Calibration of a complex hydro-ecological model through Approximate Bayesian Computation and Random Forest combined with sensitivity analysis

An automated calibration method is proposed and applied to the complex hydro-ecological model Delft3D-BLOOM which is calibrated from monitoring data of the lake Champs-sur-Marne, a small shallow urban lake in the Paris region (France). This method (ABC-RF-SA) combines Approximate Bayesian Computation (ABC) with the machine learning algorithm Random Forest (RF) and a Sensitivity Analysis (SA) of the model parameters. Three target variables are used (total chlorophyll, cyanobacteria and dissolved oxygen concentration) to calibrate 133 parameters. ABC-RF-SA is first applied on a set of simulated observations to validate the methodology. It is then applied on a real set of high-frequency observations recorded during about two weeks on the lake Champs-sur-Marne. The methodology is also compared to standard ABC and ABC-RF formulations. Only ABC-RF-SA allowed the model to reproduce the observed biogeochemical dynamics. The coupling of ABC with RF and SA thus appears crucial for its application to complex hydro-ecological models.     

Competition patterns among phytoplankton functional groups: How useful are the complex mathematical models?

Simple models have significant contribution to the development of ecological theory. However, these minimalistic modeling approaches usually focus on a small subset of the causes of a phenomenon and neglect important aspects of system dynamics. In this study, we use a complex aquatic biogeochemical model to examine competition patterns and structural shifts in the phytoplankton community under nutrient enrichment conditions. Our model simulates multiple elemental cycles (org. C, N, P, Si, O), multiple functional phytoplankton (diatoms, green algae and cyanobacteria) and zooplankton (copepods and cladocerans) groups. It also takes into account recent advances in stoichiometric nutrient recycling theory, and the zooplankton grazing term is reformulated to include algal food quality effects on zooplankton assimilation efficiency. The model provided a realistic platform to examine the functional properties (e.g., kinetics, growth strategies, intracellular storage capacity) and the abiotic conditions (temperature, nutrient loading) under which the different phytoplankton groups can dominate or can be competitively excluded in oligo, meso and eutrophic environments. Based on the results of our analysis, the intergroup variability in the minimum cell quota and maximum transport rate at the cell surface for phosphorus along with the group-specific metabolic losses can shape the structure of plankton communities. We also use classification tree analysis to elucidate aspects (e.g., relative differences in the functional group properties, critical values of the abiotic conditions, levels of the other plankton community residents) of the complex interplay among physical, chemical and biological factors that drive epilimnetic plankton dynamics. Finally, our study highlights the importance of improving the mathematical representation of phytoplankton adaptive strategies for resources procurement (e.g., regulation of transport kinetics, effects of transport kinetics on the kinetics of assimilation, relationship between assimilation and growth) to effectively link variability at the organismal level with ecosystem-scale patterns.     

Intraspecific competition in models for vegetation patterns: Decrease in resilience to aridity and facilitation of species coexistence

Patterned vegetation is a characteristic feature of many dryland ecosystems. While plant densities on the ecosystem-wide scale are typically low, a spatial self-organisation principle leads to the occurrence of alternating patches of high biomass and patches of bare soil. Nevertheless, intraspecific competition dynamics other than competition for water over long spatial scales are commonly ignored in mathematical models for vegetation patterns. In this paper, I address the impact of local intraspecific competition on a modelling framework for banded vegetation patterns. Firstly, I show that in the context of a single-species model, neglecting local intraspecific competition leads to an overestimation of a patterned ecosystem’s resilience to increases in aridity. Secondly, in the context of a multispecies model, I argue that local intraspecific competition is a key element in the successful capture of species coexistence in model solutions representing a vegetation pattern. For both models, a detailed bifurcation analysis is presented to analyse the onset, existence and stability of patterns. Besides the strengths of local intraspecific competition, also the difference between two species has a significant impact on the bifurcation structure, providing crucial insights into the complex ecosystem dynamics. Predictions on future ecosystem dynamics presented in this paper, especially on pattern onset and pattern stability, can aid the development of conservation programs.     

