Logistic Principal Component Analysis Research Articles

Comparison of Dietary Patterns and Daily Food Intake Across Kidney Disease Stages in England: An A-Posteriori Cluster Analysis

ObjectiveDietary patterns are rapidly becoming a major focus of medical nutrition therapy in CKD and the analysis of dietary patterns has emerged as a practical approach to evaluate qualitative as well as quantitative aspects of overall diet. In an a-posteriori data-driven approach, dietary patterns are based on the actual food intake of the population evaluated. Investigation of dietary patterns in CKD are not well described, and, to our knowledge, have not been conducted in a UK-based cohort. MethodsAdult participants with a diagnosed kidney condition (CKD 1-5 not requiring dialysis) were recruited into a multi-centre observational cross-sectional study. Dietary intake was assessed using the European Prospective Investigation of Cancer in Norfolk Food Frequency Questionnaire. Logistic Principal Components Analysis was used to identify food group clusters. Differences between groups were assessed using univariate general linear modelling. ResultsIn total, 696 patients were included. The mean age was 64.7 (±14.0) years, 61% of the cohort were male. Most participants were White British (89%). The mean eGFR was 36.6 (±20.9) ml/min/1.732. We found differences in food group intake across stages (e.g., greater intake of nuts and seeds intake in CKD 1-2 versus CKD 4) and across sex (e.g., females had a higher intake of fruit and vegetables versus males). Comparison with the reference cohort revealed that, overall, the CKD cohort had reduced intakes of food stuffs such as cereals and cereal products, but higher intakes of groups such as meat and meat products. There were limited differences in micronutrients, although Vitamin B2 and calcium were higher in earlier stages. ConclusionOverall, the findings from a novel a-posteriori approach underline the complex diversity of food patterns in CKD. The findings from our study may inform dieticians and other healthcare providers about the need to consider treatment modalities and stages when giving dietary recommendations.

Explaining heatwaves with machine learning

AbstractHeatwaves are known to arise from the interplay between large‐scale climate variability, synoptic weather patterns, and regional to local‐scale surface processes. Though recent research has made important progress for each individual contributing factor, ways to properly incorporate multiple or all of them in a unified analysis are still lacking. In this study, we consider a wide range of possible predictor variables from the ERA5 Reanalysis version 5 and ask how much information on heatwave occurrence in Europe can be learned from each of them. To simplify the problem, we first adapt the recently developed logistic principal component analysis to the task of compressing large binary heatwave fields to a small number of interpretable principal components. The relationships between heatwaves and various climate variables can then be learned by a neural network. Starting from the simple notion that the importance of a variable is given by its impact on the performance of our statistical model, we arrive naturally at the definition of Shapley values. Classic results of game theory show that this is the only fair way of distributing the overall success of a model among its inputs. We find a nonlinear model that explains 70% of reduced heatwave variability. The biggest individual contribution (27% of the 70%) comes from upper level geopotential; top‐level soil moisture is in second place (15%). Beyond this decomposition, Shapley interaction values enable us to quantify overlapping information and positive synergies between all pairs of predictors.

Official Statistics and Big Data Processing with Artificial Intelligence: Capacity Indicators for Public Sector Organizations

Efficient monitoring and achievement of the Sustainable Development Goals (SDGs) has increased the need for a variety of data and statistics. The massive increase in data gathering through social networks, traditional business systems, and Internet of Things (IoT)-based sensor devices raises real questions regarding the capacity of national statistical systems (NSS) for utilizing big data sources. Further, in this current era, big data is captured through sensor-based systems in public sector organizations. To gauge the capacity of public sector institutions in this regard, this work provides an indicator to monitor the processing capacity of the public sector organizations within the country (Pakistan). Some of the indicators related to measuring the capacity of the NSS were captured through a census-based survey. At the same time, convex logistic principal component analysis was used to develop scores and relative capacity indicators. The findings show that most organizations hesitate to disseminate data due to concerns about data privacy and that public sector organizations’ IT personnel are unable to deal with big data sources to generate official statistics. Artificial intelligence (AI) techniques can be used to overcome these challenges, such as automating data processing, improving data privacy and security, and enhancing the capabilities of IT human resources. This research helps to design capacity-building initiatives for public sector organizations in weak dimensions, focusing on leveraging AI to enhance the production of quality and reliable statistics.

