The necessity of connection structures in neural models of variable binding

In his review of neural binding problems, Feldman (Cogn Neurodyn 7:1–11, 2013) addressed two types of models as solutions of (novel) variable binding. The one type uses labels such as phase synchrony of activation. The other (‘connectivity based’) type uses dedicated connections structures to achieve novel variable binding. Feldman argued that label (synchrony) based models are the only possible candidates to handle novel variable binding, whereas connectivity based models lack the flexibility required for that. We argue and illustrate that Feldman’s analysis is incorrect. Contrary to his conclusion, connectivity based models are the only viable candidates for models of novel variable binding because they are the only type of models that can produce behavior. We will show that the label (synchrony) based models analyzed by Feldman are in fact examples of connectivity based models. Feldman’s analysis that novel variable binding can be achieved without existing connection structures seems to result from analyzing the binding problem in a wrong frame of reference, in particular in an outside instead of the required inside frame of reference. Connectivity based models can be models of novel variable binding when they possess a connection structure that resembles a small-world network, as found in the brain. We will illustrate binding with this type of model with episode binding and the binding of words, including novel words, in sentence structures.     

Testing the ability of species distribution models to infer variable importance

Models of species’ distributions and niches are frequently used to infer the importance of range- and niche-defining variables. However, the degree to which these models can reliably identify important variables and quantify their influence remains unknown. Here we use a series of simulations to explore how well models can 1) discriminate between variables with different influence and 2) calibrate the magnitude of influence relative to an ‘omniscient’ model. To quantify variable importance, we trained generalized additive models (GAMs), Maxent and boosted regression trees (BRTs) on simulated data and tested their sensitivity to permutations in each predictor. Importance was inferred by calculating the correlation between permuted and unpermuted predictions, and by comparing predictive accuracy of permuted and unpermuted predictions using AUC and the continuous Boyce index. In scenarios with one influential and one uninfluential variable, models failed to discriminate reliably between variables when training occurrences were < 8–64, prevalence was > 0.5, spatial extent was small, environmental data had coarse resolution and spatial autocorrelation was low, or when pairwise correlation between environmental variables was |r| > 0.7. When two variables influenced the distribution equally, importance was underestimated when species had narrow or intermediate niche breadth. Interactions between variables in how they shaped the niche did not affect inferences about their importance. When variables acted unequally, the effect of the stronger variable was overestimated. GAMs and Maxent discriminated between variables more reliably than BRTs, but no algorithm was consistently well-calibrated vis-à-vis the omniscient model. Algorithm-specific measures of importance like Maxent's change-in-gain metric were less robust than the permutation test. Overall, high predictive accuracy did not connote robust inferential capacity. As a result, requirements for reliably measuring variable importance are likely more stringent than for creating models with high predictive accuracy.     

A chromosome's protein is a dependent variable of the total metaphase protein

Chinese hamster metaphases were stained with dansyl chloride to determine protein. Using a microfluorometer, the fluorescence of chromosome number one and the total of the fluorescence of the metaphase were measured. The two measurements were found to be correlated with a coefficient of 0.640, suggesting that protein content in a single chromosome is a dependent variable of the protein content of the respective metaphase.     

Correlation between the numbers of two types of children when the family size distribution is zero-truncated negative binomial.

Let family size N be a random variable formed from two types of children B and C (B + C = N). The correlation coefficient between B and C has been recently studied by Rao et al. [1973]. We consider this problem when the frequency of fertile childless families (N = 0) is not available or cannot be estimated. Assuming the distribution of N is zero-truncated negative binomial, a formula for the correlation coefficient is derived. Applying this formula to the data of Reed and Reed [1965] gives a correlation coefficient of 0.27. The corresponding empirical coefficient for non-zero size families calculated directly from the data is 0.24.     

