Phylogenetic inference based on landmark data in 41 empirical data sets

The inference of phylogenetic hypotheses from landmark data has been questioned during the last two decades. Besides theoretical concerns, one of the limitations pointed out for the use of landmark data in phylogenetics is its (supposed) lack of information relevant to the inference of phylogenetic relationships. However, empirical analyses are scarce; there exists no previous study that systematically evaluates the phylogenetic performance of landmark data in a series of data sets. In the present study, we analysed 41 published data sets in order to assess the correspondence between the phylogenetic trees derived from landmark data and those obtained with alternative and independent sources of evidence, and determined the main factors that might affect this inference. The data sets presented a variable number of terminals (5–200) and configurations (1–14), belonging to different taxonomic groups. The results showed that for most of the data sets analysed, the trees derived from landmark data presented a low correspondence with the reference phylogenies. The results were similar irrespective of the phylogenetic method considered. Complementary analyses strongly suggested that the limited amount of evidence included in each data set (one or a few landmark configurations) is the main cause for that low correspondence: the phylogenetic analysis of eight data sets that presented three or more configurations clearly showed that the inclusion of several landmark configurations improves the results. In addition, the analyses indicated that the inclusion of landmark data from different configurations is more important than the inclusion of more landmarks from the same configuration. Based on the results presented here, we consider that the poor results previously obtained in phylogenetic analyses based on landmark data were not caused by methodological limitations, but rather due to the limited amount of evidence included in the data sets.     

Wolf in sheep's clothing: Model misspecification undermines tests of the neutral theory for life histories

Understanding the processes behind change in reproductive state along life‐history trajectories is a salient research program in evolutionary ecology. Two processes, state dependence and heterogeneity, can drive the dynamics of change among states. Both processes can operate simultaneously, begging the difficult question of how to tease them apart in practice. The Neutral Theory for Life Histories (NTLH) holds that the bulk of variations in life‐history trajectories is due to state dependence and is hence neutral: Once previous (breeding) state is taken into account, variations are mostly random. Lifetime reproductive success (LRS), the number of descendants produced over an individual's reproductive life span, has been used to infer support for NTLH in natura. Support stemmed from accurate prediction of the population‐level distribution of LRS with parameters estimated from a state dependence model. We show with Monte Carlo simulations that the current reliance of NTLH on LRS prediction in a null hypothesis framework easily leads to selecting a misspecified model, biased estimates and flawed inferences. Support for the NTLH can be spurious because of a systematic positive bias in estimated state dependence when heterogeneity is present in the data but ignored in the analysis. This bias can lead to spurious positive covariance between fitness components when there is in fact an underlying trade‐off. Furthermore, neutrality implied by NTLH needs a clarification because of a probable disjunction between its common understanding by evolutionary ecologists and its translation into statistical models of life‐history trajectories. Irrespective of what neutrality entails, testing hypotheses about the dynamics of change among states in life histories requires a multimodel framework because state dependence and heterogeneity can easily be mistaken for each other.     

Model misspecification and multipoint linkage analysis.

Pairwise linkage analysis is robust to genetic model misspecification provided dominance is correctly specified, the primary effect being inflation of the recombination fraction. By contrast, we show that multipoint analysis under misspecified models is not robust when a putative disease locus is placed between close flanking markers, with potentially spuriously negative multipoint lod scores being produced. The problem is due to incorrect attribution of segregation of a disease allele and the consequent conclusion of (unlikely) double crossovers between flanking markers. As a possible solution, we propose the use of high disease allele frequencies, as this allows probabilistically for nonsegregation (through parental homozygosity or dual matings). We show analytically and through analysis of pedigree data simulated under a two-locus heterogeneity model that using a disease allele frequency of 0.05 in the dominant case and 0.25 in the recessive case is quite robust in producing positive multipoint lod scores with close flanking markers across a broad range of conditions including varying allele frequencies, epistasis, genetic heterogeneity and phenocopies.     

Access to data and material for research: putting empirical evidence into perspective

The aim of this article is to put into critical perspective the empirical findings on secrecy and withholding in research. In other words, by taking existing empirical literature into account, it is intended that a crucial question is answered: Is secrecy and withholding in research harmful or innocuous to science? To understand how secrecy and withholding in research have affected academic science, empirical studies have been placed in the wider context of Mertonian underpinnings of the anticommons threat. The turning point in testing the effects of secrecy and withholding of data and material on scientific research was marked by statistical studies based on surveys and bibliometric measures. These two types of empirical studies have given answers to the basic question since academia was threatened by different modes of practicing science.     