A Multiobjective Evolutionary Approach for Solving Large-Scale Network Reconstruction Problems via Logistic Principal Component Analysis.

Currently, the problem of uncovering complex network structure and dynamics from time series is prominent in many fields. Despite the recent progress in this area, reconstructing large-scale networks from limited data remains a tough problem. Existing works treat connections of nodes as continuous values, leaving a challenge of setting a proper cut-off value to distinguish whether the connections exist or not. Besides, their performances on large-scale networks are far from satisfactory. Considering the reconstruction error and sparsity as two objectives, this article proposes a subspace learning-based evolutionary multiobjective network reconstruction algorithm, called SLEMO-NR, to solve the aforementioned problems. In the evolutionary process, we assume that binary-coded individuals obey the Bernoulli distribution and can use the probability and natural parameter as alternative representations. Moreover, our approach utilizes the logistic principal component analysis (LPCA) to learn a subspace containing the features of the network structure. The offspring solutions are generated in the learned subspace and then can be mapped back to the original space via LPCA. Benefitting from the alternative representations, a preference-based local search operator (PLSO) is proposed to concentrate on finding solutions approximate to the true sparsity. The experimental results on synthetic networks and six real-world networks demonstrate that, due to the well-learned network structure subspace and the preference-based strategy, our approach is effective in reconstructing large-scale networks compared to six existing methods.

Unruptured Arteriovenous Malformations in the Multidetector Computed Tomography Era: Frequency of Detection and Predictable Failures.

Objectives:While hemorrhage arising from ruptured arteriovenous malformations (AVMs) is usually evident on multidetector non-contrast computed tomography (NCCT), unruptured AVMs can be below the limits of detection. We performed a retrospective review of NCCT of patients with a proven diagnosis of unruptured AVM to determine if advances in CT technology have made them more apparent and what features predict their detection.Material and Methods:Twenty-five NCCTs met inclusion criteria of having angiography or MR proven AVM without hemorrhage, prior surgery, or other CNS disease. Demographic variables, clinical symptoms at presentation, abnormal CT imaging findings, attenuation of the superior sagittal sinus (SSS), and Spetzler-Martin grade of each AVM were recorded. We examined the relationship between AVM detection and SSS attenuation through Kruskal–Wallis test. Exploratory serial logistic principal components analysis was performed including demographics, symptoms, and CT features in the multivariate model.Results:About 80% of the NCCTs showed an abnormality while 20% were normal. All those with an identifiable abnormality showed hyperdensity (80%). Logistic regression models indicate that clustered associations between several CT features, primarily calcifications, hyperdensity, and vascular prominence significantly predicted Spetzler-Martin grade (likelihood ratio 7.7, P = 0.006). SSS attenuation was significantly lower in subjects with occult AVMs when compared to those with CT abnormalities (median 47 vs. 55 HU, P < 0.04).Conclusion:Abnormal hyperdensity was evident in all detectable cases (80%) and multiple CT features were predictive of a higher Spetzler-Martin AVM grade. Moreover, SSS attenuation less than 50 HU was significantly correlated with a false-negative NCCT.

