Predicting smear negative pulmonary tuberculosis with classification trees and logistic regression: a cross-sectional study

Abstract Background Smear negative pulmonary tuberculosis (SNPT) accounts for 30% of pulmonary tuberculosis cases reported yearly in Brazil. This study aimed to develop a prediction model for SNPT for outpatients in areas with scarce resources. Methods The study enrolled 551 patients with clinical-radiological suspicion of SNPT, in Rio de Janeiro, Brazil. The original data was divided into two equivalent samples for generation and validation of the prediction models. Symptoms, physical signs and chest X-rays were used for constructing logistic regression and classification and regression tree models. From the logistic regression, we generated a clinical and radiological prediction score. The area under the receiver operator characteristic curve, sensitivity, and specificity were used to evaluate the model's performance in both generation and validation samples. Results It was possible to generate predictive models for SNPT with sensitivity ranging from 64% to 71% and specificity ranging from 58% to 76%. Conclusion The results suggest that those models might be useful as screening tools for estimating the risk of SNPT, optimizing the utilization of more expensive tests, and avoiding costs of unnecessary anti-tuberculosis treatment. Those models might be cost-effective tools in a health care network with hierarchical distribution of scarce resources.

The Aurora A and B kinases are up-regulated in bone marrow-derived chronic lymphocytic leukemia cells and represent potential therapeutic targets

Background The malignant B cells in chronic lymphocytic leukemia receive signals from the bone marrow and lymph node microenvironments which regulate their survival and proliferation. Characterization of these signals and the pathways that propagate them to the interior of the cell is important for the identification of novel potential targets for therapeutic intervention. Design and Methods We compared the gene expression profiles of chronic lymphocytic leukemia B cells purified from bone marrow and peripheral blood to identify genes that are induced by the bone marrow microenvironment. Two of the differentially expressed genes were further studied in cell culture experiments and in an animal model to determine whether they could represent appropriate therapeutic targets in chronic lymphocytic leukemia. Results Functional classification analysis revealed that the majority of differentially expressed genes belong to gene ontology categories related to cell cycle and mitosis. Significantly up-regulated genes in bone marrow-derived tumor cells included important cell cycle regulators, such as Aurora A and B, survivin and CDK6. Down-regulation of Aurora A and B by RNA interference inhibited proliferation of chronic lymphocytic leukemia-derived cell lines and induced low levels of apoptosis. A similar effect was observed with the Aurora kinase inhibitor VX-680 in primary chronic lymphocytic leukemia cells that were induced to proliferate by CpG-oligonucleotides and interleukin-2. Moreover, VX-680 significantly blocked leukemia growth in a mouse model of chronic lymphocytic leukemia. Conclusions Aurora A and B are up-regulated in proliferating chronic lymphocytic leukemia cells and represent potential therapeutic targets in this disease.

2D-DIGE analysis of mango (Mangifera indica L.) fruit reveals major proteomic changes associated with ripening

A comparative proteomic investigation between the pre-climacteric and climacteric mango fruits (cv. Keitt) was performed to identify protein species with variable abundance during ripening. Proteins were phenol-extracted from fruits, cyanine-dye-labeled, and separated on 2D gels at pH 4-7. Total spot count of about 373 proteins spots was detected in each gel and forty-seven were consistently different between pre-climacteric and climacteric fruits and were subjected to LC-MS/MS analysis. Functional classification revealed that protein species involved in carbon fixation and hormone biosynthesis decreased during ripening, whereas those related to catabolism and the stress-response, including oxidative stress and abiotic and pathogen defense factors, accumulated. In relation to fruit quality, protein species putatively involved in color development and pulp softening were also identified. This study on mango proteomics provides an overview of the biological processes that occur during ripening. (C) 2012 Elsevier B.V. All rights reserved.

Florística e fitossociologia em parcelas permanentes da Mata Atlântica do sudeste do Brasil ao longo de um gradiente altitudinal

