Generation of an Industry-specific Physico-chemical Allocation Matrix. Application in the Dairy Industry and Implications for Systems Analysis (9 pp)

Background, Aims and Scope Allocation is required when quantifying environmental impacts of individual products from multi-product manufacturing plants. The International Organization for Standardization (ISO) recommends in ISO 14041 that allocation should reflect underlying physical relationships between inputs and outputs, or in the absence of such knowledge, allocation should reflect other relationships (e.g. economic value). Economic allocation is generally recommended if process specific information on the manufacturing process is lacking. In this paper, a physico-chemical allocation matrix, based on industry-specific data from the dairy industry, is developed and discussed as an alternative allocation method. Methods Operational data from 17 dairy manufacturing plants was used to develop an industry specific physico-chemical allocation matrix. Through an extensive process of substraction/substitution, it is possible to determine average resource use (e.g. electricity, thermal energy, water, etc) and wastewater emissions for individual dairy products within multi-product manufacturing plants. The average operational data for individual products were normalised to maintain industry confidentiality and then used as an industry specific allocation matrix. The quantity of raw milk required per product is based on the milk solids basis to account for dairy by-products that would otherwise be neglected. Results and Discussion Applying fixed type allocation methods (e.g. economic) for all input and outputs based on the quantity of product introduces order of magnitude sized deviations from physico-chemical allocation in some cases. The error associated with the quality of the whole of factory plant data or truncation error associated with setting system boundaries is insignificant in comparison. The profound effects of the results on systems analysis are discussed. The results raise concerns about using economic allocation as a default when allocating intra-industry sectoral flows (i.e. mass and process energy) in the absence of detailed technical information. It is recommended that economic allocation is better suited as a default for reflecting inter-industry sectoral flows. Conclusion The study highlights the importance of accurate causal allocation procedures that reflect industry-specific production methods. Generation of industry-specific allocation matrices is possible through a process of substitution/subtraction and optimisation. Allocation using such matrices overcomes the inherit bias of mass, process energy or price allocations for a multi-product manufacturing plant and gives a more realistic indication of resource use or emissions per product. The approach appears to be advantageous for resource use or emissions allocation if data is only available on a whole of factory basis for several plants with a similar level of technology. Recommendation and Perspective The industry specific allocation matrix approach will assist with allocation in multi-product LCAs where the level of technology in an industry is similar. The matrix will also benefit dairy manufacturing companies and help them more accurately allocate resources and impacts (i.e. costs) to different products within the one plant. It is recommended that similar physico-chemical allocation matrices be developed for other industry sectors with a view of ultimately coupling them with input-output analysis.     

