiveness and robustness of the software-aided composition procedure by a case study in the field of life science. In this study we combine different database services with different analysis services with the objective of discovering required adapters. Our experiments show that we can identify relevant adapters with high precision and recall.

Coordination of decisions in a spatial agent model

For a binary choice problem, the spatial coordination of decisions in an agent community is investigated both analytically and by means of stochastic computer simulations. The individual decisions are based on different local information generated by the agents with a finite lifetime and disseminated in the system with a finite velocity. We derive critical parameters for the emergence of minorities and majorities of agents making opposite decisions and investigate their spatial organization. We find that dependent on two essential parameters describing the local impact and the spatial dissemination of information, either a definite stable minority/majority relation (single-attractor regime) or a broad range of possible values (multi-attractor regime) occurs. In the latter case, the outcome of the decision process becomes rather diverse and hard to predict, both with respect to the share of the majority and their spatial distribution. We further investigate how a dissemination of information on different time scales affects the outcome of the decision process. We find that a more “efficient” information exchange within a subpopulation provides a suitable way to stabilize their majority status and to reduce “diversity” and uncertainty in the decision process.

ENCON: an evolutionary algorithm for the antenna placement problem

The placement of antennas is an important step in the design of mobile radio networks. We introduce a model for the antenna placement problem (APP) that addresses cover, traffic demand, interference, different parameterized antenna types, and the geometrical structure of cells. The resulting optimization problem is constrained and multiobjective. We present an evolutionary algorithm, capable of dealing with more than 700 candidate sites in the working area. The results show that the APP is tractable. The automatically generated designs enable experts to focus their efforts on the difficult parts of a network design problem.

Communication and Self-Organisation in Complex Systems: A Basic Approach

The emergence of complex behaviour in systems consisting of interacting elements is among the most fascinating phenomena of our world. Examples can be found in almost every field of today’s scientific interest, ranging from coherent pattern formation in physical and chemical systems (Feistel and Ebeling 1989; Cladis and Palffy-Muhoray 1995), to the motion of swarms of animals in biology (DeAngelis and Gross 1992) and the behaviour of social groups (Weidlich 1991; Vallacher and Nowak 1994). In the social and life sciences, it has generally been held that the evolution of social systems is determined by numerous factors —cultural, sociological, economic, political, and ecological, etc. However, in recent years, the development of the interdisciplinary ‘science of complexity’ has led to the insight that complex dynamic processes may also result from simple interactions. Moreover, at a certain level of abstraction, one can find many common features between complex structures in very different fields (Schweitzer 1997a, b).

The quest for uncertainty

The question of how to represent and process uncertainty is of fundamental importance to the scientific process, but also in everyday life. Currently there exist a lot of different calculi for managing uncertainty, each having its own advantages and disadvantages. Especially, almost all are defining the domain and structure of uncertainty values a priori, e.g., one real number, two real numbers, a finite domain, and so on, but maybe uncertainty is best measured by complex numbers, matrices or still another mathematical structure. Here we investigate the notion of uncertainty from a foundational point of view, provide an ontology and axiomatic core system for uncertainty, derive and not define the structure of uncertainty, and review the historical development of approaches to uncertainty which have led to the results presented here.

Size of neighborhood more important than temperature for stochastic local search

We investigate stochastic local search by Markov chain analysis in a high and a low dimensional discrete space. In the n-dimensional space B/sup n/ a function called Jump is considered. The analysis shows that an algorithm using a large neighborhood and never accepting worse points performs much better than any local search algorithm accepting worse points with a certain probability. We also investigate functions in the space B/sup n/ with many local optima. We compare stochastic local search using large neighborhoods with a local search using optimal temperature schedules which depend on the state of the Markov process.

From theory to practice: an evolutionary algorithm for the antenna placement problem

We give a short introduction to the results of our theoretical analysis of evolutionary algorithms. These results are used to design an algorithm for a large real-world problem: the placement of antennas for mobile radio networks. Our model for the antenna placement problem (APP) addresses cover, traffic demand, interference, different parameterized antenna types, and the geometrical structure of cells. The resulting optimization problem is constrained and multi-objective. The evolutionary algorithm derived from our theoretical analysis is capable of dealing with more than 700 candidate sites in the working area. The results show that the APP is tractable. The automatically generated designs enable experts to focus their efforts on the difficult parts of a network design problem.

