Dissimilarity Matrices Research Articles

Drivers of beta‐diversity variation inBromus erectussemi‐natural dry grasslands

AbstractQuestionsWhat are the main drivers of variation in beta‐diversity forBromus erectussemi‐natural dry grasslands of habitat 6210(*) at different scales? How should environmental variables and spatial patterns be taken into account to conserve the maximum possible beta‐diversity within the habitat?LocationCentralItaly.MethodsWe used 195 vegetation relevés distributed in three nested extents: a single mountain, a mountain chain and southernLazio. Multiple regression on distance matrices was performed using dissimilarity matrices based on: (1) species abundances as response variables; (2) spatial coordinates and environmental parameters (altitude, slope, percentage of rock and stone coverage, aspect, annual rainfall) as explanatory variables. The two groups of explanatory variables were used separately to partition the variation, and jointly to assess the relative contribution of each individual variable. Those variables found to significantly affect beta‐diversity were used to: (1) compare beta‐diversity levels between a set of randomly selected and a set of stratified relevés; and (2) analyse the habitat distribution across environmental gradients. These analyses, together with the curves describing the relationships between spatial distances and composition dissimilarities, were used to inform management decisions for the habitat.ResultsMost of the variance was explained by environmental variables, whose share was higher in the smallest and intermediate extent than in the broadest extent. Community dissimilarity increased in proportion to differences in altitude and spatial distances at every extent. Accordingly, at all the extents, the selection of relevés stratified by altitude or selected taking into account a minimum spatial distance included significantly higher levels of within‐habitat beta‐diversity, than randomly selected relevés. The relation of beta‐diversity to the variation in aspect and annual rainfall varied at different extents.ConclusionsOur results demonstrate that dry grassland management plans aimed at conserving the maximum within‐habitat beta‐diversity should take into account variation in environmental variables, among which altitude proved to be a critical factor at every extent. Also, spatial distances positively affect within‐habitat beta‐diversity levels, and scale‐dependent minimum distances among habitat patches should be taken into account when selecting patches of habitat 6210(*) to be conserved in the study area.

The relationship of tropical bird communities to tree species composition and vegetation structure along an Andean elevational gradient

AbstractAimUnderstanding patterns of species turnover along environmental gradients and their consistency across taxonomic groups is central to the study of biodiversity. We may expect congruence in diversity patterns across groups whose ranges could be influenced by species interactions. We explore associations between bird and vegetation communities in the tropical Andes to determine whether patterns of species richness and turnover in birds and trees are congruent with elevation, and whether tree species composition, vegetation structure, elevation, or a combination of these best predicts bird species composition.LocationA forested 2600‐m elevational gradient on the eastern slope of the Peruvian Andes.MethodsBird surveys and vegetation measurements were conducted at 172 points, and a subset of these were spatially matched with fourteen 1‐ha tree inventory plots. Diversity patterns were described for trees, birds, and avian foraging guilds. We used dissimilarity matrices to examine patterns of turnover and nestedness. Turnover of birds and trees was examined by comparing compositional change of adjacent plots along the gradient. Multiple regression on distance matrices was employed to determine contributions of tree species composition, vegetation structure and elevation to explaining variation in bird species composition.ResultsSpecies richness was higher for trees than for birds, and whereas diversity in both taxa decreased with elevation, tree richness showed a low‐elevation plateau before declining at higher elevations. Tree species had narrower distributions compared to birds, but patterns of turnover were largely congruent between taxa. Nestedness contributed much less to dissimilarity than turnover, although birds showed higher nestedness, particularly at high elevations. Tree species composition, elevation and vegetation structure were all important predictors of bird species composition; the best model explained 78% of bird dissimilarity across plots. Tree species composition was always included in the best models, for all birds and foraging guilds.Main conclusionsOur assessment of Andean bird and vegetation communities suggests strong correspondence, perhaps due to direct interactions or similar underlying drivers. We hypothesize that with climate change, range shifts in these groups may not occur independently. Rather, birds may have delayed upslope shifts or may be limited to high‐elevation patches where appropriate vegetation communities exist.