Allocation of energy use in petroleum refineries to petroleum products

Aim, Scope, and Background  Studies to evaluate the energy and emission impacts of vehicle/fuel systems have to address allocation of the energy use and emissions associated with petroleum refineries to various petroleum products because refineries produce multiple products. The allocation is needed in evaluating energy and emission effects of individual transportation fuels. Allocation methods used so far for petroleum-based fuels (e.g., gasoline, diesel, and liquefied petroleum gas [LPG]) are based primarily on mass, energy content, or market value shares of individual fuels from a given refinery. The aggregate approach at the refinery level is unable to account for the energy use and emission differences associated with producing individual fuels at the next sub-level: individual refining processes within a refinery. The approach ignores the fact that different refinery products go through different processes within a refinery. Allocation at the subprocess level (i.e., the refining process level) instead of at the aggregate process level (i.e., the refinery level) is advocated by the International Standard Organization. In this study, we seek a means of allocating total refinery energy use among various refinery products at the level of individual refinery processes. Main Features  We present a petroleum refinery-process-based approach to allocating energy use in a petroleum refinery to petroleum refinery products according to mass, energy content, and market value share of final and intermediate petroleum products as they flow through refining processes within a refinery. The approach is based on energy and mass balance among refining processes within a petroleum refinery. By using published energy and mass balance data for a simplified U.S. refinery, we developed a methodology and used it to allocate total energy use within a refinery to various petroleum products. The approach accounts for energy use during individual refining processes by tracking product stream mass and energy use within a refinery. The energy use associated with an individual refining process is then distributed to product streams by using the mass, energy content, or market value share of each product stream as the weighting factors. Results  The results from this study reveal that product-specific energy use based on the refinery process-level allocation differs considerably from that based on the refinery-level allocation. We calculated well-to-pump total energy use and greenhouse gas (GHG) emissions for gasoline, diesel, LPG, and naphtha with the refinery process-based allocation approach. For gasoline, the efficiency estimated from the refinery-level allocation underestimates gasoline energy use, relative to the process-level based gasoline efficiency. For diesel fuel, the well-to-pump energy use for the process-level allocations with the mass- and energy-content-based weighting factors is smaller than that predicted with the refinery-level allocations. However, the process-level allocation with the market-value-based weighting factors has results very close to those obtained by using the refinery-level allocations. For LPG, the refinery-level allocation significantly overestimates LPG energy use. For naphtha, the refinery-level allocation overestimates naphtha energy use. The GHG emission patterns for each of the fuels are similar to those of energy use. Conclusions  We presented a refining-process-level-based method that can be used to allocate energy use of individual refining processes to refinery products. The process-level-based method captures process-dependent characteristics of fuel production within a petroleum refinery. The method starts with the mass and energy flow chart of a refinery, tracks energy use by individual refining processes, and distributes energy use of a given refining process to products from the process. In allocating energy use to refinery products, the allocation method could rely on product mass, product energy contents, or product market values as weighting factors. While the mass- and energy-content-based allocation methods provide an engineering perspective of energy allocation within a refinery, the market-value-based allocation method provides an economic perspective. The results from this study show that energy allocations at the aggregate refinery level and at the refining process level could make a difference in evaluating the energy use and emissions associated with individual petroleum products. Furthermore, for the refining-process-level allocation method, use of mass — energy content- or market value share-based weighting factors could lead to different results for diesel fuels, LPG, and naphtha. We suggest that, when possible, energy use allocations should be made at the lowest subprocess level — a confirmation of the recommendation by the International Standard Organization for life cycle analyses. Outlook  The allocation of energy use in petroleum refineries at the refining process level in this study follows the recommendation of ISO 14041 that allocations should be accomplished at the subprocess level when possible. We developed a method in this study that can be readily adapted for refineries in which process-level energy and mass balance data are available. The process-level allocation helps reveal some additional energy and emission burdens associated with certain refinery products that are otherwise overlooked with the refinery-level allocation. When possible, process-level allocation should be used in life-cycle analyses.     

A coupling procedure for the discrimination of mixed data.

A procedure is described for coupling different discriminators to a new (common) decision rule using the corresponding allocation vectors only. This enables one to cope jointly with data of different structure and/or scales of measurement but without strong restrictions on the number of (especially categorical) features. The method is combined with a consequent cross-validation process securing the results reached. Examples from medical diagnostics demonstrate the usefulness of the proposed procedure, especially in comparison with the known linear discriminant analysis as judged from the error rates obtained.     

Age correction in dementia--matching to a healthy brain

In recent research, many univariate and multivariate approaches have been proposed to improve automatic classification of various dementia syndromes using imaging data. Some of these methods do not provide the possibility to integrate possible confounding variables like age into the statistical evaluation. A similar problem sometimes exists in clinical studies, as it is not always possible to match different clinical groups to each other in all confounding variables, like for example, early-onset (age<65 years) and late-onset (age≥65) patients with Alzheimer's disease (AD). Here, we propose a simple method to control for possible effects of confounding variables such as age prior to statistical evaluation of magnetic resonance imaging (MRI) data using support vector machine classification (SVM) or voxel-based morphometry (VBM). We compare SVM results for the classification of 80 AD patients and 79 healthy control subjects based on MRI data with and without prior age correction. Additionally, we compare VBM results for the comparison of three different groups of AD patients differing in age with the same group of control subjects obtained without including age as covariate, with age as covariate or with prior age correction using the proposed method. SVM classification using the proposed method resulted in higher between-group classification accuracy compared to uncorrected data. Further, applying the proposed age correction substantially improved univariate detection of disease-related grey matter atrophy using VBM in AD patients differing in age from control subjects. The results suggest that the approach proposed in this work is generally suited to control for confounding variables such as age in SVM or VBM analyses. Accordingly, the approach might improve and extend the application of these methods in clinical neurosciences.     

Gene regulatory network reconstruction by Bayesian integration of prior knowledge and/or different experimental conditions

There have been various attempts to improve the reconstruction of gene regulatory networks from microarray data by the systematic integration of biological prior knowledge. Our approach is based on pioneering work by Imoto et al. where the prior knowledge is expressed in terms of energy functions, from which a prior distribution over network structures is obtained in the form of a Gibbs distribution. The hyperparameters of this distribution represent the weights associated with the prior knowledge relative to the data. We have derived and tested a Markov chain Monte Carlo (MCMC) scheme for sampling networks and hyperparameters simultaneously from the posterior distribution, thereby automatically learning how to trade off information from the prior knowledge and the data. We have extended this approach to a Bayesian coupling scheme for learning gene regulatory networks from a combination of related data sets, which were obtained under different experimental conditions and are therefore potentially associated with different active subpathways. The proposed coupling scheme is a compromise between (1) learning networks from the different subsets separately, whereby no information between the different experiments is shared; and (2) learning networks from a monolithic fusion of the individual data sets, which does not provide any mechanism for uncovering differences between the network structures associated with the different experimental conditions. We have assessed the viability of all proposed methods on data related to the Raf signaling pathway, generated both synthetically and in cytometry experiments.     

Loss of PTEN expression in neuroendocrine pancreatic tumors

PTEN (phosphatase and tensin homologue deleted from chromosome 10) is a well established tumor suppressor gene, which was cloned to chromosome 10q23. PTEN plays an important role in controlling cell growth, apoptosis, cell adhesion, and cell migration. In various studies, a genetic change as well as loss of PTEN expression by different carcinomas has been described. To date, the role of PTEN as a differentiation marker for neuroendocrine tumors (NET) and for the loss of PTEN expression is still unknown. It is assumed that loss of PTEN expression is important for tumor progression of NETs. We hypothesize that PTEN might be used as a new prognostic marker. We report 38 patients with a NET of the pancreas. Tumor tissues were surgically resected, fixed in formalin, and embedded in paraffin. PTEN expression was evaluated by immunohistochemistry and was correlated with several clinical and pathological parameters of each individual tumor. After evaluation of our immunohistochemistry data using a modified Remmele Score, a widely accepted method for categorizing staining results for reports and statistical evaluation, staining results of PTEN expression were correlated with the clinical and pathological parameters of each individual tumor. Our data demonstrates a significant difference in survival with existence of lymph node or distant metastases. Negative patients show a significant better survival compared with positive patients. Furthermore, we show a significant difference between PTEN expression and WHO or TNM classification. Taken together, our data shows a positive correlation between WHO classification and the new TNM classification of NETs, and loss of PTEN expression as well as survival. These results strongly implicate that PTEN might be helpful as a new prognostic factor.     

Complex events initiated by individual spikes in the human cerebral cortex

Synaptic interactions between neurons of the human cerebral cortex were not directly studied to date. We recorded the first dataset, to our knowledge, on the synaptic effect of identified human pyramidal cells on various types of postsynaptic neurons and reveal complex events triggered by individual action potentials in the human neocortical network. Brain slices were prepared from nonpathological samples of cortex that had to be removed for the surgical treatment of brain areas beneath association cortices of 58 patients aged 18 to 73 y. Simultaneous triple and quadruple whole-cell patch clamp recordings were performed testing mono- and polysynaptic potentials in target neurons following a single action potential fired by layer 2/3 pyramidal cells, and the temporal structure of events and underlying mechanisms were analyzed. In addition to monosynaptic postsynaptic potentials, individual action potentials in presynaptic pyramidal cells initiated long-lasting (37 ± 17 ms) sequences of events in the network lasting an order of magnitude longer than detected previously in other species. These event series were composed of specifically alternating glutamatergic and GABAergic postsynaptic potentials and required selective spike-to-spike coupling from pyramidal cells to GABAergic interneurons producing concomitant inhibitory as well as excitatory feed-forward action of GABA. Single action potentials of human neurons are sufficient to recruit Hebbian-like neuronal assemblies that are proposed to participate in cognitive processes.     