Application of two tests of multivariate discordancy to fisheries data sets

The generalized (Mahalanobis) distance and multivariate kurtosis are two powerful tests of multivariate discordancies (outliers). Unlike the generalized distance test, the multivariate kurtosis test has not been applied as a test of discordancy to fisheries data heretofore. We applied both tests, along with published algorithms for identifying suspected causal variable(s) of discordant observations, to two fisheries data sets from Lake Erie: total length, mass, and age from 1,234 burbot, Lota lota; and 22 combinations of unique subsets of 10 morphometrics taken from 119 yellow perch, Perca flavescens. For the burbot data set, the generalized distance test identified six discordant observations and the multivariate kurtosis test identified 24 discordant observations. In contrast with the multivariate tests, the univariate generalized distance test identified no discordancies when applied separately to each variable. Removing discordancies had a substantial effect on length-versus-mass regression equations. For 500-mm burbot, the percent difference in estimated mass after removing discordancies in our study was greater than the percent difference in masses estimated for burbot of the same length in lakes that differed substantially in productivity. The number of discordant yellow perch detected ranged from 0 to 2 with the multivariate generalized distance test and from 6 to 11 with the multivariate kurtosis test. With the kurtosis test, 108 yellow perch (90.7%) were identified as discordant in zero to two combinations, and five (4.2%) were identified as discordant in either all or 21 of the 22 combinations. The relationship among the variables included in each combination determined which variables were identified as causal. The generalized distance test identified between zero and six discordancies when applied separately to each variable. Removing the discordancies found in at least one-half of the combinations (k = 5) had a marked effect on a principal components analysis. In particular, the percent of the total variation explained by second and third principal components, which explain shape, increased by 52 and 44% respectively when the discordancies were removed. Multivariate applications of the tests have numerous ecological advantages over univariate applications, including improved management of fish stocks and interpretation of multivariate morphometric data.     

Quantifying change in pelagic plankton network stability and topology based on empirical long-term data

Over the last 34 years, Lake Müggelsee has experienced concurrent warming and nutrient reduction. While the effects of environmental change on single taxonomic or physical–chemical variables have been relatively well researched in isolation, understanding how environmental change propagates through the ecological network remains a major challenge. Capitalizing on the long-term monitoring program of the German Long-Term Ecosystem Research Network site Lake Müggelsee (1979-ongoing), we identified three time periods (1979–1995; 1996–2005; 2006–2013) which differed significantly in phytoplankton biomass and relative plankton community composition. Using multivariate first order autoregressive (MAR1) modeling on 13 pelagic plankton groups and four abiotic variables, we quantified interaction networks and indicators of stability and centrality for each period. Our results suggested that the Müggelsee network was bottom-up regulated in all periods and that stability increased over time. Moreover, in all three networks, non-trophic and indirect interactions appeared to be as commonly present as trophic and direct interactions. Using network centrality measures of betweenness and closeness, we identified keystone plankton groups and groups particularly responsive to environmental change based on variation in centrality ranks over time. Given a more comprehensive understanding of the interaction network at hand, MAR1 model-derived stability and centrality measures may potentially be used as integrated ecological indicators to monitor changes in stability of lake ecosystems and to identify particularly vulnerable components of the network.     

Effects of variance-function misspecification in analysis of longitudinal data

The approach of generalized estimating equations (GEE) is based on the framework of generalized linear models but allows for specification of a working matrix for modeling within-subject correlations. The variance is often assumed to be a known function of the mean. This article investigates the impacts of misspecifying the variance function on estimators of the mean parameters for quantitative responses. Our numerical studies indicate that (1) correct specification of the variance function can improve the estimation efficiency even if the correlation structure is misspecified; (2) misspecification of the variance function impacts much more on estimators for within-cluster covariates than for cluster-level covariates; and (3) if the variance function is misspecified, correct choice of the correlation structure may not necessarily improve estimation efficiency. We illustrate impacts of different variance functions using a real data set from cow growth.     

Transaction costs of community-based forest management: empirical evidence from Tanzania

Levels of transaction costs in community‐based forest management (CBFM) in four communities adjacent to the Ambangulu mountain forests of the north‐east of Tanzania were assessed through questionnaire responses from 120 households. Costs and benefits of CBFM to the rich, medium and poor groups of forest users were estimated. Costs of CBFM were participation in forest monitoring and time spent in meetings. Benefits included forest products consumed at household level. Transaction costs relative to benefits for CBFM were found to be higher for poorer households compared with medium income and richer households. Higher income groups obtained the most net benefits followed by medium and poorer households. Community involvement in forest management may lower the transaction costs incurred by government, but a large proportion of these costs are borne by poorer members of the community. Transaction costs are critical factors in the success or failure of CBFM and need to be incorporated into policies and legislation related to community‐based natural resource management.     