Multitrait Bayesian shrinkage and variable selection models with the BGLR-R package

The BGLR-R package implements various types of single-trait shrinkage/variable selection Bayesian regressions. The package was first released in 2014, since then it has become a software very often used in genomic studies. We recently develop functionality for multitrait models. The implementation allows users to include an arbitrary number of random-effects terms. For each set of predictors, users can choose diffuse, Gaussian, and Gaussian–spike–slab multivariate priors. Unlike other software packages for multitrait genomic regressions, BGLR offers many specifications for (co)variance parameters (unstructured, diagonal, factor analytic, and recursive). Samples from the posterior distribution of the models implemented in the multitrait function are generated using a Gibbs sampler, which is implemented by combining code written in the R and C programming languages. In this article, we provide an overview of the models and methods implemented BGLR’s multitrait function, present examples that illustrate the use of the package, and benchmark the performance of the software.     

Seasonality selects for more acutely virulent parasites when virulence is density dependent

Host condition is often likely to influence parasite virulence. Furthermore, condition may often be correlated with host density, and therefore, it is important to understand the role of density-dependent virulence (DDV). We examine the consequences of DDV to the evolution of parasites in both seasonal and non-seasonal environments. In particular, we consider seasonality in host birth rate that results in a fluctuating host density and therefore a variable virulence. We show that parasites are selected for lower exploitation, and therefore lower transmission and virulence as the strength of DDV increases without seasonality. This is an important insight from our models; DDV has the opposite effect on the evolution of parasites to that of higher baseline mortality. Our key result is that although seasonality does not affect the evolution of virulence in classical models, with DDV parasites in seasonal environments are predicted to evolve to be more acute. This suggests that in more seasonal environments wildlife disease is likely to be more rather than less virulent if DDV is widespread.     

DNA methyltransferases: mechanistic models derived from kinetic analysis

The sequence-specific transfer of methyl groups from donor S-adenosyl-L-methionine (AdoMet) to certain positions of DNA-adenine or -cytosine residues by DNA methyltransferases (MTases) is a major form of epigenetic modification. It is virtually ubiquitous, except for some notable exceptions. Site-specific methylation can be regarded as a means to increase DNA information capacity and is involved in a large spectrum of biological processes. The importance of these functions necessitates a deeper understanding of the enzymatic mechanism(s) of DNA methylation. DNA MTases fall into one of two general classes; viz. amino-MTases and [C5-cytosine]-MTases. Amino-MTases, common in prokaryotes and lower eukaryotes, catalyze methylation of the exocyclic amino group of adenine ([N6-adenine]-MTase) or cytosine ([N4-cytosine]-MTase). In contrast, [C5-cytosine]-MTases methylate the cyclic carbon-5 atom of cytosine. Characteristics of DNA MTases are highly variable, differing in their affinity to their substrates or reaction products, their kinetic parameters, or other characteristics (order of substrate binding, rate limiting step in the overall reaction). It is not possible to present a unifying account of the published kinetic analyses of DNA methylation because different authors have used different substrate DNAs and/or reaction conditions. Nevertheless, it would be useful to describe those kinetic data and the mechanistic models that have been derived from them. Thus, this review considers in turn studies carried out with the most consistently and extensively investigated [N6-adenine]-, [N4-cytosine]- and [C5-cytosine]-DNA MTases.     

Pregnenolone sulfate induces NMDA receptor dependent release of dopamine from synaptic terminals in the striatum

Neuromodulators that alter the balance between lower-frequency glutamate-mediated excitatory and higher-frequency GABA-mediated inhibitory synaptic transmission are likely to participate in core mechanisms for CNS function and may contribute to the pathophysiology of neurological disorders such as schizophrenia and Alzheimer's disease. Pregnenolone sulfate (PS) modulates both ionotropic glutamate and GABA(A) receptor mediated synaptic transmission. The enzymes necessary for PS synthesis and degradation are found in brain tissue of several species including human and rat, and up to 5 nM PS has been detected in extracts of postmortem human brain. Here, we ask whether PS could modulate transmitter release from nerve terminals located in the striatum. Superfusion of a preparation of striatal nerve terminals comprised of mixed synaptosomes and synaptoneurosomes with brief-duration (2 min) pulses of 25 nM PS demonstrates that PS increases the release of newly accumulated [3H]dopamine ([3H]DA), but not [14C]glutamate or [3H]GABA, whereas pregnenolone is without effect. PS does not affect dopamine transporter (DAT) mediated uptake of [3H]DA, demonstrating that it specifically affects the transmitter release mechanism. The PS-induced [3H]DA release occurs via an NMDA receptor (NMDAR) dependent mechanism as it is blocked by D-2-amino-5-phosphonovaleric acid. PS modulates DA release with very high potency, significantly increasing [3H]DA release at PS concentrations as low as 25 pM. This first report of a selective direct enhancement of synaptosomal dopamine release by PS at picomolar concentrations via an NMDAR dependent mechanism raises the possibility that dopaminergic axon terminals may be a site of action for this neurosteroid.     

Lysogenization by bacteriophage lambda: III. - Multiplicity dependent phenomena occuring upon infection by lambda

Lysogenization by bacteriophage λ involves at least two multiplicity dependent processes [2, 3]. For the purpose of comparison, other multiplicity dependent phenomena which occur upon infection by λ have been reviewed. These include the inhibition of host'syntheses as already described by others [9] and two phenomena which are shown to be multiplicity dependent, host killing by phage unable to replicate and inhibition of cell division. It is also demonstrated that, in at least two cases (lysogenization by phage able to replicate and killing by phage unable to replicate) the multiplicity dependent character disappears at slow cellular growth rates. The significance of these results is discussed with regard to three models which are susceptible to account for multiplicity dependent phenomena in general.     

Vitamin K dependent carboxylation: determination of the stereochemical course using 4-fluoroglutamyl-containing substrate

The stereochemical course of the vitamin K dependent carboxylation has been elucidated using a (4S)-4-fluoroglutamyl-containing pentapeptide as a substrate. The absolute configuration of the [13C]-4-carboxy-4-fluoroglutamate obtained when the carboxylation was carried out with 13C-labeled sodium bicarbonate, was determined after reduction of the [13C]-4-carboxy-4-fluoroglutamyl residue into 4-fluoro-5,5'-dihydroxyleucine, hydrolysis, lactonization, and peracetylation. The absolute configuration at C-4 was determined to be S by locating the 13C label in the lactone ring of the trans isomeric lactone and in the hydroxymethyl group of the cis isomer following HPLC separation of both isomers and analysis by GC/MS/MS techniques. It follows that the vitamin K dependent carboxylation occurs with inversion of configuration.     

Kinetic manifestations of allosteric interactions in models of regulatory enzymes with "indirect" co-operativity

The shape of the plots of initial reaction rate (ν) versus initial substrate concentration ([S]₀) and versus initial concentration of allosteric effector ([F]₀) for the model of allosteric enzyme of Monod, Wyman &; Changeux (1965) and for the model of dissociating regulatory enzyme has been analysed by means of the inconstant exponent (q) for substrate or effector concentration, respectively. It has been shown that allosteric interactions in above-mentioned models with “indirect” co-operativity may be manifested not only by the sigmoidal shape of the plot of ν versus [S]₀ or ν versus [F]₀ (with one point of inflexion) but also by the increase in the magnitude of exponent q in progress of saturation process of the enzyme by the substrate or by the effector in the absence of the sigmoidal shape of these plots. It has been shown also that the plot of ν versus [S]₀ has two inflexion points when the parameters have certain definite values. One of these inflexion points (or even both at definite values of the parameters) is hardly discernible. At certain definite values of the parameters two inflexion points may be kinetically manifested by such phenomenon as “negative” co-operativity (q < 1). This is possible if one of the interconvertable enzyme forms exceeds another not only in the affinity to the substrate but also in the value of the rate constant for catalytic breakdown of the enzyme-substrate complex.     