Two generalized algorithms measuring phase–amplitude cross-frequency coupling in neuronal oscillations network

An increasing number of studies pays attention to cross-frequency coupling in neuronal oscillations network, as it is considered to play an important role in exchanging and integrating of information. In this study, two generalized algorithms, phase–amplitude coupling-evolution map approach and phase–amplitude coupling-conditional mutual information which have been developed and applied originally in an identical rhythm, are generalized to measure cross-frequency coupling. The effectiveness of quantitatively distinguishing the changes of coupling strength from the measurement of phase–amplitude coupling (PAC) is demonstrated based on simulation data. The data suggest that the generalized algorithms are able to effectively evaluate the strength of PAC, which are consistent with those traditional approaches, such as PAC-PLV and PAC-MI. Experimental data, which are local field potentials obtained from anaesthetized SD rats, have also been analyzed by these two generalized approaches. The data show that the theta–low gamma PAC in the hippocampal CA3–CA1 network is significantly decreased in the glioma group compared to that in the control group. The results, obtained from either simulation data or real experimental signals, are consistent with that of those traditional approaches PAC-MI and PAC-PLV. It may be considered as a proper indicator for the cross frequency coupling in sub-network, such as the hippocampal CA3 and CA1.     

EEG classification by learning vector quantization.

EEG classification using Learning Vector Quantization (LVQ) is introduced on the basis of a Brain-Computer Interface (BCI) built in Graz, where a subject controlled a cursor in one dimension on a monitor using potentials recorded from the intact scalp. The method of classification with LVQ is described in detail along with first results on a subject who participated in four on-line cursor control sessions. Using this data, extensive off-line experiments were performed to show the influence of the various parameters of the classifier and the extracted features of the EEG on the classification results.     

Rehabilitation after Cardiac Illness

Though the total number of patients with cardiac illness referred to a work classification clinic was relatively small (57), the results obtained in terms of patients at work in the year after assessment is encouraging. It is suggested that a comparatively simple method can help in the rehabilitation of patients, often by removing minor anxieties, and by improving communication among the various agencies interested in the welfare and employment of the individual patient.     

A mathematical model of the effects of spatio-temporal patterns of dendritic input potentials on neuronal somatic potentials

A mathematical model of a neuron was developed to investigate the effects of various spatio-temporal patterns of miniature dendritic input potentials on neuronal somatic potentials. The model treats spatio-temporal dendritic activation patterns as the input or forcing function in the non-homogeneous cable equation. The theoretical somatic potentials resulting from several different spatio-temporal input patterns were generated on an IBM 360/75 computer. The model allows the investigation of the theoretical effects of activations at several different synaptic sites, of repeated activations at one or more sites, and of changes in various parameters. The time-to-peak and amplitude of individual excitatory postsynaptic potentials, distance of synapses from the soma, and interactivation interval for repeated activations at the same synapse were among the parameters investigated. Not only the order of activations at different synaptic sites was important, but the time intervals between activations were shown to be important. For a given order of activations at different synapses, optimum time intervals between activations were demonstrated, with respect to the resulting peak somatic potentials. Possible consequences of some hypothetical learning and memory mechanisms upon neuronal excitability were discussed. It was also shown that a deterministic model can generate theoretical curves which appear to be almost of a random nature with respect to the observed numbers and amplitudes of peaks of individual EPSPs. The conditions for the appearance of extra peaks were discussed.This paper describes work done as a portion of G.M.B.'s Ph. D. thesis in biomathematics at N.C. State University, and was supported by Public Health Service Grant No. GM-678 from the National Institute of General Medical Sciences.     