Global patterns and substrate‐based mechanisms of the terrestrial nitrogen cycle

Nitrogen (N) deposition is impacting the services that ecosystems provide to humanity. However, the mechanisms determining impacts on the N cycle are not fully understood. To explore the mechanistic underpinnings of N impacts on N cycle processes, we reviewed and synthesised recent progress in ecosystem N research through empirical studies, conceptual analysis and model simulations. Experimental and observational studies have revealed that the stimulation of plant N uptake and soil retention generally diminishes as N loading increases, while dissolved and gaseous losses of N occur at low N availability but increase exponentially and become the dominant fate of N at high loading rates. The original N saturation hypothesis emphasises sequential N saturation from plant uptake to soil retention before N losses occur. However, biogeochemical models that simulate simultaneous competition for soil N substrates by multiple processes match the observed patterns of N losses better than models based on sequential competition. To enable better prediction of terrestrial N cycle responses to N loading, we recommend that future research identifies the response functions of different N processes to substrate availability using manipulative experiments, and incorporates the measured N saturation response functions into conceptual, theoretical and quantitative analyses.     

Efficiencies of maximum likelihood methods of phylogenetic inferences when different substitution models are used

Choice of a substitution model is a crucial step in the maximum likelihood (ML) method of phylogenetic inference, and investigators tend to prefer complex mathematical models to simple ones. However, when complex models with many parameters are used, the extent of noise in statistical inferences increases, and thus complex models may not produce the true topology with a higher probability than simple ones. This problem was studied using computer simulation. When the number of nucleotides used was relatively large (1000 bp), the HKY+Gamma model showed smaller d(T) topological distance between the inferred and the true trees) than the JC and Kimura models. In the cases of shorter sequences (300 bp) simpler model and search algorithm such as JC model and SA+NNI search were found to be as efficient as more complicated searches and models in terms of topological distances, although the topologies obtained under HKY+Gamma model had the highest likelihood values. The performance of relatively simple search algorithm SA+NNI was found to be essentially the same as that of more extensive SA+TBR search under all models studied. Similarly to the conclusions reached by Takahashi and Nei [Mol. Biol. Evol. 17 (2000) 1251], our results indicate that simple models can be as efficient as complex models, and that use of complex models does not necessarily give more reliable trees compared with simple models.     

Nitrogen limitation of decomposition and decay: How can it occur?

The availability of nitrogen (N) is a critical control on the cycling and storage of soil carbon (C). Yet, there are conflicting conceptual models to explain how N availability influences the decomposition of organic matter by soil microbial communities. Several lines of evidence suggest that N availability limits decomposition; the earliest stages of leaf litter decay are associated with a net import of N from the soil environment, and both observations and models show that high N organic matter decomposes more rapidly. In direct contrast to these findings, experimental additions of inorganic N to soils broadly show a suppression of microbial activity, which is inconsistent with N limitation of decomposition. Resolving this apparent contradiction is critical to representing nutrient dynamics in predictive ecosystem models under a multitude of global change factors that alter soil N availability. Here, we propose a new conceptual framework, the Carbon, Acidity, and Mineral Protection hypothesis, to understand the effects of N availability on soil C cycling and storage and explore the predictions of this framework with a mathematical model. Our model simulations demonstrate that N addition can have opposing effects on separate soil C pools (particulate and mineral‐protected carbon) because they are differentially affected by microbial biomass growth. Moreover, changes in N availability are frequently linked to shifts in soil pH or osmotic stress, which can independently affect microbial biomass dynamics and mask N stimulation of microbial activity. Thus, the net effect of N addition on soil C is dependent upon interactions among microbial physiology, soil mineralogy, and soil acidity. We believe that our synthesis provides a broadly applicable conceptual framework to understand and predict the effect of changes in soil N availability on ecosystem C cycling under global change.     

Ensemble modeling for analysis of cell signaling dynamics

Systems biology iteratively combines experimentation with mathematical modeling. However, limited mechanistic knowledge, conflicting hypotheses and scarce experimental data severely hamper the development of predictive mechanistic models in many areas of biology. Even under such high uncertainty, we show here that ensemble modeling, when combined with targeted experimental analysis, can unravel key operating principles in complex cellular pathways. For proof of concept, we develop a library of mechanistically alternative dynamic models for the highly conserved target-of-rapamycin (TOR) pathway of Saccharomyces cerevisiae. In contrast to the prevailing view of a de novo assembly of type 2A phosphatases (PP2As), our integrated computational and experimental analysis proposes a specificity factor, based on Tap42p-Tip41p, for PP2As as the key signaling mechanism that is quantitatively consistent with all available experimental data. Beyond revising our picture of TOR signaling, we expect ensemble modeling to help elucidate other insufficiently characterized cellular circuits.     

Seasonal shifts in community composition and proteome expression in a sulphur-cycling cyanobacterial mat

Seasonal changes in light and physicochemical conditions have strong impacts on cyanobacteria, but how they affect community structure, metabolism, and biogeochemistry of cyanobacterial mats remains unclear. Light may be particularly influential for cyanobacterial mats exposed to sulphide by altering the balance of oxygenic photosynthesis and sulphide-driven anoxygenic photosynthesis. We studied temporal shifts in irradiance, water chemistry, and community structure and function of microbial mats in the Middle Island Sinkhole (MIS), where anoxic and sulphate-rich groundwater provides habitat for cyanobacteria that conduct both oxygenic and anoxygenic photosynthesis. Seasonal changes in light and groundwater chemistry were accompanied by shifts in bacterial community composition, with a succession of dominant cyanobacteria from Phormidium to Planktothrix, and an increase in diatoms, sulphur-oxidizing bacteria, and sulphate-reducing bacteria from summer to autumn. Differential abundance of cyanobacterial light-harvesting proteins likely reflects a physiological response of cyanobacteria to light level. Beggiatoa sulphur oxidation proteins were more abundant in autumn. Correlated abundances of taxa through time suggest interactions between sulphur oxidizers and sulphate reducers, sulphate reducers and heterotrophs, and cyanobacteria and heterotrophs. These results support the conclusion that seasonal change, including light availability, has a strong influence on community composition and biogeochemical cycling of sulphur and O₂ in cyanobacterial mats.     

Liana optical traits increase tropical forest albedo and reduce ecosystem productivity

Lianas are a key growth form in tropical forests. Their lack of self-supporting tissues and their vertical position on top of the canopy make them strong competitors of resources. A few pioneer studies have shown that liana optical traits differ on average from those of colocated trees. Those trait discrepancies were hypothesized to be responsible for the competitive advantage of lianas over trees. Yet, in the absence of reliable modelling tools, it is impossible to unravel their impact on the forest energy balance, light competition, and on the liana success in Neotropical forests. To bridge this gap, we performed a meta-analysis of the literature to gather all published liana leaf optical spectra, as well as all canopy spectra measured over different levels of liana infestation. We then used a Bayesian data assimilation framework applied to two radiative transfer models (RTMs) covering the leaf and canopy scales to derive tropical tree and liana trait distributions, which finally informed a full dynamic vegetation model. According to the RTMs inversion, lianas grew thinner, more horizontal leaves with lower pigment concentrations. Those traits made the lianas very efficient at light interception and significantly modified the forest energy balance and its carbon cycle. While forest albedo increased by 14% in the shortwave, light availability was reduced in the understorey (?30% of the PAR radiation) and soil temperature decreased by 0.5°C. Those liana-specific traits were also responsible for a significant reduction of tree (?19%) and ecosystem (?7%) gross primary productivity (GPP) while lianas benefited from them (their GPP increased by +27%). This study provides a novel mechanistic explanation to the increase in liana abundance, new evidence of the impact of lianas on forest functioning, and paves the way for the evaluation of the large-scale impacts of lianas on forest biogeochemical cycles.     

Segregation analysis of a complex quantitative trait: approaches for identifying influential data points

BACKGROUND/AIMS: Complex traits pose a particular challenge to standard methods for segregation analysis (SA), and for such traits it is difficult to assess the ability of complex SA (CSA) to approximate the true mode of inheritance. Here we use an oligogenic Bayesian Markov chain Monte Carlo method for SA (OSA) to verify results from a single-locus likelihood-based CSA for data on a quantitative measure of reading ability. METHODS: We compared the profile likelihood from CSA, maximized over the trait allele frequency, to the posterior distribution of genotype effects from OSA to explore differences in the overall parameter estimates from SA on the original phenotype data and the same data Winsorized to reduce the potential influence of three outlying data points. RESULTS: Bayesian OSA revealed two modes of inheritance, one of which coincided with the QTL model from CSA. Winsorizing abolished the model originally estimated by CSA; both CSA and OSA identified only the second OSA model. CONCLUSION: Differences between the results from the two methods alerted us to the presence of influential data points, and identified the QTL model best supported by the data. Thus, the Bayesian OSA proved a valuable tool for assessing and verifying inheritance models from CSA.     

The effect of recycling on plant competitive hierarchies

Evidence from field studies suggests that some plant species enhance their persistence by reinforcing patterns of N availability through differences in litter quality. Using mathematical models of nutrient flow, we explore whether and how recycling affects plant growth, competition, and coexistence and whether it leads to positive feedbacks. Two mechanisms are considered: the ability of plants to access two forms of soil N, complex (e.g., organic) and simple (e.g., nitrate), and the effect of density-dependent limitation of growth. Except in the trivial case of limitation by N in one form without density dependence, differences in litter quality can prevent the establishment of competitors. Feedback can, conversely, facilitate the invasion of competitors. At equilibrium, the rate of decomposition does not affect the outcome of competition. Species affect their long-term persistence if they alter the fraction of nitrogen that is returned to the soil and becomes available for plant uptake. Increasing the fraction of N that is recycled favors specialists in complex nitrogen and species that suppress the growth of others at high nitrogen availability. Increasing the rate of microbial decomposition of complex nitrogen favors specialists in simple nitrogen.     

A Predictive Framework for Integrating Disparate Genomic Data Types Using Sample-Specific Gene Set Enrichment Analysis and Multi-Task Learning

Understanding the root molecular and genetic causes driving complex traits is a fundamental challenge in genomics and genetics. Numerous studies have used variation in gene expression to understand complex traits, but the underlying genomic variation that contributes to these expression changes is not well understood. In this study, we developed a framework to integrate gene expression and genotype data to identify biological differences between samples from opposing complex trait classes that are driven by expression changes and genotypic variation. This framework utilizes pathway analysis and multi-task learning to build a predictive model and discover pathways relevant to the complex trait of interest. We simulated expression and genotype data to test the predictive ability of our framework and to measure how well it uncovered pathways with genes both differentially expressed and genetically associated with a complex trait. We found that the predictive performance of the multi-task model was comparable to other similar methods. Also, methods like multi-task learning that considered enrichment analysis scores from both data sets found pathways with both genetic and expression differences related to the phenotype. We used our framework to analyze differences between estrogen receptor (ER) positive and negative breast cancer samples. An analysis of the top 15 gene sets from the multi-task model showed they were all related to estrogen, steroids, cell signaling, or the cell cycle. Although our study suggests that multi-task learning does not enhance predictive accuracy, the models generated by our framework do provide valuable biological pathway knowledge for complex traits.     

Disturbance Decouples Biogeochemical Cycles Across Forests of the Southeastern US

Biogeochemical cycles are inherently linked through the stoichiometric demands of the organisms that cycle the elements. Landscape disturbance can alter element availability and thus the rates of biogeochemical cycling. Nitrification is a fundamental biogeochemical process positively related to plant productivity and nitrogen loss from soils to aquatic systems, and the rate of nitrification is sensitive to both carbon and nitrogen availability. Yet how these controls influence nitrification rates at the landscape scale is not fully elucidated. We, therefore, sampled ten watersheds with different disturbance histories in the southern Appalachian Mountains to examine effects on potential net nitrification rates. Using linear mixed model selection (AIC), we narrowed a broad suite of putative explanatory variables into a set of models that best explained landscape patterns in potential net nitrification. Forest disturbance history determined whether nitrification and nitrogen mineralization were correlated, with the effect apparently mediated by microbially available carbon. Undisturbed forests had higher available carbon, which uncoupled potential net nitrification from potential net nitrogen mineralization. In contrast, disturbed watersheds had lower available carbon, and nitrification rates were strongly correlated to those of nitrogen mineralization. These data suggest that a history of disturbance at the landscape scale reduces soil carbon availability, which increases ammonium availability to nitrifiers at the micro-scale. Landscape-level soil carbon availability then appears to determine the coupling of autotrophic (nitrification) and heterotrophic (nitrogen mineralization) biogeochemical processes, and hence the relationship between carbon and nitrogen cycling in soils.     