Pollen–vegetation relationships in the central Caspian (Hyrcanian) forests of northern Iran

We studied pollen–vegetation relationships along an altitudinal transect in the Caspian (Hyrcanian) forests of northern Iran. We collected surface samples from 20 plots of 400 or 625m2 in the major forest communities and compared pollen percentages of the most abundant plant taxa with the vegetation abundances (cover and basal areas) in the pollen plots and in areas of different radii (100, 250, 500 and 1000m) around the pollen plots using both linear regression and qualitative methods. The results show that anemophilous taxa (i.e. Quercus, Fagus, and Carpinus) in general produce more pollen relative to entomophilous taxa (i.e. Acer, Parrotia, Diospyros, and Hedera). Taxa with light pollen grains, such as Quercus and Carpinus, show large positive intercepts in linear regression analysis, while heavier pollen grains or those possessing a patterned exine, such as Diospyros, Parrotia, Acer, and Hedera show small or even negative intercepts. Larger intercepts indicate that a higher proportion of pollen in the samples arrives from outside the area covered by the vegetation analysis. Our findings indicate that light pollen grains are deposited more slowly than heavy pollen grains. Dissimilarity matrices using the Bray–Curtis index showed that the overall relationship between species composition of vegetation plots and pollen assemblages is better for the areas with 100m radius than for the pollen plots.

Which environmental variables should I use in my biodiversity model?

Appropriate selection of environmental variables is critical to the performance of biodiversity models, but has received less attention than the choice of modelling method. Online aggregators of biological and environmental data, such as the Global Biodiversity Information Facility and the Atlas of Living Australia, necessitate a rational approach to variable selection. We outline a set of general principles for systematically identifying, compiling, evaluating and selecting environmental variables for a biodiversity model. Our approach aims to maximise the information obtained from the analysis of biological records linked to a potentially large suite of spatial environmental variables. We demonstrate the utility of this structured framework through case studies with Australian vascular plants: regional modelling of a species distribution, continent-wide modelling of species compositional turnover and environmental classification. The approach is informed by three components of a biodiversity model: (1) an ecological framework or conceptual model, (2) a data model concerning availability, resolution and variable selection and (3) a method for analysing data. We expand the data model in structuring the problem of choosing environmental variables. The case studies demonstrate a structured approach for the: (1) cost-effective compilation of variables in the context of an explicit ecological framework for the study, attribute accuracy and resolution; (2) evaluation of non-linear relationships between variables using knowledge of their derivation, scatter plots and dissimilarity matrices; (3) selection and grouping of variables based on hypotheses of relative ecological importance and perceived predictor effectiveness; (4) systematic testing of variables as predictors through the process of model building and refinement and (5) model critique, inference and synthesis using direct gradient analysis to evaluate the shape of response curves in the context of ecological theory by presenting predictions in both geographic and environmental space.

Using cladistic characters to predict functional variety: experiments using early gnathostomes

ABSTRACT An increasing number of paleobiological studies use preexisting cladistic data sets to estimate shifting levels of morphological disparity over time, and investigate the relative positions of taxa within a multivariate morphospace. Although these patterns are derived from characters selected to reconstruct phylogenetic relationships, it is sometimes suggested that they also reflect functional or ecological variation. We test this assertion using a purpose-built matrix of functional mandibular traits and 24 published cladistic matrices targeting early jawed vertebrates. We find that distance matrices calculated from functional and cladistic data are significantly correlated, but that substantial variation in functional dissimilarity goes unexplained by variation in cladistic dissimilarity. Agreement sequentially improves, but not drastically, when cladistic characters are limited to those pertaining to the skull and then mandible. When dissimilarity matrices are subjected to principal coordinate analysis to yield multivariate morphospaces, we find no clear relationship between phylogenetically independent contrasts for scores on lower functional and cladistic coordinate axes. However, we do find that a majority of variation on the first axis of ordinations based on cladistic characters can be explained by phylogeny, even though the distance matrices underlying these ordinations are phenetic. A resampling procedure reveals a positive, but weak, correlation between cladistic and functional disparity measures for equivalent sets of taxa. Although our results show nonrandom relationships between disparity patterns extracted from functional and cladistic character sets, we conclude that these relationships are not sufficiently strong to allow clear inferences about functional variation to be drawn from data assembled for phylogenetic reconstruction.