An experiment on animal re-identification from video

In the face of the global concern about climate change and endangered ecosystems, monitoring individual animals is of paramount importance. Computer vision methods for animal recognition and re-identification from video or image collections are a modern alternative to more traditional but intrusive methods such as tagging or branding. While there are many studies reporting results on various animal re-identification databases, there is a notable lack of comparative studies between different classification methods. In this paper we offer a comparison of 25 classification methods including linear, non-linear and ensemble models, as well as deep learning networks. Since the animal databases are vastly different in characteristics and difficulty, we propose an experimental protocol that can be applied to a chosen data collections. We use a publicly available database of five video clips, each containing multiple identities (9 to 27), where the animals are typically present as a group in each video frame. Our experiment involves five data representations: colour, shape, texture, and two feature spaces extracted by deep learning. In our experiments, simpler models (linear classifiers) and just colour feature space gave the best classification accuracy, demonstrating the importance of running a comparative study before resorting to complex, time-consuming, and potentially less robust methods.     

A model of the long-term response of carbon allocation and productivity of forests to increased CO2concentration and nitrogen deposition

We present a simple theoretical analysis of the long term response of forest growth and carbon allocation to increased atmospheric [CO₂] and N deposition. Our analysis is based on a recent model which predicts that plant light-use efficiency increases with [CO₂] but is independent of plant N supply. We combine that model with simple assumptions for nitrogen fluxes in the soil. A quasi-equilibrium analysis of the short term tree and soil pools is then used to develop a simple graphical depiction of the long term carbon and nitrogen supply constraints on total growth, stem growth and foliar allocation. Our results suggest that long-term growth responses to [CO₂] and N deposition depend strongly on the extent to which stem allocation and foliage allocation are coupled. At one extreme (‘no coupling’), when stem allocation is fixed and independent of foliage allocation, there is no response of total growth or stem growth to increased [CO₂] unless N deposition increases. At the other extreme (‘linear coupling’), when stem allocation is proportional to foliage allocation, there is a significant long-term increase in total growth following a doubling of [CO₂], even when N deposition is unchanged, but stem growth decreases because of a long-term decrease in foliage allocation. For both types of coupling, total growth and stem growth increase with increasing N deposition. In the case of linear coupling, however, the N deposition response of stem growth is significantly larger than that of total growth, because of a long-term increase in foliage allocation. We compare our results with those obtained previously from an alternative model of canopy light-use efficiency involving a dependence on the foliar N:C ratio in addition to [CO₂]. Our results highlight the need for more experimental information on (i) the extent to which canopy light-use efficiency is independent of N supply, and (ii) the relationship between foliage allocation and stem allocation.     

LCA impact assessment categories

A technical framework is presented to evaluate the strengths and the limitations of LCA impact assessment categories to yield accurate, useful results. The framework integrates the inherent characteristics of life-cycle inventory (LCI) data sets, characteristics of individual impact categories, how impact categories are defined, and the models used to characterize different categories. The sources for uncertainty in impact assessment are derived from the basic LCI procedures and the complexity of environmental processes and mechanisms. The noteworthy LCI procedures are: (1) the collection and aggregation of data across a comprehensive product system, (2) co-product and recycling allocation for releases and resources, and (3) the conversion of these data by functional unit calculations. These operations largely remove spatial and temporal considerations, resulting in analytical and interpretive limitations that vary in magnitude for different impact assessment categories. The framework shows two groups of categories where LCA results may be insufficient for making comparisons: (1) categories that involve local and/or transient processes and (2) categories that involve non-mass loading, biological parameters, such as biodiversity, habitat alteration, and toxicity. The framework also shows that how impact categories are defined complicates their use. Some categories are based on objective stressor-effect networks using known environmental mechanisms. In contrast, other categories are defined using various levels of subjective judgment to address either highly complex or unknown mechanisms. Finally, the framework shows that differences in the quality and detail of information provided by various models used during characterization also influence the accuracy and usefulness of the results. In summary, the framework indicates that (1) the various uncertainties in each individual category have a a number of different technical origins and that (2) the degree of uncertainty varies significantly between categories. As a result, interpretation and valuation cannot presume an equivalency of processes or merit behind numerical values for different categories. The framework can be used to initially identify and track these uncertainties to improve LCA impact assessment interpretation and application.     

