Dataset Information

Neutral Zone Classifiers Using a Decision-Theoretic Approach With Application to DNA Array Analyses.

ABSTRACT: Two-class neutral zone classifiers were recently proposed for use in microbial community profiling applications. These classifiers allow a region of neutrality for cases where probe hybridization outcomes are too ambiguous to have adequate confidence in assigning a "binding" or "no binding" result. In this paper, we generalize the idea of neutral zone classifiers to an arbitrary number of classes and apply it to improve the process of microbial community profiling by considering a third class for the outcome of probe hybridization experiments, "partial binding." We introduce a family of class distributions that uses a mixture of Gaussian distributions as a model for a Box-Cox power transformation of the raw intensity measurements. Stratified cross-validation analyses are used to assess the efficacy of the proposed three-class neutral zone classifier. This article has supplementary material online.

SUBMITTER: Yu H

PROVIDER: S-EPMC3137885 | biostudies-literature | 2010 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Neutral Zone Classifiers Using a Decision-Theoretic Approach With Application to DNA Array Analyses.

Yu Hua H Jeske Daniel R DR Ruegger Paul P Borneman James J

Journal of agricultural, biological, and environmental statistics 20101201 4

Two-class neutral zone classifiers were recently proposed for use in microbial community profiling applications. These classifiers allow a region of neutrality for cases where probe hybridization outcomes are too ambiguous to have adequate confidence in assigning a "binding" or "no binding" result. In this paper, we generalize the idea of neutral zone classifiers to an arbitrary number of classes and apply it to improve the process of microbial community profiling by considering a third class fo ...[more]

PMID: 21769245

Similar Datasets

Project description:The human microbiome consists of five major regional biomes distributed in or on our five body sites including skin, oral, lung, gut, and reproductive tract. Its biogeography (the spatial and temporal distribution of its biodiversity) has far-reaching implications to our health and diseases. Nevertheless, we currently have very limited understanding on the mechanisms shaping the biogeography, since it is often rather difficult to determine the relative importance of drift, dispersal, speciation, and selection, the four processes (mechanisms) determining the patterns of microbial biogeography and community dynamics according to a recent synthesis in community ecology and biogeography. To disentangle these mechanisms, I utilize multisite neutral (MSN) model and niche-neutral hybrid (NNH) model to analyze large number of truly multisite microbiome samples covering all five major human microbiome habitats, including 699 metacommunities and 5,420 local communities. Approximately 89% of metacommunities and 92% local communities exhibit patterns indistinguishable from neutral, and 20% indistinguishable from niche-neutral hybrid model, indicating the relative significance of stochastic neutral forces versus deterministic niche selection in shaping the biogeography of human microbiome. These findings cast supporting evidence to van der Gast's revision to classic Bass-Becking doctrine of microbial biogeography: "Some things are everywhere and some things are not. Sometimes the environment selects and sometimes it doesn't," offering the first educated guess for "some" and "sometimes" in the revised doctrine. Furthermore, the logistic/Cox regression models describing the relationships among community neutrality, niche differentiation, and key community/species characteristics (including community diversity, community/species dominance, speciation, and migration rates) were constructed to quantitatively describe the niche-neutral continuum and the influences of community/species properties on the continuum.

Project description:MotivationArtificial intelligence, trained via machine learning (e.g. neural nets, random forests) or computational statistical algorithms (e.g. support vector machines, ridge regression), holds much promise for the improvement of small-molecule drug discovery. However, small-molecule structure-activity data are high dimensional with low signal-to-noise ratios and proper validation of predictive methods is difficult. It is poorly understood which, if any, of the currently available machine learning algorithms will best predict new candidate drugs.ResultsThe quantile-activity bootstrap is proposed as a new model validation framework using quantile splits on the activity distribution function to construct training and testing sets. In addition, we propose two novel rank-based loss functions which penalize only the out-of-sample predicted ranks of high-activity molecules. The combination of these methods was used to assess the performance of neural nets, random forests, support vector machines (regression) and ridge regression applied to 25 diverse high-quality structure-activity datasets publicly available on ChEMBL. Model validation based on random partitioning of available data favours models that overfit and 'memorize' the training set, namely random forests and deep neural nets. Partitioning based on quantiles of the activity distribution correctly penalizes extrapolation of models onto structurally different molecules outside of the training data. Simpler, traditional statistical methods such as ridge regression can outperform state-of-the-art machine learning methods in this setting. In addition, our new rank-based loss functions give considerably different results from mean squared error highlighting the necessity to define model optimality with respect to the decision task at hand.Availability and implementationAll software and data are available as Jupyter notebooks found at https://github.com/owatson/QuantileBootstrap.Supplementary informationSupplementary data are available at Bioinformatics online.

Dataset Information

Neutral Zone Classifiers Using a Decision-Theoretic Approach With Application to DNA Array Analyses.

Publications

Neutral Zone Classifiers Using a Decision-Theoretic Approach With Application to DNA Array Analyses.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets