High Predictive Capacity Research Articles

Who Smells? Forecasting Taste and Odor in a Drinking Water Reservoir.

Taste and odor problems can impede public trust in drinking water and impose major costs on water utilities. The ability to forecast taste and odor events in source waters, in advance, is shown for the first time in this paper. This could allow water utilities to adapt treatment, and where effective treatment is not available, consumers could be warned. A unique 24-year time series, from an important drinking water reservoir in Saskatchewan, Canada, is used to develop forecasting models of odor using chlorophyll a, turbidity, total phosphorus, temperature, and the following odor producing algae taxa: Anabaena spp., Aphanizemenon spp., Oscillatoria spp., Chlorophyta, Cyclotella spp., and Asterionella spp. We demonstrate, using linear regression and random forest models, that odor events can be forecast at 0-26 week time lags, and that the models are able to capture a significant increase in threshold odor number in the mid-1990 s. Models with a fortnight time-lag show a high predictive capacity (R(2) = 0.71 for random forest; 0.52 for linear regression). Predictive skill declines for time lags from 0 to 15 weeks, then increases again, to R(2) values of 0.61 (random forest) and 0.48 (linear regression) at a 26-week lag. The random forest model is also able to provide accurate forecasting of TON levels requiring treatment 12 weeks in advance-93% true positive rate with a 0% false positive rate. Results of the random forest model demonstrate that phytoplankton taxonomic data outperform chlorophyll a in terms of predictive importance.

Physiological, anthropometric, strength, and muscle power characteristics correlates with running performance in young runners.

The purpose of this study was to investigate the relationship between physiological, anthropometric, strength, and muscle power variables and a 5-km time trial (5kmT) in young runners. Twenty-three runners volunteered to participate in this study. Height, body mass, body fat, and fat-free mass (FFM) were measured. The subjects underwent laboratory testing to determine maximal oxygen uptake ((Equation is included in full-text article.)), velocity at ventilatory threshold (VVT), running economy (RE), velocity associated with maximal oxygen uptake ((Equation is included in full-text article.)), and peak velocity (PV). Peak torque, total work, and power were measured by an isokinetic dynamometer at 60°·s(-1) and 240°·s(-1) angular velocities. Right and left knee flexor and extensor torques were evaluated. Finally, the participants performed a 5kmT. Multiple regression and correlation analysis were used to determine the variables that significantly related to 5kmT. Strength and muscle power variables did not correlate with 5kmT. However, most physiological variables were associated with 5kmT. Velocity at ventilatory threshold alone explains 40% of the variance in 5kmT. The addition of the RE at speed 11.2 km·h(-1) (RE11.2) and FFM to the prediction equation allowed for 71% of the adjusted variance in 5kmT to be predicted. These results show that strength and muscle power variables are not good predictors of 5kmT; however, the physiological variables presented high prediction capacity in the 5kmT. Moreover, the anthropometric measures showed significant influence in performance prediction.

Predictive spatial modelling of seasonal bottlenose dolphin (Tursiops truncatus) distributions in the Mississippi Sound

Abstract Spatial distribution models (SDMs) have been useful for improving management of species of concern in many areas. This study was designed to model the spatial distribution of bottlenose dolphins among seasons of the year in the Mississippi Sound within the northern Gulf of Mexico. Models were constructed by integrating presence locations of dolphins acquired from line‐transect sampling from 2011–2013 with maps of environmental conditions for the region to generate a likelihood of dolphin occurrence for winter (January–March), spring (April–June), summer (July–September), and autumn (October–December) using maximum entropy. Models were successfully generated using the program MaxEnt and had high predictive capacity for all seasons (AUC (area under curve) > 0.8). Distinct seasonal shifts in spatial distribution were evident including increased predicted occurrence in deepwater habitats during the winter, limited predicted occurrence in the western Mississippi Sound in winter and spring, widespread predicted occurrence over the entire region during summer, and a distinct westward shift of predicted occurrence in autumn. The most important environmental predictors used in SDMs were distance to shore, salinity, and nitrates, but variable importance differed considerably among seasons. Geographic shifts in predicted occurrence probably reflect both direct effects of changing environmental conditions and subsequent changes in prey availability and foraging efficiency. Overall, seasonal models helped to identify preferred habitats for dolphins among seasons of the year and can be used to inform management of this protected species in the northern Gulf of Mexico. Copyright © 2015 John Wiley & Sons, Ltd.