Este trabalho resume os dados de florística e fitossociologia de 11, das 14 parcelas de 1 ha, alocadas ao longo do gradiente altitudinal da Serra do Mar, São Paulo, Brasil. As parcelas começam na cota 10 m (Floresta de Restinga da Praia da Fazenda, município de Ubatuba) e estão distribuídas até a cota 1100 m (Floresta Ombrófila Densa Montana da Trilha do rio Itamambuca, município de São Luis do Paraitinga) abrangendo os Núcleos Picinguaba e Santa Virgínia do Parque Estadual da Serra do Mar. Na Restinga o solo é Neossolo Quartzarênico francamente arenoso, enquanto que na encosta o solo é um Cambisolo Háplico Distrófico argilo-arenoso, sendo que todas as parcelas apresentaram solo ácido (pH 3 – 4) com alta diluição de nutrientes e alta saturação de alumínio. Na Restinga e no sopé da encosta o clima é Tropical/Subtropical Úmido (Af/Cfa), sem estação seca, com precipitação média anual superior a 2.200 mm e temperatura média anual de 22 °C. Subindo a encosta mantêm-se a média de precipitação, mas há um gradativo resfriamento, de forma que a 1.100 m o clima é Subtropical Úmido (Cfa/Cfb), sem estação seca, com temperatura média anual de 17 °C. Destaca-se ainda que, quase diariamente, a parte superior da encosta, geralmente acima de 400 m, é coberta por uma densa neblina. Nas 14 parcelas foram marcados, medidos e amostrados 21.733 indivíduos com DAP ≥ 4,8 cm, incluindo árvores, palmeiras e fetos arborescentes. O número médio de indivíduos amostrados nas 14 parcelas foi de 1.264 ind.ha–1 (± 218 EP de 95%). Dentro dos parâmetros considerados predominaram as árvores (71% FOD Montana a 90% na Restinga), seguidas de palmeiras (10% na Restinga a 25% na FOD Montana) e fetos arborescentes (0% na Restinga a 4% na FOD Montana). Neste aspecto destaca-se a FOD Terras Baixas Exploradas com apenas 1,8% de palmeiras e surpreendentes 10% de fetos arborescentes. O dossel é irregular, com altura variando de 7 a 9 m, raramente as árvores emergentes chegam a 18 m, e a irregularidade do dossel permite a entrada de luz suficiente para o desenvolvimento de centenas de espécies epífitas. Com exceção da FOD Montana, onde o número de mortos foi superior a 5% dos indivíduos amostrados, nas demais fitofisionomias este valor ficou abaixo de 2,5%. Nas 11 parcelas onde foi realizado o estudo florístico foram encontradas 562 espécies distribuídas em 195 gêneros e 68 famílias. Apenas sete espécies – Euterpe edulis Mart. (Arecaceae), Calyptranthes lucida Mart. ex DC. e Marlierea tomentosa Cambess (ambas Myrtaceae), Guapira opposita (Vell.) Reitz (Nyctaginaceae), Cupania oblongifolia Mart. (Sapindaceae) e as Urticaceae Cecropia glaziovii Snethl. e Coussapoa microcarpa (Schott) Rizzini – ocorreram da Floresta de Restinga à FOD Montana, enquanto outras 12 espécies só não ocorreram na Floresta de Restinga. As famílias com o maior número de espécies são Myrtaceae (133 spp), Fabaceae (47 spp), 125 Fitossociologia em parcelas permanentes de Mata Atlântica http://www.biotaneotropica.org.br/v12n1/pt/abstract?article+bn01812012012 http://www.biotaneotropica.org.br Biota Neotrop., vol. 12, no. 1 Introdução A Mata Atlântica sensu lato (Joly et al. 1999) é a segunda maior floresta tropical do continente americano (Tabarelli et al. 2005). A maior parte dos Sistemas de Classificação da vegetação brasileira reconhece que no Domínio Atlântico (sensu Ab’Saber 1977) esse bioma pode ser dividido em dois grandes grupos: a Floresta Ombrófila Densa, típica da região costeira e das escarpas serranas com alta pluviosidade (Mata Atlântica – MA – sensu stricto), e a Floresta Estacional Semidecidual, que ocorre no interior, onde a pluviosidade, além de menor, é sazonal. Na região costeira podem ocorrer também Manguezais (Schaeffer-Novelli 2000), ao longo da foz de rios de médio e grande porte, e as Restingas (Scarano 2009), crescendo sobre a planície costeira do quaternário. No topo das montanhas, geralmente acima de 1500 m, estão os Campos de Altitude (Ribeiro & Freitas 2010). Em 2002, a Fundação SOS Mata Atlântica em parceria com o INPE (Instituto..., 2002) realizaram um levantamento que indica que há apenas 7,6% da cobertura original da Mata Atlântica (s.l.). Mais recentemente Ribeiro et al. (2009) refinaram a estimativa incluindo fragmentos menores, que não haviam sido contabilizados, e concluíram que resta algo entre 11,4 e 16% da área original. Mesmo com esta fragmentação, o mosaico da Floresta Atlântica brasileira possui um dos maiores níveis de endemismos do mundo (Myers et al. 2000) e cerca da metade desses remanescentes de grande extensão estão protegidos na forma de Unidades de Conservação (Galindo & Câmara 2005). Entre os dois centros de endemismo reconhecidos para a MA (Fiaschi & Pirani 2009), o bloco das regiões sudeste/sul é o que conserva elementos da porção sul de Gondwana (Sanmartin & Ronquist 2004), tido como a formação florestal mais antiga do Brasil (Colombo & Joly 2010). Segundo Hirota (2003), parte dos remanescentes de MA está no estado de São Paulo, onde cerca de 80% de sua área era coberta por florestas (Victor 1977) genericamente enquadradas como Mata Atlântica “sensu lato” (Joly et al. 1999). Dados de Kronka et al. (2005) mostram que no estado restam apenas 12% de área de mata e menos do que 5% são efetivamente florestas nativas pouco antropizadas. Nos 500 anos de fragmentação e degradação das formações naturais, foram poupadas apenas as regiões serranas, principalmente a fachada da Serra do Mar, por serem impróprias para práticas agrícolas. Usando o sistema fisionômico-ecológico de classificação da vegetação brasileira adotado pelo IBGE (Veloso et al. 1991), a Floresta Ombrófila Densa, na área de domínio da Mata Atlântica, foi subdividida em quatro faciações ordenadas segundo a hierarquia topográfica, que refletem fisionomias de acordo com as variações das faixas altimétricas e latitudinais. No estado de São Paulo, na latitude entre 16 e 24 °S temos: 1) Floresta Ombrófila Densa das Terras Baixas - 5 a 50 m de altitude; 2) Floresta Ombrófila Densa Submontana – no sopé da Serra do Mar, com cotas de altitude variando entre 50 e 500 m; 3) Floresta Ombrófila Densa Montana – recobrindo a encosta da Serra do Mar propriamente dita, em altitudes que variam de 500 a 1.200 m; 4) Floresta Ombrófila Densa Altimontana – ocorrendo no topo da Serra do Mar, acima dos limites estabelecidos para a formação montana, onde a vegetação praticamente deixa de ser arbórea, pois predominam os campos de altitude. Nas últimas três décadas muita informação vem sendo acumulada sobre a composição florística e a estrutura do estrato arbóreo dos remanescentes florestais do estado, conforme mostram as revisões de Oliveira-Filho & Fontes (2000) e Scudeller et al. (2001). Em florestas tropicais este tipo de informação, assim como dados sobre a riqueza de espécies, reflete não só fatores evolutivos e biogeográficos, como também o histórico de perturbação, natural ou antrópica, das respectivas áreas (Gentry 1992, Hubbell & Foster 1986). A síntese dessas informações tem permitido a definição de unidades fitogeográficas com diferentes padrões de riqueza de espécies e apontam para uma diferenciação, entre as florestas paulistas, no sentido leste/oeste (Salis et al. 1995, Torres et al. 1997, Santos et al. 1998). Segundo Bakker et al. (1996) um método adequado para acompanhar e avaliar as mudanças na composição das espécies e dinâmica da floresta ao longo do tempo é por meio de parcelas permanentes (em inglês Permanent Sample Plots –PSPs). Essa metodologia tem sido amplamente utilizada em estudos de longa duração em florestas tropicais, pois permite avaliar a composição e a estrutura florestal e monitorar sua mudança no tempo (Dallmeier 1992, Condit 1995, Sheil 1995, Malhi et al. 2002, Lewis et al. 2004). Permite avaliar também as consequências para a floresta de problemas como o aquecimento global e a poluição atmosférica (Bakker et al. 1996). No Brasil os projetos/programas que utilizam a metodologia de Parcelas Permanentes tiveram origem, praticamente, com o Projeto Rubiaceae (49) e Lauraceae (49) ao longo de todo gradiente da FOD e Monimiaceae (21) especificamente nas parcelas da FOD Montana. Em termos de número de indivíduos as famílias mais importantes foram Arecaceae, Rubiaceae, Myrtaceae, Sapotaceae, Lauraceae e na FOD Montana, Monimiaceae. Somente na parcela F, onde ocorreu exploração de madeira entre 1960 e 1985, a abundância de palmeiras foi substituída pelas Cyatheaceae. O gradiente estudado apresenta um pico da diversidade e riqueza nas altitudes intermediárias (300 a 400 m) ao longo da encosta (índice de Shannon-Weiner - H’ - variando de 3,96 a 4,48 nats.indivíduo–1). Diversas explicações para este resultado são apresentadas neste trabalho, incluindo o fato dessas altitudes estarem nos limites das expansões e retrações das diferentes fitofisionomias da FOD Atlântica durante as flutuações climáticas do Pleistoceno. Os dados aqui apresentados demonstram a extraordinária riqueza de espécies arbóreas da Floresta Ombrófila Densa Atlântica dos Núcleos Picinguaba e Santa Virgínia do Parque Estadual da Serra do Mar, reforçando a importância de sua conservação ao longo de todo o gradiente altitudinal. A diversidade desta floresta justifica também o investimento de longo prazo, através de parcelas permanentes, para compreender sua dinâmica e funcionamento, bem como monitorar o impacto das mudanças climáticas nessa vegetação.