Survival Analysis of COVID-19 Patients With Symptoms Information by Machine Learning Algorithms

In this study, a survival analysis of the time to death caused by coronavirus disease 2019 is presented. The analysis of a dataset from the East Asian region with a focus on data from the Philippines revealed that the hazard of time to death was associated with the symptoms and background variables of patients. Machine learning algorithms, i.e., dimensionality reduction and boosting, were used along with conventional Cox regression. Machine learning algorithms solved the diverging problem observed when using traditional Cox regression and improved performance by maximizing the concordance index (C-index). Logistic principal component analysis for dimensionality reduction was significantly efficient in addressing the collinearity problem. In addition, to address the nonlinear pattern, a higher C-index was achieved using extreme gradient boosting (XGBoost). The results of the analysis showed that the symptoms were statistically significant for the hazard rate. Among the symptoms, respiratory and pneumonia symptoms resulted in the highest hazard level, which can help in the preliminary identification of high-risk patients. Among various background variables, the influence of age, chronic disease, and their interaction were identified as significant. The use of XGBoost revealed that the hazards were minimized during middle age and increased for younger and older people without any chronic diseases, with only the elderly having a higher risk of chronic disease. These results imply that patients with respiratory and pneumonia symptoms or older patients should be given medical attention.

Dimensionality reduction for binary data through the projection of natural parameters

Principal component analysis (PCA) for binary data, known as logistic PCA, has become a popular alternative to dimensionality reduction of binary data. It is motivated as an extension of ordinary PCA by means of a matrix factorization, akin to the singular value decomposition, that maximizes the Bernoulli log-likelihood. We propose a new formulation of logistic PCA which extends Pearson’s formulation of a low dimensional data representation with minimum error to binary data. Our formulation does not require a matrix factorization, as previous methods do, but instead looks for projections of the natural parameters from the saturated model. Due to this difference, the number of parameters does not grow with the number of observations and the principal component scores on new data can be computed with simple matrix multiplication. We derive explicit solutions for data matrices of special structure and provide a computationally efficient algorithm for solving for the principal component loadings. Through simulation experiments and an analysis of medical diagnoses data, we compare our formulation of logistic PCA to the previous formulation as well as ordinary PCA to demonstrate its benefits.

Logistic principal component analysis via non-convex singular value thresholding

Multivariate binary data is becoming abundant in current biological research. Logistic principal component analysis (PCA) is one of the commonly used tools to explore the relationships inside a multivariate binary data set by exploiting the underlying low rank structure. We re-expressed the logistic PCA model based on the latent variable interpretation of the generalized linear model on binary data. The multivariate binary data set is assumed to be the sign observation of an unobserved quantitative data set, on which a low rank structure is assumed to exist. However, the standard logistic PCA model (using exact low rank constraint) is prone to overfitting, which could lead to the divergence of some estimated parameters towards infinity. We propose to fit a logistic PCA model through non-convex singular value thresholding to alleviate the overfitting issue. An efficient Majorization-Minimization algorithm is implemented to fit the model and a missing value based cross validation (CV) procedure is introduced for the model selection. Our experiments on realistic simulations of imbalanced binary data and low signal-to-noise ratio show that the CV error based model selection procedure is successful in selecting the proposed model. Furthermore, the selected model demonstrates superior performance in recovering the underlying low rank structure compared to models with convex nuclear norm penalty and exact low rank constraint. Finally, a binary copy number aberration data set is used to illustrate the proposed methodology in practice.

A multivariate statistical approach for the estimation of the ethnic origin of unknown genetic profiles in forensic genetics

DNA typing and genetic profile data interpretation are among the most relevant topics in forensic science; among other applications, genetic profile’s capability to distinguish biogeographic information about population groups, subgroups and affiliations have been largely explored in the last decade. In fact, for investigative and intelligence purposes, it is extremely useful to identify subjects and estimate their biogeographic origins by examining the recovered DNA profiles from evidence on a crime scene. Current approaches for BiogeoGraphic Ancestry (BGA) estimation using STRs profiles are usually based on Bayesian methods, which quantify the evidence in terms of likelihood ratio, supporting or not the hypothesis that a certain profile belongs to a specific ethnic group. The present study provides an alternative approach to the likelihood ratio method that involves multivariate data analysis strategies for the estimation of multiple populations. Starting from the well-known NIST US autosomal STRs dataset involving African-American, Asian, and Caucasian individuals, and moving towards further and more geographically restricted populations (such as Northern Africans vs sub-Saharan Africans, Afghans vs Iraqis and Italians vs Romanians), powerful multivariate techniques such as Sparse and Logistic Principal Component Analysis (SL-PCA), Sparse Partial Least Squares-Discriminant Analysis (sPLS-DA) and Support Vector Machines (SVM) were employed and their discriminating power was also compared. Both sPLS-DA and SVM techniques provided robust classifications, yielding high sensitivity and specificity models capable of discriminating populations on ethnic basis. This application may represent a powerful and dynamic tool for law enforcement agencies whenever a standard autosomal STR profile is obtained from the biological evidence collected at a crime scene or recovered during mass-disaster and missing person investigations.