Inferring complex phylogenies using parsimony: an empirical approach using three large DNA data sets for angiosperms

To explore the feasibility of parsimony analysis for large data sets, we conducted heuristic parsimony searches and bootstrap analyses on separate and combined DNA data sets for 190 angiosperms and three outgroups. Separate data sets of 18S rDNA (1,855 bp), rbcL (1,428 bp), and atpB (1,450 bp) sequences were combined into a single matrix 4,733 bp in length. Analyses of the combined data set show great improvements in computer run times compared to those of the separate data sets and of the data sets combined in pairs. Six searches of the 18S rDNA + rbcL + atpB data set were conducted; in all cases TBR branch swapping was completed, generally within a few days. In contrast, TBR branch swapping was not completed for any of the three separate data sets, or for the pairwise combined data sets. These results illustrate that it is possible to conduct a thorough search of tree space with large data sets, given sufficient signal. In this case, and probably most others, sufficient signal for a large number of taxa can only be obtained by combining data sets. The combined data sets also have higher internal support for clades than the separate data sets, and more clades receive bootstrap support of > or = 50% in the combined analysis than in analyses of the separate data sets. These data suggest that one solution to the computational and analytical dilemmas posed by large data sets is the addition of nucleotides, as well as taxa.     

Quinean social skills: empirical evidence from eye-gaze against information encapsulation

Since social skills are highly significant to the evolutionary success of humans, we should expect these skills to be efficient and reliable. For many Evolutionary Psychologists efficiency entails encapsulation: the only way to get an efficient system is via information encapsulation. But encapsulation reduces reliability in opaque epistemic domains. And the social domain is darkly opaque: people lie and cheat, and deliberately hide their intentions and deceptions. Modest modularity [Currie and Sterelny (2000) Philos Q 50:145–160] attempts to combine efficiency and reliability. Reliability is obtained by placing social skills in un-encapsulated central cognition; efficiency by having the social system sensitive to encapsulated socially tagged cues. In this paper, I argue that this approach fails. I focus on eye-gaze as a plausible example of a socially significant encapsulated cue. I demonstrate contra modest modularity that eye-gaze is subject to influence from central cognition.

Latitudinal gradients in niche breadth: empirical evidence from haematophagous ectoparasites

Aim  We searched for relationships between latitude and both the geographic range size and host specificity of fleas parasitic on small mammals. This provided a test for the hypothesis that specialization is lower, and thus niche breadth is wider, in high-latitude species than in their counterparts at lower latitudes.
Location  We used data on the host specificity and geographic range size of 120 Palaearctic flea species (Siphonaptera) parasitic on small mammals (Soricomorpha, Lagomorpha and Rodentia). Data on host specificity were taken from 33 regions, whereas data on geographic ranges covered the entire distribution of the 120 species.
Methods  Our analyses controlled for the potentially confounding effects of phylogenetic relationships among flea species by means of the independent-contrasts method. We used regressions and structural equation modelling to determine whether the latitudinal position of the geographic range of a flea covaried with either the size of its range or its host specificity. The latter was measured as the number of host species used, as well as by an index providing the average (and variance in) taxonomic distinctness among the host species used by a flea.
Results  Geographic range size was positively correlated with the position of the centre of the range; in other words, fleas with more northerly distributions had larger geographic ranges. Although the number of host species used by a flea did not vary with latitude, both the mean taxonomic distinctness among host species used and its variance increased significantly towards higher latitudes.
Main conclusions  The results indicate that niche breadth in fleas, measured in terms of both its spatial (geographic range size) and biological (host specificity) components, increases at higher latitudes. These findings are compatible with the predictions of recent hypotheses about latitudinal gradients.     

Treasures and traps in genome-wide data sets: case examples from yeast

Since the publication of the Saccharomyces cerevisiae genome sequence, much effort has been dedicated to developing high-throughput techniques to generate comprehensive information about the function and dynamics of all genes in this yeast's genome. These techniques have generated data sets that typically contain large amounts of reliable and valuable biological information. Nevertheless, there are also uncertainties that are associated with such large-scale studies, which we discuss in this review. These uncertainties increase with the complexity of the organism under study. On the basis of the results from yeast, we should learn much from human and mouse genomic data sets. However, as with yeast data sets, they might also contain misleading results.     