Structural and Functional Insights from the Metagenome of an Acidic Hot Spring Microbial Planktonic Community in the Colombian Andes

A taxonomic and annotated functional description of microbial life was deduced from 53 Mb of metagenomic sequence retrieved from a planktonic fraction of the Neotropical high Andean (3,973 meters above sea level) acidic hot spring El Coquito (EC). A classification of unassembled metagenomic reads using different databases showed a high proportion of Gammaproteobacteria and Alphaproteobacteria (in total read affiliation), and through taxonomic affiliation of 16S rRNA gene fragments we observed the presence of Proteobacteria, micro-algae chloroplast and Firmicutes. Reads mapped against the genomes Acidiphilium cryptum JF-5, Legionella pneumophila str. Corby and Acidithiobacillus caldus revealed the presence of transposase-like sequences, potentially involved in horizontal gene transfer. Functional annotation and hierarchical comparison with different datasets obtained by pyrosequencing in different ecosystems showed that the microbial community also contained extensive DNA repair systems, possibly to cope with ultraviolet radiation at such high altitudes. Analysis of genes involved in the nitrogen cycle indicated the presence of dissimilatory nitrate reduction to N2 (narGHI, nirS, norBCDQ and nosZ), associated with Proteobacteria-like sequences. Genes involved in the sulfur cycle (cysDN, cysNC and aprA) indicated adenylsulfate and sulfite production that were affiliated to several bacterial species. In summary, metagenomic sequence data provided insight regarding the structure and possible functions of this hot spring microbial community, describing some groups potentially involved in the nitrogen and sulfur cycling in this environment. Citation: Jimenez DJ, Andreote FD, Chaves D, Montana JS, Osorio-Forero C, et al. (2012) Structural and Functional Insights from the Metagenome of an Acidic Hot Spring Microbial Planktonic Community in the Colombian Andes. PLoS ONE 7(12): e52069. doi:10.1371/journal.pone.0052069

Seasonal effects on the relationship between photosynthesis and leaf carbohydrates in orange trees

To understand the effect of summer and winter on the relationships between leaf carbohydrate and photosynthesis in citrus trees growing in subtropical conditions, 'Valencia' orange trees were subjected to external manipulation of their carbohydrate concentration by exposing them to darkness and evaluating the maximal photosynthetic capacity. In addition, the relationships between carbohydrate and photosynthesis in the citrus leaves were studied under natural conditions. Exposing the leaves to dark conditions decreased the carbohydrate concentration and increased photosynthesis in both seasons, which is in accordance with the current model of carbohydrate regulation. Significant negative correlations were found between total non-structural carbohydrates and photosynthesis in both seasons. However, non-reducing sugars were the most important carbohydrate that apparently regulated photosynthesis on a typical summer day, whereas starch was important on a typical winter day. As a novelty, photosynthesis stimulation by carbohydrate consumption was approximately three times higher during the summer, i.e. the growing season. Under subtropical conditions, citrus leaves exhibited relatively high photosynthesis and high carbohydrate levels on the summer day, as well as a high nocturnal consumption of starch and soluble sugars. A positive association was determined between photosynthesis and photoassimilate consumption/exportation, even in leaves showing a high carbohydrate concentration. This paper provides evidence that photosynthesis in citrus leaves is regulated by an increase in sink demand rather than by the absolute carbohydrate concentration in leaves.

Height-diameter relationships of tropical Atlantic moist forest trees in southeastern Brazil

Site-specific height-diameter models may be used to improve biomass estimates for forest inventories where only diameter at breast height (DBH) measurements are available. In this study, we fit height-diameter models for vegetation types of a tropical Atlantic forest using field measurements of height across plots along an altitudinal gradient. To fit height-diameter models, we sampled trees by DBH class and measured tree height within 13 one-hectare permanent plots established at four altitude classes. To select the best model we tested the performance of 11 height-diameter models using the Akaike Information Criterion (AIC). The Weibull and Chapman-Richards height-diameter models performed better than other models, and regional site-specific models performed better than the general model. In addition, there is a slight variation of height-diameter relationships across the altitudinal gradient and an extensive difference in the stature between the Atlantic and Amazon forests. The results showed the effect of altitude on tree height estimates and emphasize the need for altitude-specific models that produce more accurate results than a general model that encompasses all altitudes. To improve biomass estimation, the development of regional height-diameter models that estimate tree height using a subset of randomly sampled trees presents an approach to supplement surveys where only diameter has been measured.