Characterization of Freshwater Diatom Communities: Comparing Taxonomic and Genetic-Fingerprinting Approaches.

Benthic diatom assemblages from five sampling sites located on two rivers were characterized simultaneously by means of traditional microscopic observations and PCR-DGGE fingerprinting with primers specifically designed for Bacillariophyceae. Community structure, richness, and diversity assessed by both methods were compared. Diatom lists obtained from morphological identification were separated into subsets, depending on (i) the taxonomic level considered (genus, species, variety) and, for each of them, (ii) the relative abundance (RA) of each component (the whole data set, RA > 1%, RA>2%). These data were then compared to genetic fingerprinting data. Clusters based on taxonomic composition and DGGE banding patterns were very similar, showing good correspondence of community structure between the two methods. Data were compared by linear regressions between indices (richness, diversity) and by Mantel tests on dissimilarity matrices generated for each community composition data set. Statistical analysis indicated that the most reliable correlations with fingerprinting were obtained for genera representing more than 1% RA or species representing more than 2% RA. The results reveal that the PCR-DGGE protocol described here offers a satisfactory alternative for performing preliminary screening of coarse differences in diatom global community structure between samples. It can be regarded as a good complement to taxonomic analyses, which still remain necessary to detect precise changes in richness and diversity, especially when considering species with low abundance in natural assemblages.

Relational partitioning fuzzy clustering algorithms based on multiple dissimilarity matrices

This paper introduces fuzzy clustering algorithms that can partition objects taking into account simultaneously their relational descriptions given by multiple dissimilarity matrices. The aim is to obtain a collaborative role of the different dissimilarity matrices to get a final consensus partition. These matrices can be obtained using different sets of variables and dissimilarity functions. These algorithms are designed to furnish a partition and a prototype for each fuzzy cluster as well as to learn a relevance weight for each dissimilarity matrix by optimizing an adequacy criterion that measures the fit between the fuzzy clusters and their representatives. These relevance weights change at each algorithm iteration and can either be the same for all fuzzy clusters or different from one fuzzy cluster to another. Experiments with real-valued data sets from the UCI Machine Learning Repository as well as with interval-valued and histogram-valued data sets show the usefulness of the proposed fuzzy clustering algorithms.

Harpoon or Bait? A Comparison of Various Metrics in Fishing for Sequence Patterns

The use of sequence analysis in the social sciences has significantly increased during the last decade or two. Sequence analysis explores and describes trajectories and “fishes for patterns” (Abbott, 2000). Many dissimilarity metrics exist in various domains (bioinformatics, data mining, etc.); therefore a crucial and pervasive issue in papers using sequence analysis is robustness. To what extent do the various techniques lead to consistent and converging results? What kinds of patterns are more easily fished out by each of the metrics? Here we propose a systematic comparison of about ten metrics that have been used in the social science literature, based on the examination of dissimilarity matrices computed from a simulated sequence data set including various patterns that sociologists can try to identify. This should help scholars in picking the method best suited to their data design and inquiry objectives.

Voice identity discrimination in schizophrenia

Voices provide a wealth of socially-relevant information, including cues to a speaker's identity and emotion. Deficits recognising emotion from voice have been extensively described in schizophrenia, and linked specifically to auditory hallucinations (AH), but relatively little attention has been given to examining the ability to analyse speaker identity. Hence, the current study assessed the ability to discriminate between different speakers in people with schizophrenia (including 33 with and 32 without AH) compared to 32 healthy controls. Participants rated the degree of perceived identity similarity of pairs of unfamiliar voices pronouncing three-syllable words. Multidimensional scaling of the dissimilarity matrices was performed and the resulting dimensions were interpreted, a posteriori, via correlations with acoustic measures relevant to voice identity. A two-dimensional perceptual space was found to be appropriate for both schizophrenia patients and controls, with axes corresponding to the average fundamental frequency (F0) and formant dispersion (Df). Patients with schizophrenia did not differ from healthy controls in their reliance on F0 in differentiating voices, suggesting that the ability to use pitch-based cues for discriminating voice identity may be relatively preserved in schizophrenia. On the other hand, patients (both with and without AH) made less use of Df in discriminating voices compared to healthy controls. This distorted pattern of responses suggests some potentially important differences in voice identity processing in schizophrenia. Formant dispersion has been linked to perceptions of dominance, masculinity, size and age in healthy individuals. These findings open some interesting new directions for future research.