Application of adaptive neuro-fuzzy inference system (ANFIS) to estimate the biochemical oxygen demand (BOD) of Surma River

This paper describes the application of adaptive neuro-fuzzy inference system (ANFIS) to estimate the biochemical oxygen demand (BOD) of Surma River of Bangladesh. The data sets consist of 10 water quality parameters which include pH, alkalinity (mg/L as CaCO3), hardness, total solids (TS), total dissolved solids (TDS), potassium (K+), PO4−3 (mg/l), NO3− (mg/l), BOD (mg/l) and DO (mg/l). The performance of the ANFIS models was assessed through the correlation coefficient (R), mean squared error (MSE), mean absolute error (MAE) and Nash model efficiency (E). Study results show that the adaptive neuro-fuzzy inference system is able to predict the biochemical oxygen demand with reasonable accuracy, suggesting that the ANFIS model is a valuable tool for river water quality estimation. The result shows that, ANFIS-I has a high prediction capacity of BOD compared with ANFIS-II. The results also suggest that ANFIS method can be successfully applied to establish river water quality prediction model.

The mathematical limits of genetic prediction for complex chronic disease

BackgroundAttempts at predicting individual risk of disease based on common germline genetic variation have largely been disappointing. The present paper formalises why genetic prediction at the individual level is and...

Quantitative structure activity relationship and binding investigation of N-alkyl glycine amides as inhibitors of Leukotriene A4 hydrolase

The leukotriene A4 hydrolase (LTA4H) is a bifunctional zinc enzyme that catalyzes the final (rate-limiting) step in the synthesis of leukotriene B4 (LTB4), which is involved in several diseases. Many pharmaceutical attempts to exploit the LTA4H/LTB4 pathway have been unsatisfactory, hence, the development of new inhibitory drugs is essential. This paper describes the generation of a quantitative structure–activity relationship (QSAR) model on a series of 50 N-alkyl glycine amides with experimentally defined IC50. In addition, the optimized molecular structures of the inhibitors were docked into the active site of the enzyme to identify the enzyme-ligand interactions and quantify the estimated free energy of binding (ΔGbind). A simple four-descriptor QSAR model with high predictive capacity was obtained. The statistic parameters of the model are: regression coefficient (Rtest) of 0.714 and a standard deviation (Stest) of 0.696. The predicted inhibitory activity of 85 new N-alkyl glycine amides compounds was obtained with this QSAR model and these compounds were docked into LTA4H. Ten of the compounds present predicted IC50 values lower than 10 nM and binding poses and affinity values similar to the natural ligand (leukotriene A4), turning them into suitable candidates for experimental assays.

Identification of a Molecular Signature to Predict the Progression of Kidney Fibrosis Post Transplantation.