Soil-nutrient availability modifies the response of young pioneer and late successional trees to elevated carbon dioxide in a Brazilian tropical environment

The aim of this work was to determine the impact of three levels of [CO2] and two levels of soil-nutrient availability on the growth and physiological responses of two tropical tree species differing in their ecological group: Croton urucurana Baillon, a pioneer (P), and also Cariniana legalis (Martius) Kuntze, a late succession (LS). We aimed to test the hypothesis that P species have stronger response to elevated [CO2] than LS species as a result of differences in photosynthetic capacity and growth kinetics between both functional groups. Seedlings of both species were grown in open-top-chambers under high (HN) or low (LN) soil-nutrient supply and exposed to ambient (380 mu mol mol(-1)) or elevated (570 and 760 mu mol mol(-1)) [CO2]. Measurements of gas exchange, chlorophyll a fluorescence, seedling biomass and allocation were made after 70 days of treatment. Results suggest that elevated [CO2] significantly enhances the photosynthetic rates (A) and biomass production in the seedlings of both species, but that soil-nutrient supply has the potential to modify the response of young tropical trees to elevated [CO2]. In relation to plants grown in ambient [CO2], the P species grown under 760 mu mol mol(-1) [CO2] showed increases of 28% and 91% in A when grown in LN and HN, respectively. In P species grown under 570 mu mol mol(-1) [CO2], A increased by 16% under HN, but there was no effect in LN. In LS species, the enhancement of A by effect of 760 mu mol mol(-1) [CO2] was 30% and 70% in LN and HN, respectively. The exposure to 570 mu mol mol(-1) [CO2] stimulated A by 31% in HN, but was no effect in LN. Reductions in stomatal conductance (g(s)) and transpiration (E), as a result of elevated [CO2] were observed. Increasing the nutrient supply from low to high increased both the maximum rate of carboxylation (V-cmax) and maximum potential rate of electron transport (J(max)). As the level of [CO2] increased, both the V-cmax and the J(max) were found to decrease, whereas the J(max)/V-cmax ratio increased. In the LS species, the maximum efficiency of PSII (F-v/F-m) was higher in the 760 mu mol mol(-1) [CO2] treatment relative to other [CO2] treatments. The results suggest that when grown under HN and the highest [CO2], the performance of the P species C. urucurana, in terms of photosynthesis and biomass enhancement, is better than the LS species C. legalis. However, a larger biomass is allocated to roots when C. legalis seedlings were exposed to elevated [CO2]. This response would be an important strategy for plant survival and productivity of the LS species under drought stresses conditions on tropical environments in a global-change scenario. (C) 2011 Elsevier B.V. All rights reserved.

Incorporating label dependency into the binary relevance framework for multi-label classification

In multi-label classification, examples can be associated with multiple labels simultaneously. The task of learning from multi-label data can be addressed by methods that transform the multi-label classification problem into several single-label classification problems. The binary relevance approach is one of these methods, where the multi-label learning task is decomposed into several independent binary classification problems, one for each label in the set of labels, and the final labels for each example are determined by aggregating the predictions from all binary classifiers. However, this approach fails to consider any dependency among the labels. Aiming to accurately predict label combinations, in this paper we propose a simple approach that enables the binary classifiers to discover existing label dependency by themselves. An experimental study using decision trees, a kernel method as well as Naive Bayes as base-learning techniques shows the potential of the proposed approach to improve the multi-label classification performance.

Evolving decision trees with beam search-based initialization and lexicographic multi-objective evaluation

Decision tree induction algorithms represent one of the most popular techniques for dealing with classification problems. However, traditional decision-tree induction algorithms implement a greedy approach for node splitting that is inherently susceptible to local optima convergence. Evolutionary algorithms can avoid the problems associated with a greedy search and have been successfully employed to the induction of decision trees. Previously, we proposed a lexicographic multi-objective genetic algorithm for decision-tree induction, named LEGAL-Tree. In this work, we propose extending this approach substantially, particularly w.r.t. two important evolutionary aspects: the initialization of the population and the fitness function. We carry out a comprehensive set of experiments to validate our extended algorithm. The experimental results suggest that it is able to outperform both traditional algorithms for decision-tree induction and another evolutionary algorithm in a variety of application domains.