A Cluster Validity Framework Based on Induced Partition Dissimilarity

We describe a new cluster validity framework (CVF) that compares structure in the data (in dissimilarity form) to the structure of dissimilarity matrices induced by a matrix transformation of the partition being tested. As part of this framework, we show two possible cluster validation measures: one, visual cluster validity, that that uses visual comparison and another one, correlation cluster validity, based on correlation. Unlike many existing measures, the measures we propose can be applied to crisp or soft partitions obtained by any relational or object data clustering algorithm. We illustrate the new measures and compare them to several well-known existing measures using real and artificial data sets.

Genetic Diversity and Population Structure in a European Collection of Rice

ABSTRACTIn southern Europe, rice (Oryza sativa L.) is grown as an irrigated crop in river deltas where it plays an important role in the agroecological equilibrium through soil desalinization. In these regions, rice is at the northern limit of its natural area of adaptation. Special cultivars are needed for these challenging conditions. Using model‐based and distance‐based approaches, we analyzed the genetic structure of the European Rice Germplasm Collection (ERGC), which is composed of 425 accessions, using 25 simple sequence repeat (SSR) markers. We compared it with a reference set of 50 accessions that are representative of the diversity of O. sativa. Most of the ERGC accessions (89%) clustered with japonica types. The ERGC japonica accessions were classified into three groups: one group close to rice types of tropical origin that are found in the United States and Argentina and two groups of temperate origin showing less differentiation. The three japonica groups could be characterized according to their grain type and maturity class, which are the most strongly selected traits in European breeding programs. We extracted a core collection of 250 japonica accessions and characterized it using 70 single nucleotide polymorphisms (SNPs). The SSR and SNP dissimilarity matrices coincided reasonably well and for the best‐supported structure, the percentages of admixture were highly correlated. The core collection can be used as an association panel to search for alleles of interest for temperate areas or as a training population for genomic selection.

Analysis of barcode sequences by means of compression-based methods

Motivations. The key idea of DNA barcode initiative is to identify, for each group of species belonging to different kingdoms of life, a short DNA sequence that can act as a true taxon barcode. DNA barcode represents a valuable type of information that can be integrated with ecological, genetic, and morphological data in order to obtain a more consistent taxonomy. Recent studies have shown that, for the animal kingdom, the mitochondrial gene cytochrome c oxidase I (COI), about 650 bp long, can be used as a barcode sequence for identification and taxonomic purposes of animals. The analysis of DNA barcode sequences is carried out with well known bioinformatics techniques: for example the most common approach to create a phylogenetic tree for a group of species uses the multi-alignment of genetic sequences, the computation of a dissimilarity matrix, using one of the current available evolutionary distance model, and finally the building of a tree by means of hierarchical algorithms such as Unweighted Pair Group Method with Arithmetic Mean (UPGMA) and Neighbor Joining (NJ). In the present work we aims at introducing the use of an alignment-free approach in order to make taxonomic analysis of barcode sequences. Our approach is based on the use of two compression-based versions of non-computable Universal Similarity Metric (USM) class of distances. This way we try to overcome some flaws of classic techniques, such as the time-consuming and parameter-dependent alignment procedure and the use of stochastic evolutionary distance models, that do not represent a distance metric. Methods. Universal Similarity Metric represents a class of distance measures based on the non-computable Kolmogorov complexity. That means it needs some approximation in order to be used. USM is said to be "universal" because it can be applied for computing a distance matrix among input data belonging to very different application domains. In fact it has been used for the analysis of text, images, music. In bioinformatics, it has been applied for obtaining phylogenetic trees from complete mythocondrial genome of mammalian species. Our purpose is to justify the employ of USM also for the analysis of short DNA barcode sequences, showing USM is able to correctly extract taxonomic information among those kind of sequences. We, then, downloaded from Barcode of Life data System database (BOLD) 20 datasets of barcode sequences belonging to different animal species. For each dataset we computed dissimilarity matrices by means of two compression-based approximation of USM, namely Normalized Compression Distance (NCD) and its conditional compression version. In both cases we used GenCompres compressor, that is a dictionary-based compressor suited to work with DNA sequences. From those matrices we built, using UPGMA and NJ algorithms, phylogenetic trees of every dataset and compared them, in terms of topology preservation, with the trees obtained through Kimura 2-parameter evolutionary distance model. Results. Experimental tests aim to evaluate the quality of phylogenetic reconstruction in terms of both topological similarity and differences in the relative branch length. As regard the tree similarity, we obtain good results with a percentage of similarity between evolutionary and compression-based tree greater than 82% and, for the most of datasets, between 90% and 100%. Lower results are for datasets having an high percentage of sequences with ambiguous bases. We detect the same trend for differences in the relative branch of trees, except that poorer results are reached by those datasets containing some COI-5P gene sequences longer than the other ones.