Chronic injury in kidney transplants remains a major cause of graft loss. The aim of the study was to identify a predictive gene set capable of classifying renal grafts at risk for progressive injury due to fibrosis. The Genomics of Chronic Allograft Rejection (GoCAR) study is a prospective, multicenter study. Biopsies obtained prospectively from renal allograft recipients (n=159) with stable renal function 3 months after transplantation were analyzed for gene expression by microarray. Genes correlating with 12-month Chronic Allograft Damage Index (CADI) but not 3 month CADI were identified, and analysis performed on 101 recipients with 12-month CADI to determine the predictive capacity of this gene set relative to current clinical and histopathology parameters. We identified a 13 gene set that was independently predictive for the development of chronic allograft damage at 1 year. The high predictive capacity of the gene set (0.947) was superior to clinical indicators (AUC 0.73). While histological parameters failed to discriminate between grafts in which the CADI progressed from non-progressors, the predictive gene set accurately differentiated between these biopsies (AUC=0.987). Testing on an independent cohort within GoCAR (n=45) by qPCR and three independent publically available microarray expression gene sets validated our gene sets ability to predict renal allograft injury. A gene set obtained at 3 months from stably functioning renal allografts correlated with the progression of established markers for chronic allograft damage at 12 months and proved superior to clinico-pathological variables currently used in clinical practice to in identifying renal transplant recipients at risk of graft loss.

OECD validation study to assess intra- and inter-laboratory reproducibility of the zebrafish embryo toxicity test for acute aquatic toxicity testing

The OECD validation study of the zebrafish embryo acute toxicity test (ZFET) for acute aquatic toxicity testing evaluated the ZFET reproducibility by testing 20 chemicals at 5 different concentrations in 3 independent runs in at least 3 laboratories. Stock solutions and test concentrations were analytically confirmed for 11 chemicals. Newly fertilised zebrafish eggs (20/concentration and control) were exposed for 96h to chemicals. Four apical endpoints were recorded daily as indicators of acute lethality: coagulation of the embryo, lack of somite formation, non-detachment of the tail bud from the yolk sac and lack of heartbeat. Results (LC50 values for 48/96h exposure) show that the ZFET is a robust method with a good intra- and inter-laboratory reproducibility (CV<30%) for most chemicals and laboratories. The reproducibility was lower (CV>30%) for some very toxic or volatile chemicals, and chemicals tested close to their limit of solubility. The ZFET is now available as OECD Test Guideline 236.Considering the high predictive capacity of the ZFET demonstrated by Belanger et al. (2013) in their retrospective analysis of acute fish toxicity and fish embryo acute toxicity data, the ZFET is ready to be considered for acute fish toxicity for regulatory purposes.

Accurate and Robust Genomic Prediction of Celiac Disease Using Statistical Learning

Practical application of genomic-based risk stratification to clinical diagnosis is appealing yet performance varies widely depending on the disease and genomic risk score (GRS) method. Celiac disease (CD), a common immune-mediated illness, is strongly genetically determined and requires specific HLA haplotypes. HLA testing can exclude diagnosis but has low specificity, providing little information suitable for clinical risk stratification. Using six European cohorts, we provide a proof-of-concept that statistical learning approaches which simultaneously model all SNPs can generate robust and highly accurate predictive models of CD based on genome-wide SNP profiles. The high predictive capacity replicated both in cross-validation within each cohort (AUC of 0.87–0.89) and in independent replication across cohorts (AUC of 0.86–0.9), despite differences in ethnicity. The models explained 30–35% of disease variance and up to ∼43% of heritability. The GRS's utility was assessed in different clinically relevant settings. Comparable to HLA typing, the GRS can be used to identify individuals without CD with ≥99.6% negative predictive value however, unlike HLA typing, fine-scale stratification of individuals into categories of higher-risk for CD can identify those that would benefit from more invasive and costly definitive testing. The GRS is flexible and its performance can be adapted to the clinical situation by adjusting the threshold cut-off. Despite explaining a minority of disease heritability, our findings indicate a genomic risk score provides clinically relevant information to improve upon current diagnostic pathways for CD and support further studies evaluating the clinical utility of this approach in CD and other complex diseases.