The ideal free distribution when the resource is variable

On the basis of the ideal free distribution (IFD) model, twostochastic models that incorporate the uncertainty of the informationused for decision making were considered to investigate theeffects of the variability in the resource supply rate on theIFD under continuous input conditions. In the uncertain-informationmodel, competitors cannot trace the variation of the supplyrate and use the expectation of the supply rate or previouspayoffs for decision making. Both submodels predict matchingof means, in which the average number of competitors for eachpatch is proportional to the average supply rate in the patch.In the perfect-information model, competitors continuously knowand trace the environment conditions. Numerical predictionsdepend on the relative size of the resource variance betweenpatches. When the resource variance in the good patch is sufficientlylarger than that in the poor patch, it predicts undermatchingof means; when the variance of the supply rate for each patchis small and proportional to the average of the supply ratein the patch, it predicts matching of means; and when the resourcevariance in the poor patch is larger than (or equal to) thatin the good patch, it predicts overmatching of means. Theseresults indicate the importance of clarifying the assumptionon the uncertainty in information for decision making and thetype of the resource variance for the test of the IFD underconditions where the resource supply rate is stochastic.     

Effects of cocaine on sodium dependent dopamine uptake in rat striatal synaptosomes

Initial velocity of uptake of dopamine (DA) has been measured in the presence of 1M cocaine as a function of both [DA] and [Na]. Although DA uptake is overwhelmingly dependent on sodium, it appears that a small amount of DA uptake takes place in the absence of sodium. This contrasts with a previous study of the sodium dependence of uptake without cocaine (referred to below as control), in which uptake was found to be 100% sodium dependent. The data were fitted to several rapid equilibrium models and the minimal best fit model identified. The interaction of transporter (C), DA (S), and Na⁺ (Na) in this present model is identical to the reaction scheme found previously to fit control data (no cocaine). Whereas the control model required translocation only as CNa₂S, in the presence of cocaine (I), two additional translocated species are required to fit the data (CS and CNaS). Another previous study of the interaction of carrier and cocaine at a constant [Na]₀ predicted that cocaine interacts with a transporter site other than the DA binding site and that uptake takes place as CS and CSI. The present results are consistent with the assumption that the CS and CNaS forms of the present model are actually CSI and CNaSI, since they are required to fit a model of the sodium dependence in the presence of cocaine, but are not required in the absence of cocaine.     

Logit effect values and efficiency of significance tests when a noncollapsible variable is omitted

For linear models, randomization of the assignment to the levels of a variable A is a sufficient condition to obtain correct estimates of the effect of A on another variable B. For logit models, the omission of a relevant variable Z often results in invalid estimates of the effect of A and in a loss of power of the significance test. The values of this underestimation and of the loss of power are evaluated here asymptotically and from simulation studies. Consequences for practical use of logit models in randomized settings are underlined.     

Multiple imputation in the presence of an incomplete binary variable created from an underlying continuous variable

Multiple imputation (MI) is used to handle missing at random (MAR) data. Despite warnings from statisticians, continuous variables are often recoded into binary variables. With MI it is important that the imputation and analysis models are compatible; variables should be imputed in the same form they appear in the analysis model. With an encoded binary variable more accurate imputations may be obtained by imputing the underlying continuous variable. We conducted a simulation study to explore how best to impute a binary variable that was created from an underlying continuous variable. We generated a completely observed continuous outcome associated with an incomplete binary covariate that is a categorized version of an underlying continuous covariate, and an auxiliary variable associated with the underlying continuous covariate. We simulated data with several sample sizes, and set 25% and 50% of data in the covariate to MAR dependent on the outcome and the auxiliary variable. We compared the performance of five different imputation methods: (a) Imputation of the binary variable using logistic regression; (b) imputation of the continuous variable using linear regression, then categorizing into the binary variable; (c, d) imputation of both the continuous and binary variables using fully conditional specification (FCS) and multivariate normal imputation; (e) substantive-model compatible (SMC) FCS. Bias and standard errors were large when the continuous variable only was imputed. The other methods performed adequately. Imputation of both the binary and continuous variables using FCS often encountered mathematical difficulties. We recommend the SMC-FCS method as it performed best in our simulation studies.     

Can optimality models and an ‘optimality research program’ help us understand some plant–fungal relationships?