Cross-dialect and cross-language differences in perception of vowels: A multidimensional scaling study

This study examines the perceptual responses to vowels in two regional varieties of American English typical of Central Ohio (OH) and Western North Carolina (NC). Listeners are monolingual local speakers of each dialect and Mandarin-English bilinguals in Columbus OH. The questions are (1) how English listeners perceive acoustic variations in vowel quality differences not found in their own dialect and (2) whether bilingual listeners are sensitive to such differences not found in their native language. In a multidimensional scaling study, we investigate whether listeners make these perceptual decisions on the basis of acoustic properties or categorical properties of these vowels. Each listener was presented with two sets of 13 vowels /i ɪ e ɛ æ u ʊ o ɔ ɑ ɔɪ aɪ aʊ/ in a /hVd/ syllable. One set was produced by an OH male speaker and the other by a NC male speaker. Listeners rated all possible vowel pairs from one set only on a nine-point similarity/dissimilarity scale (there were two subsets of listeners in each group). The resulting dissimilarity matrices were analyzed using INDSCAL. Results will be discussed in terms of differences in perceptual dimensions (vowel coordinates and subject weights) as a function of dialect and L2 background.

Modelling niche and neutral dynamics: on the ecological interpretation of variation partitioning results

Many studies have attempted to disentangle the effects of neutral and niche‐mediated processes on community composition using partial Mantel tests and variance partitioning on dissimilarity matrices. Recently, doubts about the reliability of these methods have emerged. Here we explore how the results are affected by three confounding factors that may affect ecological data to different degrees: spatial autocorrelation of the environmental variables, length of the compositional gradient, and sampling noise. We document that the statistical hypotheses tested in these methods may or may not correspond to the ecological hypotheses of interest. A major discrepancy emerges if a large proportion of sampling units in the analysed dataset share no species, in which case compositional dissimilarities saturate to a fixed maximum value although explanatory dissimilarities do not. With increasing dissimilarity saturation, the explanatory power of regression models decrease, which may lead to the erroneous conclusion that the ecological processes represented by the explanatory variables are not operating. A survey of recent literature suggests that there is a general lack of awareness of this problem, although it appears to affect > 10% of relevant studies. Our simulations show that if dissimilarity saturation is due to a long ecological gradient, extended dissimilarities essentially solve the problem for any degree of saturation. Using distances from a hybrid multidimensional scaling alleviates the saturation problem when the degree of saturation is < 60%. However, neither correction method can provide a solution to problems caused by insufficient sampling. How the presence of multiple explanatory gradients in combination with sampling noise affects overall analysis performance remains to be clarified.