Análise discriminante da maturação puberal de jovens do sexo masculino, a partir das características antropométricas

The relationship between anthropometric variables and maturation stages is important for a more detailed monitoring of pubertal development and may provide a suitable tool for clinical diagnosis. The aim of this study was to analyze the predictive contribution of anthropometric variables to pubertal maturation using multivariate discri-minant analysis. A total of 190 boys aged 8 to 18 years, from public and private schools in Natal, Brazil, participated in the study. Thirty-two anthropometric variables were measured and pubertal maturation was evaluated objectively by observing pubic hair development. Measures of central tendency and dispersion, inferential analysis of variance, and multi-variate discriminant analysis were used for statistical analysis. Pubertal advancement was accompanied by significant changes in anthropometric variables (p&lt;0.05). Discriminant analysis identified eight variables with a high predictive capacity of pubertal maturation: age, sitting height, biacromial breadth, acromiale-radiale, trochanterion-tibiale laterale and tibiale laterale bone lengths, and abdominal and forearm girths. The anthropometric variables showed a high correlation with the classification of pubertal maturation, dem-onstrating a high predictive level among them. These findings indicate the possibility of developing a predictive equation for pubertal stages.

Modelo de adsorción molecular para compuestos orgánicos sobre TiO2 – P25 mediante el método de afinidad protónica

Un modelo predictivo para adsorción molecular de componentes orgánicos sobre partículas de TiO2-P25, fueformulado. El modelo es una extensión de un método teórico de caracterización de superficies denominadoMétodo de distribución de Afinidad Protónica (PDA), el cual se obtiene a partir del análisis de datos de titulaciónpotenciométrica, ZPC del material e isotermas adsorción del tipo Langmuir-Hinshelwood. La naturaleza químicadel componente en el modelo es establecida por el pH inicial y su peso molecular. Se encontró que para el TiO2-P25el ZPC es 6.8. El modelo fue validado con datos experimentales de ácido dicloroacético (DCA), 4-clorofenol (4-CP) y fenol (PH) a diferentes condiciones de operación. Se encontró que el modelo tiene alta capacidad predictiva,estabilidad numérica y robustez en la obtención de las soluciones. Para los datos experimentales disponibles, elcoeficiente de correlación fue de 0.97

From exploitation to conservation: habitat models using whaling data predict distribution patterns and threat exposure of an endangered whale

AbstractAimSufficient data to describe spatial distributions of rare and threatened populations are typically difficult to obtain. For example, there are minimal modern offshore sightings of the endangered southern right whale, limiting our knowledge of foraging grounds and habitat use patterns. Using historical exploitation data of southern right whales (SRW), we aim to better understand their seasonal offshore distribution patterns in relation to broad‐scale oceanography, and to predict their exposure to shipping traffic and response to global climate change.LocationAustralasian region between 130° W and 100° E, and 30° S and 55° S.MethodsWe model 19th century whaling data with boosted regression trees to determine functional responses of whale distribution relative to environmental factors. Habitat suitability maps are generated and we validate these predictions with independent historical and recent sightings. We identify areas of increased risk of ship‐strike by integrating predicted whale distribution maps with shipping traffic patterns. We implement predicted ocean temperatures for the 2090–2100 decade in our models to predict changes in whale distribution due to climate change.ResultsTemperature in the upper 200 m, distance from the subtropical front, mixed layer depth, chlorophyll concentration and distance from ridges are the most consistent and influential predictors of whale distribution. Validation tests of predicted distributions determined generally high predictive capacity. We identify two areas of increased risk of vessel strikes and predict substantial shifts in habitat suitability and availability due to climate change.Main conclusionsOur results represent the first quantitative description of the offshore foraging habitat of SRW. Conservation applications include identifying areas and causes of threats to SRW, generating effective mitigation strategies, and directing population monitoring and research efforts. Our study demonstrates the benefits of incorporating unconventional datasets such as historical exploitation data into species distribution models to inform management and help combat biodiversity loss.