Computational mate choice: theory and empirical evidence

The present review is based on the thesis that mate choice results from information-processing mechanisms governed by computational rules and that, to understand how females choose their mates, we should identify which are the sources of information and how they are used to make decisions. We describe mate choice as a three-step computational process and for each step we present theories and review empirical evidence. The first step is a perceptual process. It describes the acquisition of evidence, that is, how females use multiple cues and signals to assign an attractiveness value to prospective mates (the preference function hypothesis). The second step is a decisional process. It describes the construction of the decision variable (DV), which integrates evidence (private information by direct assessment), priors (public information), and value (perceived utility) of prospective mates into a quantity that is used by a decision rule (DR) to produce a choice. We make the assumption that females are optimal Bayesian decision makers and we derive a formal model of DV that can explain the effects of preference functions, mate copying, social context, and females' state and condition on the patterns of mate choice. The third step of mating decision is a deliberative process that depends on the DRs. We identify two main categories of DRs (absolute and comparative rules), and review the normative models of mate sampling tactics associated to them. We highlight the limits of the normative approach and present a class of computational models (sequential-sampling models) that are based on the assumption that DVs accumulate noisy evidence over time until a decision threshold is reached. These models force us to rethink the dichotomy between comparative and absolute decision rules, between discrimination and recognition, and even between rational and irrational choice. Since they have a robust biological basis, we think they may represent a useful theoretical tool for behavioural ecologist interested in integrating proximate and ultimate causes of mate choice.     

Hamilton goes empirical: estimation of inclusive fitness from life-history data

Hamilton's theory of kin selection is one of the most important advances in evolutionary biology since Darwin. Central to the kin-selection theory is the concept of inclusive fitness. However, despite the importance of inclusive fitness in evolutionary theory, empirical estimation of inclusive fitness has remained an elusive task. Using the concept of individual fitness, I present a method for estimating inclusive fitness and its components for diploid organisms with age-structured life histories. The method presented here: (i) allows empirical estimation of inclusive fitness from life-history data; (ii) simultaneously considers all components of fitness, including timing and magnitude of reproduction; (iii) is consistent with Hamilton's definition of inclusive fitness; and (iv) adequately addresses shortcomings of existing methods of estimating inclusive fitness. I also demonstrate the application of this new method for testing Hamilton's rule.     

Population size and the risk of local extinction: empirical evidence from rare plants

Due to habitat fragmentation many plant species today occur mainly in small and isolated populations. Modeling studies predict that small populations will be threatened more strongly by stochastic processes than large populations, but there is little empirical evidence to support this prediction for plants. We studied the relationship between size of local populations (number of flowering plants) and survival over ten years for 359 populations of eight short-lived, threatened plants in northern Germany ( Lepidium campestre , Thlaspi perfoliatum , Rhinanthus minor , R . serotinus , Melampyrum arvense , M . nemorosum , Gentianella ciliata and G . germanica ). Overall, 27% of the populations became extinct during the study period. Probability of survival of a local population increased significantly with its size in all but one species ( R. minor ). However, estimated population sizes required for 90% probability of survival over 10 years varied widely among species. Survival probability increased with decreasing distance to the nearest conspecific population in R . serotinus , but not in the other species. The mean annual growth rate of surviving populations differed greatly between species, but was only for G . germanica significantly lower than 1, suggesting that there was no general deterministic decline in the number of plants due to deteriorating habitat conditions. We conclude that the extinction of populations was at least partly due to stochastic processes. This is supported by the fact that in all species a considerable proportion of small populations survived and developed into large populations.     

Matching data sets from two different spatial samples

Abstract. Methods for coupling two data sets (species composition and environmental variables for example) are well known and often used in ecology. All these methods require that variables of the two data sets have been recorded at the same sample stations. But if the two data sets arise from different sample schemes, sample locations can be different. In this case, scientists usually transform one data set to conform with the other one that is chosen as a reference. This inevitably leads to some loss of information. We propose a new ordination method, named spatial‐RLQ analysis, for coupling two data sets with different spatial sample techniques. Spatial‐RLQ analysis is an extension of co‐inertia analysis and is based on neighbourhood graph theory and classical RLQ analysis. This analysis finds linear combinations of variables of the two data sets which maximize the spatial cross‐covariance. This provides a co‐ordination of the two data sets according to their spatial relationships. A vegetation study concerning the forest of Chizé (western France) is presented to illustrate the method.     

Predicting protein function from protein/protein interaction data: a probabilistic approach

MOTIVATION:The development of experimental methods for genome scale analysis of molecular interaction networks has made possible new approaches to inferring protein function. This paper describes a method of assigning functions based on a probabilistic analysis of graph neighborhoods in a protein-protein interaction network. The method exploits the fact that graph neighbors are more likely to share functions than nodes which are not neighbors. A binomial model of local neighbor function labeling probability is combined with a Markov random field propagation algorithm to assign function probabilities for proteins in the network. RESULTS: We applied the method to a protein-protein interaction dataset for the yeast Saccharomyces cerevisiae using the Gene Ontology (GO) terms as function labels. The method reconstructed known GO term assignments with high precision, and produced putative GO assignments to 320 proteins that currently lack GO annotation, which represents about 10% of the unlabeled proteins in S. cerevisiae.