The Representation of Biological Classes in the Human Brain

Evidence of category specificity from neuroimaging in the human visual system is generally limited to a few relatively coarse categorical distinctions-e.g., faces versus bodies, or animals versus artifacts-leaving unknown the neural underpinnings of fine-grained category structure within these large domains. Here we use fMRI to explore brain activity for a set of categories within the animate domain, including six animal species-two each from three very different biological classes: primates, birds, and insects. Patterns of activity throughout ventral object vision cortex reflected the biological classes of the stimuli. Specifically, the abstract representational space-measured as dissimilarity matrices defined between species-specific multivariate patterns of brain activity-correlated strongly with behavioral judgments of biological similarity of the same stimuli. This biological class structure was uncorrelated with structure measured in retinotopic visual cortex, which correlated instead with a dissimilarity matrix defined by a model of V1 cortex for the same stimuli. Additionally, analysis of the shape of the similarity space in ventral regions provides evidence for a continuum in the abstract representational space-with primates at one end and insects at the other. Further investigation into the cortical topography of activity that contributes to this category structure reveals the partial engagement of brain systems active normally for inanimate objects in addition to animate regions.

Genomic diversity among sorghum genotypes with resistance to sorghum shoot fly, Atherigona soccata

Host plant resistance is one of the important components for management of sorghum shoot fly, Atherigona soccata. The levels of resistance in cultivated germplasm are low to moderate, and therefore, it is important to identify sorghum genotypes with diverse mechanisms of resistance based on physico-chemical and or molecular markers. We assessed the genetic diversity of 15 sorghum genotypes with different levels of resistance/susceptibility to shoot fly, A. soccata using 93 sorghum simple sequence repeat (SSR) primer pairs and simultaneously characterized for 15 morpho-biochemical traits associated with shoot fly resistance. Of these 93 SSR primer pairs, amplification products from 79, thought to correspond to single-copy loci distributed across all ten sorghum chromosome pairs, showed good polymorphism across the 15 sorghum genotypes. The polymorphic information content (PIC) values of these 79 SSR markers ranged from 0.06 to 0.86. The Principal Coordinate Analyses (PCoA) and cluster analyses based on dissimilarity matrices derived from SSR based allelic variation (Neighbor-Joining distance) and variation in 15 morpho-biochemical traits (based on Gower’s distance), revealed grouping of most susceptible genotypes in single cluster. The improved breeding lines grouped with resistant or susceptible genotypes, based on shared pedigree. Based on these results, three resistant accessions viz., IS 1054, IS 1057 and IS 4664 were found diverse to IS 18551, which is widely used as shoot fly resistance donor. These diverse sources, after further characterization for resistance mechanisms, can be used in breeding programs for improving shoot fly resistance.

Inverse MDS: Inferring Dissimilarity Structure from Multiple Item Arrangements

The pairwise dissimilarities of a set of items can be intuitively visualized by a 2D arrangement of the items, in which the distances reflect the dissimilarities. Such an arrangement can be obtained by multidimensional scaling (MDS). We propose a method for the inverse process: inferring the pairwise dissimilarities from multiple 2D arrangements of items. Perceptual dissimilarities are classically measured using pairwise dissimilarity judgments. However, alternative methods including free sorting and 2D arrangements have previously been proposed. The present proposal is novel (a) in that the dissimilarity matrix is estimated by “inverse MDS” based on multiple arrangements of item subsets, and (b) in that the subsets are designed by an adaptive algorithm that aims to provide optimal evidence for the dissimilarity estimates. The subject arranges the items (represented as icons on a computer screen) by means of mouse drag-and-drop operations. The multi-arrangement method can be construed as a generalization of simpler methods: It reduces to pairwise dissimilarity judgments if each arrangement contains only two items, and to free sorting if the items are categorically arranged into discrete piles. Multi-arrangement combines the advantages of these methods. It is efficient (because the subject communicates many dissimilarity judgments with each mouse drag), psychologically attractive (because dissimilarities are judged in context), and can characterize continuous high-dimensional dissimilarity structures. We present two procedures for estimating the dissimilarity matrix: a simple weighted-aligned-average of the partial dissimilarity matrices and a computationally intensive algorithm, which estimates the dissimilarity matrix by iteratively minimizing the error of MDS-predictions of the subject’s arrangements. The Matlab code for interactive arrangement and dissimilarity estimation is available from the authors upon request.