Making solomonoff induction effective: or: you can learn what you can bound

The notion of effective learnability is analyzed by relating it to the proof-theoretic strength of an axiom system which is used to derive totality proofs for recursive functions. The main result, the generator-predictor theorem, states that an infinite sequence of bits is learnable if the axiom system proves the totality of a recursive function which dominates the time function of the bit sequence generating process. This result establishes a tight connection between learnability and provability, thus reducing the question of what can be effectively learned to the foundational questions of mathematics with regard to set existence axioms. Results of reverse mathematics are used to illustrate the implications of the generator-predictor theorem by connecting a hierarchy of axiom systems with increasing logical strength to fast growing functions. Our results are discussed in the context of the probabilistic universal induction framework pioneered by Solomonoff, showing how the integration of a proof system into the learning process leads to naturally defined effective instances of Solomonoff induction. Finally, we analyze the problem of effective learning in a framework where the time scales of the generator and the predictor are coupled, leading to a surprising conclusion.

A Simple 3-Parameter Model for Cancer Incidences

We propose a simple 3-parameter model that provides very good fits for incidence curves of 18 common solid cancers even when variations due to different locations, races, or periods are taken into account. From a data perspective, we use model selection (Akaike information criterion) to show that this model, which is based on the Weibull distribution, outperforms other simple models like the Gamma distribution. From a modeling perspective, the Weibull distribution can be justified as modeling the accumulation of driver events, which establishes a link to stem cell division based cancer development models and a connection to a recursion formula for intrinsic cancer risk published by Wu et al. For the recursion formula a closed form solution is given, which will help to simplify future analyses. Additionally, we perform a sensitivity analysis for the parameters, showing that two of the three parameters can vary over several orders of magnitude. However, the shape parameter of the Weibull distribution, which corresponds to the number of driver mutations required for cancer onset, can be robustly estimated from epidemiological data.

SwitchFinder – a novel method and query facility for discovering dynamic gene expression patterns

BackgroundBiological systems and processes are highly dynamic. To gain insights into their functioning time-resolved measurements are necessary. Time-resolved gene expression data captures temporal behaviour of the genes genome-wide under various biological conditions: in response to stimuli, during cell cycle, differentiation or developmental programs. Dissecting dynamic gene expression patterns from this data may shed light on the functioning of the gene regulatory system. The present approach facilitates this discovery. The fundamental idea behind it is the following: there are change-points (switches) in the gene behaviour separating intervals of increasing and decreasing activity, whereas the intervals may have different durations. Elucidating the switch-points is important for the identification of biologically meanigfull features and patterns of the gene dynamics.ResultsWe developed a statistical method, called SwitchFinder, for the analysis of time-series data, in particular gene expression data, based on a change-point model. Fitting the model to the gene expression time-courses indicates switch-points between increasing and decreasing activities of each gene. Two types of the model - based on linear and on generalized logistic function - were used to capture the data between the switch-points. Model inference was facilitated with the Bayesian methodology using Markov chain Monte Carlo (MCMC) technique Gibbs sampling. Further on, we introduced features of the switch-points: growth, decay, spike and cleft, which reflect important dynamic aspects. With this, the gene expression profiles are represented in a qualitative manner - as sets of the dynamic features at their onset-times. We developed a Web application of the approach, enabling to put queries to the gene expression time-courses and to deduce groups of genes with common dynamic patterns.SwitchFinder was applied to our original data - the gene expression time-series measured in neuroblastoma cell line upon treatment with all-trans retinoic acid (ATRA). The analysis revealed eight patterns of the gene expression responses to ATRA, indicating the induction of the BMP, WNT, Notch, FGF and NTRK-receptor signaling pathways involved in cell differentiation, as well as the repression of the cell-cycle related genes.ConclusionsSwitchFinder is a novel approach to the analysis of biological time-series data, supporting inference and interactive exploration of its inherent dynamic patterns, hence facilitating biological discovery process. SwitchFinder is freely available at https://newbioinformatics.eu/switchfinder.