Variability in the role of the gasbladder in fish audition

The teleost gasbladder is believed to aid in fish audition by transferring pressure components of incoming sound to the inner ears. This idea is primarily based on both anatomical observations of the mechanical connection between the gasbladder and the ear, followed by physiological experiments by various researchers. The gasbladder movement has been modeled mathematically as a pulsating bubble. This study is extending the previous work on fish with a physical coupling of the gasbladder and ear by investigating hearing in two species (the blue gourami Trichogaster trichopterus, and the oyster toadfish Opsanus tau) without a mechanical linkage. An otophysan specialist (the goldfish Carassius auratus) with mechanical coupling, is used as the control. Audiograms were obtained with acoustically evoked potentials (e.g., auditory brainstem response) from intact fish and from the same individuals with their gasbladders deflated. In blue gourami and oyster toadfish, removal of gas did not significantly change thresholds, and evoked potentials had similar waveforms. In goldfish thresholds increased by 33–55 dB (frequency dependent) after deflation, and major changes in evoked potentials were observed. These results suggest that the gasbladder may not serve an auditory enhancement function in teleost fishes that lack mechanical coupling between the gasbladder and the inner ear. Accepted: 28 February 2000     

Factors controlling resource allocation in mountain birch

We present a comprehensive analysis of factors affecting resource allocation and crown formation in a subarctic birch tree, Betula pubescens ssp. czerepanovii (Orlova) Hämet-Ahti. Using biomass measurements and digitized data on tree architecture, we investigated several hypotheses on various factors that may modify plant growth. We also analyzed the extent to which different mechanisms operate at different scales, ranging from individual shoots to the whole branches or trees. Different factors affected allocation at different levels of organization. Stem age had a minor effect, suggesting that similar control mechanisms operate at all stages of development. Fates of individual shoots were affected by their local growing conditions as indicated, for example, by the dependence of long shoot production on light. Buds formed in the current long shoots were likely to become new long shoots. In the innermost crown parts, radial growth had priority compared to long shoot production. Elongation of individual long shoots was controlled by two conflicting factors. Long distance from the roots suppressed growth, probably indicating costs associated with resource transportation, whereas a high level of light augmented growth. In contrast, growth of entire branches was not so clearly related to the availability of resources, but showed limitation due to allometric scaling. This set a relationship between the maximum long shoot number and the overall branch size, and may indicate allometric constraints to the way a tree is constructed. Strict allometric relationships existed also between other structural traits of mountain birch, most of them similar at all levels of branching hierarchy. However, despite the upper level restrictions set by allometry, source-sink interactions and localized responses of individual shoots operated as local processes that directed allocation towards the most favourable positions. This may be a mechanism for achieving efficient tree architecture in terms of resource intake and costs of transportation.     

Extracting glycan motifs using a biochemicallyweighted kernel

Carbohydrates, or glycans, are one of the most abundant and structurally diverse biopolymers constitute the third major class of biomolecules, following DNA and proteins. However, the study of carbohydrate sugar chains has lagged behind compared to that of DNA and proteins, mainly due to their inherent structural complexity. However, their analysis is important because they serve various important roles in biological processes, including signaling transduction and cellular recognition. In order to glean some light into glycan function based on carbohydrate structure, kernel methods have been developed in the past, in particular to extract potential glycan biomarkers by classifying glycan structures found in different tissue samples. The recently developed weighted qgram method (LK-method) exhibits good performance on glycan structure classification while having limitations in feature selection. That is, it was unable to extract biologically meaningful features from the data. Therefore, we propose a biochemicallyweighted tree kernel (BioLK-method) which is based on a glycan similarity matrix and also incorporates biochemical information of individual q-grams in constructing the kernel matrix. We further applied our new method for the classification and recognition of motifs on publicly available glycan data. Our novel tree kernel (BioLK-method) using a Support Vector Machine (SVM) is capable of detecting biologically important motifs accurately while LK-method failed to do so. It was tested on three glycan data sets from the Consortium for Functional Glycomics (CFG) and Kyoto Encyclopedia of Genes and Genomes (KEGG) GLYCAN and showed that the results are consistent with the literature. The newly developed BioLK-method also maintains comparable classification performance with the LK-method. Our results obtained here indicate that the incorporation of biochemical information of q-grams further shows the flexibility and capability of the novel kernel in feature extraction, which may aid in the prediction of glycan biomarkers.     