Three notions of tropical rank for symmetric matrices

We introduce and study three different notions of tropical rank for symmetric and dissimilarity matrices in terms of minimal decompositions into rank 1 symmetric matrices, star tree matrices, and tree matrices. Our results provide a close study of the tropical secant sets of certain nice tropical varieties, including the tropical Grassmannian. In particular, we determine the dimension of each secant set, the convex hull of the variety, and in most cases, the smallest secant set which is equal to the convex hull.

Too good to be true: pitfalls of using mean Ellenberg indicator values in vegetation analyses

AbstractQuestionMean Ellenberg indicator values (EIVs) inherit information about compositional similarity, because during their calculation species abundances (or presence–absences) are used as weights. Can this similarity issue actually be demonstrated, does it bias results of vegetation analyses correlating mean EIVs with other aspects of species composition and how often are biased studies published?MethodsIn order to separate information on compositional similarity possibly present in mean EIVs, a new variable was introduced, calculated as a weighted average of randomized species EIVs. The performance of these mean randomized EIVs was compared with that of the mean real EIVs on the one hand and random values (randomized mean EIVs) on the other. To demonstrate the similarity issue, differences between samples were correlated with dissimilarity matrices based on various indices. Next, the three mean EIV variables were tested in canonical correspondence analysis (CCA), detrended correspondence analysis (DCA), analysis of variance (ANOVA) between vegetation clusters, and in regression on species richness. Subsequently, a modified permutation test of significance was proposed, taking the similarity issue into account. In addition, an inventory was made of studies published in the Journal of Vegetation Science and Applied Vegetation Science between 2000 and 2010 likely reporting biased results due to the similarity issue.ResultsUsing mean randomized EIVs, it is shown that compositional similarity is inherited into mean EIVs and most resembles the inter‐sample distances in correspondence analysis, which itself is based on iterative weighted averaging. The use of mean EIVs produced biased results in all four analysis types examined: unrealistic (too high) explained variances in CCA, too many significant correlations with ordination axes in DCA, too many significant differences between cluster analysis groups and too high coefficients of determination in regressions on species richness. Modified permutation tests provided ecologically better interpretable results. From 95 studies using Ellenberg indicator values, 36 reported potentially biased results.ConclusionsNo statistical inferences should be made in analyses relating mean EIVs with other variables derived from the species composition as this can produce highly biased results, leading to misinterpretation. Alternatively, a modified permutation test using mean randomized EIVs can sometimes be used.

Integração de dados em estudos de diversidade genética de tomateiro

O objetivo deste trabalho foi integrar dados de caracteres quantitativos, multicategóricos, moleculares e fitopatológicos para a avaliação da diversidade genética de subamostras de tomateiro do Banco de Germoplasma de Hortaliças da Universidade Federal de Viçosa (BGH-UFV). Foram utilizados dados de 67 subamostras de tomateiro do BGH-UFV, caracterizadas quanto a 19 caracteres quantitativos, 30 multicategóricos, 52 locos ISSR e à reação a três patógenos (Alternaria solani, Pseudomonas syringae pv. tomato e Tomato yellow spot virus). Inicialmente, a avaliação da diversidade entre as subamostras foi realizada para cada conjunto de caracteres individualmente, e indicou que a diversidade baseada em qualquer um dos conjuntos de dados não reflete a diversidade dos demais. Para a integração dos dados, codificaram-se os de natureza quantitativa em multicategóricos, por meio de cinco estratégias diferentes. A estratégia de divisão equitativa da amplitude dos dados em três classes foi a mais indicada, com correlação de 0,78 entre as matrizes de dissimilaridade dos dados codificados e originais. A análise de diversidade genética a partir da integração dos dados resultou em grupos com maior correspondência às origens das subamostras de tomateiro avaliadas, o que indica que a integração de dados de diferentes naturezas pode ser realizada com êxito pela conversão dos dados quantitativos em multicategóricos.