QSAR Models for Prediction of Binding and Inhibitory Properties of [(E)-2- R-vinyl]benzene Derivatives with Therapeutic Effects against Helicobacter pylori

Helicobacter pylori is a gram-negative bacterium that infects the luminal surface of the human gastric epithelium. Around one half of the world’s population is thought to be infected by this bacterium, which is able to develop diseases such as peptic ulcer or gastric cancer. Eradication of Helicobacter pylori is becoming increasingly difficult due to resistance to common antibiotics. In previous work we have shown that an essential protein, flavodoxin, constitutes a target for the development of novel, specific antibiotics against infection caused by this microorganism, and we have described compounds sharing the [(E)-2-R-vinyl]benzene scaffold that exhibit bactericidal properties against Helicobacter pylori cultures. Based on the affinity and activity of 24 such compounds we have now developed QSAR models for affinity and minimal inhibitory concentration that will guide the improvement of antibacterial compounds based in the [(E)-2-R-vinyl]benzene scaffold. The two models show high statistical correlation and predictive capacity. Discovering novel chemicals with specific antimicrobial properties against Helicobacter pylori , and presumably not affected by existing resistances, will be of great help for the treatment of the diseases associated with this bacteriu

Citrullinemia stimulation test in the evaluation of the intestinal function.

Citrullinemia is been reported as a quantitative parameter of the enterocyte mass and function. The objective of this research is to analyse the value of fasting and stimulated citrullinemias in the intestinal function evaluation. A case-control study was undertaken, including 11 patients with short bowel syndrome, 13 patients submitted to malabsorptive bariatric surgery and 11 healthy controls. Plasma levels of amino acids were determined, before and after a stimulation test with oral Lglutamine, by ion exchange chromatography. Citrullinemia was inferior in short bowel patients (28,6 ± 11,3 versus 35,5 ± 11 in operated obese versus 32,2 ± 6,6 μmol/L in controls; n.s.) and lower than 25,5 μmol/L in 54,5% of them (versus 16,7%; p = 0,041; accuracy = 74%; odds ratio = 3, 95%CI 1,2-7,6). ΔCitrullinemia80 (relative variation of citrullinemia at the 80th minute of test) was lower in short bowel patients; its diagnostic accuracy was similar to baseline citrullinemia and also not significant. ΔCitrullinemia80 revealed a high predictive capacity of a short bowel inferior or equal to 50 cm (auR.O.C. = 82,3%; 95%CI 61,7-102,8; p = 0,038). In short bowel syndrome context, citrullinemia stimulation test with oral L-glutamine is feasible and it may improve the predictive capacity of severity. Further investigation is required to determine its clinical relevance and applicability.

Perfil cognitivo de la dependencia emocional en estudiantes universitarios en Medellín, Colombia

El objetivo de este estudio fue la identificación del perfil cognitivo de los estudiantes universitarios con dependencia emocional. Ésta se plantea como un patrón orientado al otro como fuente de satisfacción y seguridad personal, donde el perfil cognitivo les llevaría a interpretar los hechos de manera diferente, basándose en cubrir sus necesidades emocionales insatisfechas. Se tomó una muestra aleatoria estratificada de 569 estudiantes universitarios. Los resultados arrojaron un modelo que explica el 89.3 % de la variabilidad, con alta capacidad predictiva (sensibilidad 91.4 %, especificidad 97.7 %), conformado por el esquema de desconfianza/abuso, creencias centrales del trastorno de la personalidad paranoide y dependiente, distorsión cognitiva de falacia de cambio y estrategia hipodesarrollada de afrontamiento de autonomía.

