Collective Inference Research Articles

A Network Inference Workflow Applied to Virulence-Related Processes in Salmonella typhimurium

Inference of the structure of mRNA transcriptional regulatory networks, protein regulatory or interaction networks, and protein activation/inactivation-based signal transduction networks are critical tasks in systems biology. In this article we discuss a workflow for the reconstruction of parts of the transcriptional regulatory network of the pathogenic bacterium Salmonella typhimurium based on the information contained in sets of microarray gene-expression data now available for that organism and describe our results obtained by following this workflow. The primary tool is one of the network-inference algorithms deployed in the Software Environment for Biological Network Inference (SEBINI). Specifically, we selected the algorithm called context likelihood of relatedness (CLR), which uses the mutual information contained in the gene-expression data to infer regulatory connections. The associated analysis pipeline automatically stores the inferred edges from the CLR runs within SEBINI and, upon request, transfers the inferred edges into either Cytoscape or the plug-in Collective Analysis of Biological Interaction Networks (CABIN) tool for further postanalysis of the inferred regulatory edges. The following article presents the outcome of this workflow, as well as the protocols followed for microarray data collection, data cleansing, and network inference. Our analysis revealed several interesting interactions, functional groups, metabolic pathways, and regulons in S. typhimurium.

Classification in Networked Data: A Toolkit and a Univariate Case Study

This paper is about classifying entities that are interlinked with entities for which the class is known. After surveying prior work, we present NetKit, a modular toolkit for classification in networked data, and a case-study of its application to networked data used in prior machine learning research. NetKit is based on a node-centric framework in which classifiers comprise a local classifier, a relational classifier, and a collective inference procedure. Various existing node-centric relational learning algorithms can be instantiated with appropriate choices for these components, and new combinations of components realize new algorithms. The case study focuses on univariate network classification, for which the only information used is the structure of class linkage in the network (i.e., only links and some class labels). To our knowledge, no work previously has evaluated systematically the power of class-linkage alone for classification in machine learning benchmark data sets. The results demonstrate that very simple network-classification models perform quite well---well enough that they should be used regularly as baseline classifiers for studies of learning with networked data. The simplest method (which performs remarkably well) highlights the close correspondence between several existing methods introduced for different purposes---that is, Gaussian-field classifiers, Hopfield networks, and relational-neighbor classifiers. The case study also shows that there are two sets of techniques that are preferable in different situations, namely when few versus many labels are known initially. We also demonstrate that link selection plays an important role similar to traditional feature selection.

A bias/variance decomposition for models using collective inference

Bias/variance analysis is a useful tool for investigating the performance of machine learning algorithms. Conventional analysis decomposes loss into errors due to aspects of the learning process, b...

Implicações da utilização do conceito de risco para a prática em APS

A produção do conhecimento e as ações de saúde têm relação direta com as bases conceituais e metodológicas dadas pela Epidemiologia. Três temas relevantes do conceito de risco são problematizados neste texto: as construções de inferências coletivas a partir da apreciação de agregados; a necessidade de redução da complexidade dos fenômenos estudados e seu isolamento de influências externas; o descarte da subjetividade na construção dos modelos que prezam pela objetividade, linearidade e rigidez. As implicações deste pensamento científico são analisadas em suas potencialidades e, especialmente, em suas limitações para a prática clínica na Atenção Primária em Saúde.

Heterogeneity in fecundability studies: issues and modelling

Modelization of fecundability stepped recently from demography and population-based contexts to reproductive biology and treatment of infertility. This created a strong call for flexibility and robustness. Indeed, explained and unexplained heterogeneities are non-negligible sources of bias that result in false conclusions as to the determinants of fertility or to the success rates of reproductive techniques, among other examples. There are two main sources of heterogeneity: biological heterogeneity and heterogeneity of sexual behaviour. A uniform presentation of time-to-pregnancy and Barrett-Marshall models is proposed to enlighten their similarities and differences in modelling heterogeneity of fecundability. Mixed models for fecundability studies are presented as tools to allow for unexplained heterogeneity and to quantify heterogeneity of the effect of observed factors and variability of size of this unexplained heterogeneity between subpopulations. Some criteria for the modelling strategy in fecundability studies are suggested with emphasis on the unit-treatment additivity criterion. The strong and complex selection process resulting from heterogeneity is described as well as the selection and cross-selection processes of observed and unobserved fecundability factors. Consequences regarding data collection and statistical inference are discussed. In the current context, a consensus setting general rules for data collection and statistical analysis would be useful to compare the results and increase the reliability of these results in medical practice.

Glaciological reconstruction of the Laurentide Ice Sheet: physical processes and modelling challenges

Current understanding of Pleistocene ice-sheet history is based on collective inferences from three separate avenues of study: (1) the geologic and paleoceanographic records, (2) the isostatic record, and (3) the behaviour of contemporary glaciers and ice sheets. The geologic record provides good constraint on the areal extent of former ice sheets, while isostatic deflection patterns provide important information about late-glacial ice-sheet thickness. The picture emerging from geologic and isostatic deductions is suggestive of a thin and mobile Laurentide Ice Sheet relative to present-day Greenland and Antarctica. We model Laurentide Ice Sheet evolution through a glacial cycle to explore the glaciological mechanisms that are required to replicate the geologic and isostatic evidence. A number of glaciological processes important to the ice-sheet evolution are not fully understood, including marine-based ice dynamics, iceberg calving, rheologic properties of ice, and basal flow dynamics. We present a spectrum of glacial cycle simulations with different treatments of poorly constrained physical processes. We conclude that glaciological model reconstructions can only be reconciled with the late-glacial geologic record of a thin, low-sloping Laurentide Ice Sheet by invoking (1) extremely deformable ice, (2) widespread basal flow, or (3) paleoclimate-ice-sheet fluctuations which give last glacial maximum ice sheets that are far from equilibrium.

Race in the making: cognition, culture, and the child's construction of human kinds

Representing race - universal and comparative perspectives mining history for psychological wisdom - rethinking racial thinking domain specificity and the study of race do children have a theory of race? race, language, and collect-ive inference the appearance of race - perception in the construction of racial categories the cultural biology of race.

A Critique of Steven Lukes' `Power: A Radical View'

Steven Lukes' monograph Power: A Radical View (Lukes 1974) offers a critical analysis of pluralist and nondecision theories of power, leading on to a suggested `three-dimensional view' of power. This view stresses exercises of power that (a) do not entail observable conflicts, but rather latent conflicts; (b) are a `function of collective forces and social arrangements' (p. 22). In particular, Lukes advances the self-confessedly problematic ideas that such exercises may (1) involve inaction; (2) be unconscious; (3) be wielded by identifiable groups or institutions. However, there are difficulties in Lukes' `real interests' approach with its claims to be both `empirically applicable' and `essentially contested'. Second, Lukes' acceptance of the orientation of the pluralist and nondecision writers as the point of departure in his attempt to supersede them creates a fundamental antimony between his individualist illustrative methodology and his collective inferences. Third, flaws in propositions (1) and (2) above are highlighted. Finally it is suggested that Lukes' position in the structural determinism versus elite voluntarism argument between Poulantzas and Miliband conflicts violently with (3) above, so dividing the essay against itself.