Controlling the time evolution of mAb N‐linked glycosylation ‐ Part II: Model‐based predictions

N‐linked glycosylation is known to be a crucial factor for the therapeutic efficacy and safety of monoclonal antibodies (mAbs) and many other glycoproteins. The nontemplate process of glycosylation is influenced by external factors which have to be tightly controlled during the manufacturing process. In order to describe and predict mAb N‐linked glycosylation patterns in a CHO‐S cell fed‐batch process, an existing dynamic mathematical model has been refined and coupled to an unstructured metabolic model. High‐throughput cell culture experiments carried out in miniaturized bioreactors in combination with intracellular measurements of nucleotide sugars were used to tune the parameter configuration of the coupled models as a function of extracellular pH, manganese and galactose addition. The proposed modeling framework is able to predict the time evolution of N‐linked glycosylation patterns during a fed‐batch process as a function of time as well as the manipulated variables. A constant and varying mAb N‐linked glycosylation pattern throughout the culture were chosen to demonstrate the predictive capability of the modeling framework, which is able to quantify the interconnected influence of media components and cell culture conditions. Such a model‐based evaluation of feeding regimes using high‐throughput tools and mathematical models gives rise to a more rational way to control and design cell culture processes with defined glycosylation patterns. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 32:1135–1148, 2016     

Continuum model of fibroblast-driven wound contraction: inflammation-mediation.

We propose a mathematical model to aid the understanding of how events in wound healing are orchestrated to result in wound contraction. Ultimately, a validated model could provide a predictive means for enhancing or mitigating contraction as is appropriate for managing a particular wound. The complex nature of wound healing and the lack of a modeling framework which can account for both the relevant cell biology and biomechanics are major reasons for the absence of models to date. Here we adapt a model originally proposed by Murray and co-workers to show how cell traction forces can result in spatial patterns of cell aggregates since it offers a framework for understanding how traction exerted by wound fibroblasts drives wound contraction. Since it is a continuum model based on conservation laws which reflect assumed cell and tissue properties, it is readily extended to account for emerging understanding of the cell biology of wound healing and its relationship to inflammation. We consider various sets of assumed properties, based on current knowledge, within a base model of dermal wound healing and compare predictions of the rate and extent of wound contraction to published experimental results.     

Using a new framework of two-phase generalized additive models to incorporate prey abundance in spatial distribution models of juvenile slender lizardfish in Haizhou Bay,China

The predictive skill of species distribution models depends on the quality and quantity of input information. In addition to the physical environmental variables, prey availability is also one of the main drivers regulating spatial distribution of marine species. However, prey distribution data have rarely been considered in habitat models due to the lack of information on non-commercial prey species. This may lead to an incomplete view of species distributions and biased model predictions. In this study, we developed a new framework of two-phase generalized additive models (GAMs) based on the Tweedie distribution to incorporate the predicted prey abundance as covariates in habitat models, and applied this framework to juvenile slender lizardfish Saurida elongata in Haizhou Bay, China. This study demonstrated that the predictive skill of habitat models could be greatly improved through incorporating prey abundance as explanatory variables. The importance of prey distribution data in the habitat model confirms the essentiality of including prey data while modelling species distribution. Spatial overlap and GAM analysis demonstrated that not all dominant prey can be selected as potential explanatory variables and only those prey species showing high spatiotemporal occurrences with predators should be incorporated. The framework derived in this study could be extended to other marine organisms to improve the predictive skill of habitat models and enhance our understanding of the ecological mechanisms underlying the distribution of marine species.     

陆地生物地球化学模型的应用和发展

以TEM和DNDC模型为例,在分析国外生物地球化学模型发展基础上,按照模拟方法,应用目的,元素类型,生态系统类型和空间尺度等对现有的生物地球化学模型进行了分类,对生物地球化学模型的基本框架（植物,大气和土壤3个组分及植物－大气,植物－土壤和土壤－大气界面等3个界面）,以及内部基本过程（物理的,化学的和生物的过程）进行了总结分析,对目前生物地球化学模型建立中的几个问题（如跨尺度问题,地理信息系统（GIS）和遥感技术结合,考虑人类活动的影响和生物地球化学模型的比较研究）的研究动态进行了评价。