Adele Cutler

Random Forests for Classification in Ecology

Classification procedures are some of the most widely used statistical methods in ecology. Random forests (RF) is a new and powerful statistical classifier that is well established in other disciplines but is relatively unknown in ecology. Advantages of RF compared to other statistical classifiers include (1) very high classification accuracy; (2) a novel method of determining variable importance; (3) ability to model complex interactions among predictor variables; (4) flexibility to perform several types of statistical data analysis, including regression, classification, survival analysis, and unsupervised learning; and (5) an algorithm for imputing missing values. We compared the accuracies of RF and four other commonly used statistical classifiers using data on invasive plant species presence in Lava Beds National Monument, California, USA, rare lichen species presence in the Pacific Northwest, USA, and nest sites for cavity nesting birds in the Uinta Mountains, Utah, USA. We observed high classification accuracy in all applications as measured by cross-validation and, in the case of the lichen data, by independent test data, when comparing RF to other common classification methods. We also observed that the variables that RF identified as most important for classifying invasive plant species coincided with expectations based on the literature.

An application of Random Forests to a genome-wide association dataset: Methodological considerations & new findings

BackgroundAs computational power improves, the application of more advanced machine learning techniques to the analysis of large genome-wide association (GWA) datasets becomes possible. While most traditional statistical methods can only elucidate main effects of genetic variants on risk for disease, certain machine learning approaches are particularly suited to discover higher order and non-linear effects. One such approach is the Random Forests (RF) algorithm. The use of RF for SNP discovery related to human disease has grown in recent years; however, most work has focused on small datasets or simulation studies which are limited.ResultsUsing a multiple sclerosis (MS) case-control dataset comprised of 300 K SNP genotypes across the genome, we outline an approach and some considerations for optimally tuning the RF algorithm based on the empirical dataset. Importantly, results show that typical default parameter values are not appropriate for large GWA datasets. Furthermore, gains can be made by sub-sampling the data, pruning based on linkage disequilibrium (LD), and removing strong effects from RF analyses. The new RF results are compared to findings from the original MS GWA study and demonstrate overlap. In addition, four new interesting candidate MS genes are identified, MPHOSPH9, CTNNA3, PHACTR2 and IL7, by RF analysis and warrant further follow-up in independent studies.ConclusionsThis study presents one of the first illustrations of successfully analyzing GWA data with a machine learning algorithm. It is shown that RF is computationally feasible for GWA data and the results obtained make biologic sense based on previous studies. More importantly, new genes were identified as potentially being associated with MS, suggesting new avenues of investigation for this complex disease.

The transmission disequilibrium test suggests that HLA-DR4 and DR13 are linked to autism spectrum disorder.

We have evaluated possible contributions of HLA-DRB1 alleles to autism spectrum disorder (ASD) in 103 families of Caucasian descent. The DR4 allele occurred more often in probands than controls (0.007), whereas the DR13,14 alleles occurred less often in probands than controls (p = 0.003). The transmission disequilibrium test (TDT) indicated that the ASD probands inherited the DR4 allele more frequently than expected (p = 0.026) from the fathers. The TDT also revealed that fewer DR13 alleles than expected were inherited from the mother by ASD probands (p = 0.006). We conclude that the TDT results suggest that DR4 and DR13 are linked to ASD. Reasons for the parental inheritance of specific alleles are poorly understood but coincide with current genetic research noting possible parent-of-origin effects in autism.

Prospective study of Dietary Approaches to Stop Hypertension– and Mediterranean-style dietary patterns and age-related cognitive change: the Cache County Study on Memory, Health and Aging

BACKGROUND Healthy dietary patterns may protect against age-related cognitive decline, but results of studies have been inconsistent. OBJECTIVE We examined associations between Dietary Approaches to Stop Hypertension (DASH)- and Mediterranean-style dietary patterns and age-related cognitive change in a prospective, population-based study. DESIGN Participants included 3831 men and women ≥65 y of age who were residents of Cache County, UT, in 1995. Cognitive function was assessed by using the Modified Mini-Mental State Examination (3MS) ≤4 times over 11 y. Diet-adherence scores were computed by summing across the energy-adjusted rank-order of individual food and nutrient components and categorizing participants into quintiles of the distribution of the diet accordance score. Mixed-effects repeated-measures models were used to examine 3MS scores over time across increasing quintiles of dietary accordance scores and individual food components that comprised each score. RESULTS The range of rank-order DASH and Mediterranean diet scores was 1661-25,596 and 2407-26,947, respectively. Higher DASH and Mediterranean diet scores were associated with higher average 3MS scores. People in quintile 5 of DASH averaged 0.97 points higher than those in quintile 1 (P = 0.001). The corresponding difference for Mediterranean quintiles was 0.94 (P = 0.001). These differences were consistent over 11 y. Higher intakes of whole grains and nuts and legumes were also associated with higher average 3MS scores [mean quintile 5 compared with 1 differences: 1.19 (P < 0.001), 1.22 (P < 0.001), respectively]. CONCLUSIONS Higher levels of accordance with both the DASH and Mediterranean dietary patterns were associated with consistently higher levels of cognitive function in elderly men and women over an 11-y period. Whole grains and nuts and legumes were positively associated with higher cognitive functions and may be core neuroprotective foods common to various healthy plant-centered diets around the globe.