Identification of characteristic plant co-occurrences in neotropical secondary montane forests

Aims Inferring environmental conditions from characteristic patterns of plant co-occurrences can be crucial for the development of conservation strategies concerning secondary neotropical forests. However, no methodological agreement has been achieved so far regarding the identification and classification of characteristic groups of vascular plant species in the tropics. This study examines botanical and, in particular, statistical aspects to be considered in such analyses. Based on these, we propose a novel data-driven approach for the identification of characteristic plant co-occurrences in neotropical secondary mountain forests.Methods Floristic inventory data were gathered in secondary tropical mountain forests in Ecuador. Vegetation classification was performed by coupling locally adaptive isometric feature mapping, a non-linear ordination method and fuzzy- c -means clustering. This approach was designed for dealing with underlying non-linearities and uncertainties in the inventory data.Important findings The results indicate that the applied non-linear mapping in combination with fuzzy classification of species occurrence allows an effective identification of characteristic groups of co-occurring species as fuzzy-defined clusters. The selected species indicated groups representing characteristic life-form distributions, as they correspond to various stages of forest regeneration. Combining the identified 'characteristic species groups' with meta-information derived from accompanying studies indicated that the clusters can also be related to habitat conditions. In conclusion, we identified species groups either characteristic of different stages of forest succession after clear-cutting or of impact by fire or a landslide. We expect that the proposed data-mining method will be useful for vegetation classification where no a priori knowledge is available.     

Statistical inference for simultaneous clustering of gene expression data

Current methods for analysis of gene expression data are mostly based on clustering and classification of either genes or samples. We offer support for the idea that more complex patterns can be identified in the data if genes and samples are considered simultaneously. We formalize the approach and propose a statistical framework for two-way clustering. A simultaneous clustering parameter is defined as a function theta=Phi(P) of the true data generating distribution P, and an estimate is obtained by applying this function to the empirical distribution P(n). We illustrate that a wide range of clustering procedures, including generalized hierarchical methods, can be defined as parameters which are compositions of individual mappings for clustering patients and genes. This framework allows one to assess classical properties of clustering methods, such as consistency, and to formally study statistical inference regarding the clustering parameter. We present results of simulations designed to assess the asymptotic validity of different bootstrap methods for estimating the distribution of Phi(P(n)). The method is illustrated on a publicly available data set.     

Socioeconomic determinants of rates of consultation in general practice based on fourth national morbidity survey of general practices.

OBJECTIVE--To identify the socioeconomic determinants of consultation rates in general practice. DESIGN--Analysis of data from the fourth national morbidity survey of general practices (MSGP4) including sociodemographic details of individual patients and small area statistics from the 1991 census. Multilevel modelling techniques were used to take account of both individual patient data and small area statistics to relate socioeconomic and health status factors directly to a measure of general practitioner workload. RESULTS--Higher rates of consultations were found in patients who were classified as permanently sick, unemployed (especially those who became unemployed during the study year), living in rented accommodation, from the Indian subcontinent, living with a spouse or partner (women only), children living with two parents (girls only), and living in urban areas, especially those living relatively near the practice. When characteristics of individual patients are known and controlled for the role of "indices of deprivation" is considerably reduced. The effect of individual sociodemographic characteristics were shown to vary between different areas. CONCLUSIONS--Demographic and socioeconomic factors can act as powerful predictors of consultation patterns. Though it will always be necessary to retain some local planning discretion, the sets of coefficients estimated for individual level factors, area level characteristics, and for practice groupings may be sufficient to provide an indicative level of demand for general medical services. Although the problems in using socioeconomic data from individual patients would be substantial, these results are relevant to the development of a resource allocation formula for general practice.