Random forest algorithm yields accurate quantitative prediction models of benthic light at intertidal sites affected by toxic Lyngbya majuscula blooms

It is shown that targeted high frequency monitoring and modern machine learning methods lead to highly predictive models of benthic light flux. A state-of-the-art machine learning technique was used in conjunction with a high frequency data set to calibrate and test predictive benthic lights models for sites dominated by the toxic cyanobacterium Lyngbya majuscula. Datasets of incident light flux, waters depth, wind speed, wind direction and benthic light flux were collected at two sites, Deception Bay and Eastern Banks, in Moreton Bay, Queensland, that are affected by the toxic marine cyanobacteria L. majuscula. Random forest models were calibrated and validated using the dataset. The models predict benthic photosynthetically active radiation (PAR) based on wind speed, wind direction, water depth and incident light. The two benthic light models produced have high predictive capacity with R2 values of the order 0.8 for both calibration and validation. Wind speed and/or direction influenced the prediction of benthic PAR in the model at each site, suggesting that wind-driven sediment resuspension is an important process in controlling water clarity and hence benthic PAR at both sites; a result consistent with many previous examples from coastal regions. In particular, wind direction had a marked affect on benthic PAR at the partially sheltered Deception Bay site compared to the fully exposed Eastern Banks site. Consistent with other Harmful Algae Bloom (HAB) species, eutrophication has been identified as a common link between sites affected by L. majuscula blooms. However, we show here that at least one major bloom of L. majuscula in Deception Bay coincided with light conditions for optimal L. majuscula photosynthesis, supporting the hypothesis that toxic blooms are a result of a confluence of causes rather than just one. We argue that at eutrophied sites, variation in parameters such as temperature and light rather than nutrients are more likely to cause blooms. This study demonstrates that collection of high frequency variables related to aquatic light flux combined with state of the art machine learning techniques, such as the random forest algorithm, can produce highly accurate models of light flux at shallow coastal sites. These models provide management with the beginnings of an early warning system and scientists with a practical tool for catching the bloom at the incipient stage rather than after the bloom has already developed.

Bagging k-dependence probabilistic networks: An alternative powerful fraud detection tool

Fraud is a global problem that has required more attention due to an accentuated expansion of modern technology and communication. When statistical techniques are used to detect fraud, whether a fraud detection model is accurate enough in order to provide correct classification of the case as a fraudulent or legitimate is a critical factor. In this context, the concept of bootstrap aggregating (bagging) arises. The basic idea is to generate multiple classifiers by obtaining the predicted values from the adjusted models to several replicated datasets and then combining them into a single predictive classification in order to improve the classification accuracy. In this paper, for the first time, we aim to present a pioneer study of the performance of the discrete and continuous k-dependence probabilistic networks within the context of bagging predictors classification. Via a large simulation study and various real datasets, we discovered that the probabilistic networks are a strong modeling option with high predictive capacity and with a high increment using the bagging procedure when compared to traditional techniques.

Molecular Prognostic Prediction for Locally Advanced Nasopharyngeal Carcinoma by Support Vector Machine Integrated Approach

BackgroundAccurate prognostication of locally advanced nasopharyngeal carcinoma (NPC) will benefit patients for tailored therapy. Here, we addressed this issue by developing a mathematical algorithm based on support vector machine (SVM) through integrating the expression levels of multi-biomarkers.Methodology/Principal FindingsNinety-seven locally advanced NPC patients in a randomized controlled trial (RCT), consisting of 48 cases serving as training set and 49 cases as testing set of SVM models, with 5-year follow-up were studied. We designed SVM models by selecting the variables from 38 tissue molecular biomarkers, which represent 6 tumorigenesis signaling pathways, and 3 EBV-related serological biomarkers. We designed 3 SVM models to refine prognosis of NPC with 5-year follow-up. The SVM1 displayed highly predictive sensitivity (sensitivity, specificity were 88.0% and 81.9%, respectively) by integrating the expression of 7 molecular biomarkers. The SVM2 model showed highly predictive specificity (sensitivity, specificity were 84.0% and 94.5%, respectively) by grouping the expression level of 12 molecular biomarkers and 3 EBV-related serological biomarkers. The SVM3 model, constructed by combination SVM1 with SVM2, displayed a high predictive capacity (sensitivity, specificity were 88.0% and 90.3%, respectively). We found that 3 SVM models had strong power in classification of prognosis. Moreover, Cox multivariate regression analysis confirmed these 3 SVM models were all the significant independent prognostic model for overall survival in testing set and overall patients.Conclusions/SignificanceOur SVM prognostic models designed in the RCT displayed strong power in refining patient prognosis for locally advanced NPC, potentially directing future target therapy against the related signaling pathways.