A deterministic algorithm for global optimization

We present an algorithm for finding the global maximum of a multimodal, multivariate function for which derivatives are available. The algorithm assumes a bound on the second derivatives of the function and uses this to construct an upper envelope. Successive function evaluations lower this envelope until the value of the global maximum is known to the required degree of accuracy. The algorithm has been implemented in RATFOR and execution times for standard test functions are presented at the end of the paper.

Information Ratios for Validating Mixture Analyses

Abstract Determining the number of components in a mixture of distributions is an important but difficult problem. This article introduces a procedure called minimum information ratio estimation and validation (MIREV), which is based on a ratio of Fisher information matrices. The smallest eigenvalue of the information ratio matrix is used to determine the number of components. A measure of uncertainty may be obtained using a bootstrap technique. Simulations illustrate the effectiveness of the procedure. For mixtures of exponential families, an expression for the observed information ratio matrix provides insight to the success of the procedure. Cluster analysis attempts to identify and characterize subpopulations believed to be present in a population. A wide variety of methods, are available, including criterion optimization, hierarchical methods, and various heuristic methods. Criterion optimization techniques, such as mixture analysis, fuzzy clustering, and partitioning methods are popular because they...

DNA Macroarray Profiling of Lactococcus lactis subsp. lactis IL1403 Gene Expression during Environmental Stresses

ABSTRACT This report describes the use of an oligonucleotide macroarray to profile the expression of 375 genes in Lactococcus lactis subsp. lactis IL1403 during heat, acid, and osmotic stress. A set of known stress-associated genes in IL1403 was used as the internal control on the array. Every stress response was accurately detected using the macroarray, compared to data from previous reports. As a group, the expression patterns of the investigated metabolic genes were significantly altered by heat, acid, and osmotic stresses. Specifically, 13 to 18% of the investigated genes were differentially expressed in each of the environmental stress treatments. Interestingly, the methionine biosynthesis pathway genes (metA-metB1 and metB2-cysK) were induced during heat shock, but methionine utilization genes, such as metK, were induced during acid stress. These data provide a possible explanation for the differences between acid tolerance mechanisms of L. lactis strains IL1403 and MG1363 reported previously. Several groups of transcriptional responses were common among the stress treatments, such as repression of peptide transporter genes, including the opt operon (also known as dpp) and dtpT. Reduction of peptide transport due to environmental stress will have important implications in the cheese ripening process. Although stress responses in lactococci were extensively studied during the last decade, additional information about this bacterium was gained from the use of this metabolic array.

Minimum Hellinger distance estimation for finite mixture models

Abstract Minimum Hellinger distance estimates are considered for finite mixture models when the exact forms of the component densities are unknown in detail but are thought to be close to members of some parametric family. Minimum Hellinger distance estimates are asymptotically efficient if the data come from a member of the parametric family and are robust to certain departures from the parametric family. A new algorithm is introduced that is similar to the EM algorithm a specialized adaptive density estimate is also introduced. Standard measures of robustness are discussed some difficulties are noted. The robustness and asymptotic efficiency of the estimators are illustrated using simulations.

Random forests for microarrays.

Random Forests is a powerful multipurpose tool for predicting and understanding data. If gene expression data come from known groups or classes (e.g., tumor patients and controls), Random Forests can rank the genes in terms of their usefulness in separating the groups. When the groups are unknown, Random Forests uses an intrinsic measure of the similarity of the genes to extract useful multivariate structure, including clusters. This chapter summarizes the Random Forests methodology and illustrates its use on freely available data sets.

Information-Based Validity Functionals for Mixture Analysis

Model identification is an essential but often neglected component of sound statistical inference. In mixture analysis, data are assumed to be sampled from a distribution with density