In this paper we suggest that the field of fungal ecology may benefit from the use of optimality models in the context of an ‘optimality research program’ (ORP). An ORP is a research program in the sense of modern philosopher of science Lakatos' [1978. The Methodology of Scientific Research Programmes: philosophical papers, vol. 1. Cambridge University Press, Cambridge] seminal work. An optimality research program has a lengthy history and record of success in the field of behavioural ecology, but has been seldom employed in fungal ecology. We discuss the ORP and provide some examples of how optimality models may be useful in fungal ecology. We suggest that such an approach may benefit experimental fungal ecologists by: providing a framework for organizing knowledge; generating hypotheses; helping in the planning of experiments; aiding in the interpretation of results; and directing the next steps of an experimental research program. We illustrate these benefits by sketching out how an ORP might be used to answer some fundamental questions about the interactions between host plants and arbuscular mycorrhizal fungi.     

Imputation and variable selection in linear regression models with missing covariates

Across multiply imputed data sets, variable selection methods such as stepwise regression and other criterion-based strategies that include or exclude particular variables typically result in models with different selected predictors, thus presenting a problem for combining the results from separate complete-data analyses. Here, drawing on a Bayesian framework, we propose two alternative strategies to address the problem of choosing among linear regression models when there are missing covariates. One approach, which we call "impute, then select" (ITS) involves initially performing multiple imputation and then applying Bayesian variable selection to the multiply imputed data sets. A second strategy is to conduct Bayesian variable selection and missing data imputation simultaneously within one Gibbs sampling process, which we call "simultaneously impute and select" (SIAS). The methods are implemented and evaluated using the Bayesian procedure known as stochastic search variable selection for multivariate normal data sets, but both strategies offer general frameworks within which different Bayesian variable selection algorithms could be used for other types of data sets. A study of mental health services utilization among children in foster care programs is used to illustrate the techniques. Simulation studies show that both ITS and SIAS outperform complete-case analysis with stepwise variable selection and that SIAS slightly outperforms ITS.     

cAMP dependent chloride conductance is not different in cystic fibrosis fibroblasts

A chloride conductive pathway has been demonstrated in human skin fibroblasts and a defective cAMP dependent activation of this conductance in Cystic Fibrosis (CF) fibroblasts has been also reported. Chloride transport by the same reported method was studied in normal and CF skin fibroblasts. The stimulation of this pathway was not obtained consistently by the addition of dibutyryl cAMP. The addition of prostaglandin E1 (PGE1) increased the intracellular [cAMP] but did not increase the conductivity of the pathway consistently. Neither the basal nor the dibutyryl cAMP or the PGE1 stimulated chloride conductance differed significantly in CF fibroblasts.     

Parameterising variable assimilation efficiency in predator–prey models

Our understanding of the dynamics of predator–prey systems has relied heavily on the use of models based on the standard Lotka–Volterra (LV) framework, dating back over 80 years. Although these models have been repeatedly analysed and refined since their initial inception, the way they describe the predator's growth rate has received surprisingly little attention; typically it is simply assumed that the predator's growth rate is linearly related to its ingestion rate according to a constant assimilation efficiency, e. However, for many consumers e is known to decrease at high prey densities. Models that ignore variable assimilation efficiencies overlook potentially important non‐linearities, affecting the validity of predictions relating to conservation, invasion biology and pest control. Directly quantifying the relationship between e and prey abundance is, however, difficult. An alternative approach (the independent‐response, IR, approach) is to not assume any direct link between the predator's functional response (the relationship between ingestion rate and prey abundance) and its growth response. This flexibility is invaluable when parameterising models from data; providing the model‐fitting process is constrained to ensure that e never exceeds 1, this approach allows considerable insight into whether, and how, e varies with prey density. Here we examine the synergistic value of combining the IR and LV approaches. We illustrate these concepts through analysis of published functional and growth response data and show that, in many cases, e does vary with prey abundance. This paper is the first recognition that these two complementary approaches can be combined into a single framework that allows the relationship between a predator's functional and growth responses to emerge during the parameterisation process, thereby acting as a compromise between restrictive models that require this relationship to be defined a priori, and completely unrestrained models that allow assimilation efficiencies to exceed 1.