Sustainability and environmental management in the Mexican vegetable sector

In terms of sustainability, the Mexican horticultural sector faces great challenges as well as opportunities because of its strong environmental and social impact. The main environmental problems in Mexico are related to inefficient water use and to deforestation. Of growing importance is soil pollution because of the inappropriate use of agrochemicals, greenhouse gas emissions and soil erosion. Environmental problems are more severe in arid or semiarid regions, such as the state of Zacatecas, where 87.7% of the cultivated area is rainfed, while the rest is irrigated with water extracted from 34 aquifers, 14 of which are over-exploited. This study determines the relationship between the decision to adopt an integral program of care and protection of the environment in vegetable production units (PUs) and the characteristics of the PUs and growers, their views on sustainability, their agricultural practices and the drivers and barriers to adoption. Moreover, the likelihood of adoption of this program was determined. Information was obtained through a questionnaire applied to owners of PUs and was processed using two logistic regression models and principal component analysis. The likelihood of adopting the program was 74.2%, determined by level of education, view on sustainability, profitability of the PU, internal aspects of the organization and sustainable agricultural practices. For the design of an integral program of care and protection of the environment in the horticulture sector, we suggest direct support to the growers to establish sustainable agricultural practices, integrating standards of food safety, biodiversity and environmental management.

Dimension reduction and surrogate based topology optimization of periodic structures

The frequency bandgap of periodic structures has many potential applications. Topology optimization of the unit cell of periodic structures offers great potential to design periodic structures with desirable bandgap characteristics. However, topology optimization typically involves many design variables stemming from discretization of the unit cell. This creates computational challenges for both optimization (i.e., high-dimensional discrete design variables) and the calculation of frequency bandgaps for a given design (i.e., finer discretization requires higher computational effort). To address these challenges, this paper proposes an efficient dimension reduction and surrogate based approach for topology optimization of periodic structures. Using information from a set of reference topologies with more desirable bandgap characteristics, dimension reduction technique (i.e., logistic principal component analysis) is used to establish a low-dimensional representation of different topologies in latent design space. To reduce computational effort in calculation of bandgap, Kriging surrogate model is built with respect to the low-dimensional latent continuous design variables, and used within efficient global optimization to efficiently and adaptively identify the optimal topology. The effectiveness and great efficiency of the proposed approach are verified through an example on topology optimization of 2D periodic structures to maximize the in-plane frequency bandgaps.

Evaluation of methods for adjusting population stratification in genome-wide association studies: Standard versus categorical principal component analysis.

Unaccounted population stratification can lead to false-positive findings and can mask the true association signals in identification of disease-related genetic variants. The computational simplicity of principal component analysis (PCA) makes it a widely used method for population stratification adjustment. However, given that genotype data are generally represented by numerical values 0, 1, and 2, corresponding to the number of minor alleles, it is more reasonable to consider genotype data as categorical data. Because PCA is inherently only suitable for continuous variables, it is not appropriate to directly apply PCA on genotype data. Second, although common variants have been extensively studied, little is known about the stratification of rare variants and its impact on association tests. Over the last decade, there has been a shift in the genome-wide association studies toward studying low-frequency (minor allele frequency [MAF] between 0.01 and 0.05) and rare (MAF less than 0.01) variants, which are now widely reputed as complex trait determinants. The fact that rare variants are not stratified in the same way as common variants necessitates the development of statistical methods that can capture stratification patterns for low-frequency and rare variants. To address these limitations, we investigate performances of generalized PCA and similarity-matrix-based PCA methods to detect underlying structures for rare and common variants. We demonstrate, through simulated and real datasets, that a special case of generalized PCA (i.e., logistic PCA) is able to adjust for population stratification in rare variants much more effectively than standard PCA while their performances are comparable for common variants.