Development of a fuzzy model for predicting ground vibration caused by rock blasting in surface mining

Ground vibration is an integral part of the rock blasting process in surface mines, which may cause severe damages to structures and plants in the nearby environment. Therefore, its prediction plays an important role in the minimization of environmental impacts. The peak particle velocity (PPV) is an important predictor for ground vibration. In this paper, first a fuzzy logic model was developed to predict PPV based on collected data from blasting events in Sarcheshmeh copper mine, located in the southwest of Iran. The predictive fuzzy model was implemented on the fuzzy logic toolbox of MATLAB using the Mamdani algorithm. Then, the PPV was predicted by conventional empirical predictors used in blasting practice and also by multiple regression analysis. Finally, a comparative analysis between the results obtained by the fuzzy model and common vibration predictors was carried out. The results indicated the high predictive capacity of fuzzy model, which can be used as a reliable predictor of ground vibration for the studied mine.

DELIMITAÇÃO DE MICRORREGIÕES AGROCLIMÁTICAS E SUAS RELAÇÕES COM O POTENCIAL PRODUTIVO DA CULTURA DO EUCALIPTO

O objetivo desta pesquisa foi delimitar zonas agroclimáticas homogêneas propícias ao cultivo da cultura do eucalipto. A área estudada abrange do extremo sul da Bahia à região central-serrana do Espírito Santo. A partir de dados de temperatura e pluviosidade de 64 estações meteorológicas, foi calculado o balanço hídrico sequencial decendial, para estimativa da evapotranspiração potencial (ETP), evapotranspiração real (ETR), índice de satisfação de necessidade por água (ISNA = ETR/ETP) e índice de satisfação por necessidade de água com base em análise frequencial de ocorrência de valores superiores a 0,65 (ISNA65). Por espacialização, foram gerados mapas das médias anuais dessas variáveis e dos períodos seco e chuvoso da área de estudo, a partir da seleção dos interpoladores de maior capacidade preditiva, eleitos por validação cruzada. O zoneamento foi elaborado utilizando-se análise de agrupamento, a partir das variáveis ISNA65, ETR do período seco e temperatura média anual. A zona litorânea apresentou as melhores condições agroclimáticas, com os maiores valores médios de ISNA65 (67,8%) e de ETR (391 mm), e a região oeste do Espírito Santo apresentou as piores condições agroclimáticas.Palavras-chave: Zoneamento agroclimático; análise de agrupamento; sistema de informações geográficas; krigagem. AbstractDelimitation of agroclimatic microregions and its relationships with the productive potential of eucalyptus culture. The objective of this research was to delimit agroclimatic favorable zones to the cultivation of eucalyptus. The study area encloses the extreme south of Bahia State to the Espirito Santo`s central-mountainous, Brazil. Using data from 64 weather stations and based on temperature and pluviometric data, water balance was calculated, in order to estimate the variables potential evapotranspiration (ETP), actual evapotranspiration (ETR), crop water requirement index (ISNA = ETR/ETP) and crop water requirement index based on analysis of the frequency of highest values to 0,65 (ISNA65). Spatial maps were generated by the annual averages of these variables for the dry and rainy periods of the study area from the selection of interpolation of higher predictive capacity, elected by cross-validation. The agroclimatic zoning was elaborated using cluster analysis, based on the variables: ISNA65 and ETR of the dry period and annual temperature. The littoral zone wich are the best agroclimatic conditions during the dry period and is characterized by high values of ISNA65 (67,8%) and ETR (391) and the west region of Espirito Santo, showed the worst agroclimatic conditions.Keywords: Agroclimatic zoning; cluster analysis; geographic information system; kriging.