Factors driving university choice: a principal component analysis on Italian institutions

ABSTRACT When investigating students’ motivations to enrol in university, a wide range of elements related to the overall student experience should be taken into account. The current study moves from this point to analyse students’ choice factors from a survey completed by 27,504 students across 23 Italian institutions by means of a logistic principal component analysis. Results confirm the presence of multiple factors jointly influencing students’ choice, with geographical proximity, job opportunities in the region, university reputation and ease of access opposing one another. Aggregating results at the institutional level, students’ distribution proves to be highly heterogeneous across universities, which are selected for different contextual factors even within the same region. From this, a managerial tool is provided to position students’ population and derive strategic implications. This increases universities’ awareness and enables them to better focus on the main served population or to target a different one. Finally, policy considerations are reported.

The Charlson Index Is Insufficient to Control for Comorbidities in a National Trauma Registry

BackgroundThe Charlson Comorbidity Index (CCI) is frequently used to control for confounding by comorbidities in observational studies, but its performance as such has not been studied. We evaluated the performance of CCI and an alternative summary method, logistic principal component analysis (LPCA), to adjust for comorbidities, using as an example the association between insurance and mortality. Materials and methodsUsing all admissions in the National Trauma Data Bank 2010-2015, we extracted mortality, payment method, and 36 International Classification of Disease, Ninth Revision–derived comorbidities. We estimated odds ratios (ORs) for the association between uninsured status and mortality before and after adjusting for CCI, LPCA, and separate covariates. We also calculated standardized mean differences (SMDs) of comorbidity variables before and after weighting the sample using inverse probability of treatment weights for CCI, LPCA, and separate covariates. ResultsIn 4,936,880 admissions, most (68.3%) had at least one comorbidity. Considerable imbalance was observed in the unweighted sample (mean SMD = 0.086, OR = 1.17), which was almost entirely eliminated by inverse probability of treatment weights on separate covariates (mean SMD = 0.012, OR = 1.36). The CCI performed similarly to the unweighted sample (mean SMD = 0.080, OR = 1.25), whereas two LPCA axes were better able to control for confounding (mean SMD = 0.04, OR = 1.31). Using covariate adjustment, the CCI accounted for 56.1% of observed confounding, whereas two LPCA axes accounted for 91.3%. ConclusionsThe use of the CCI to adjust for confounding may result in residual confounding, and alternative strategies should be considered. LPCA may be a viable alternative to adjusting for each comorbidity when samples are small or positivity assumptions are violated.

Dimensionality Reduction to Reveal Urban Truck Driver Activity Patterns

This paper studies the activity profiles of truck drivers in urban areas. Finding repeating dynamical patterns is important in understanding freight behaviors, and aids freight-friendly planning. In the digital age, data on truck drivers is becoming more available with heterogeneous demographic and work profiles. By synthesizing such pervasive data and applying machine learning concepts, this paper proposes to identify signature travel activity patterns via dimensionality reduction. Based on driver survey data, truck drivers’ behaviors are represented as longitudinal activity sequences. Dimensionality reduction and activity reconstruction via principal components analysis (PCA), logistic PCA, and autoencoder were conducted to reveal fundamental activity features and approximate the underlying data-generating function. In the driver survey dataset, 243 truck drivers in Singapore reported their daily activities for 1,099 weekdays. This study found that PCA produced the most faithful reconstruction of drivers’ activities. When projecting the input data down from 2,592 to 82 dimensions, PCA explained 77% of variances with a reconstruction error of 0.99%. Logistic PCA is a useful extension of PCA to study the pattern of a single activity. It captures the variation of infrequent activities such as truck queuing, which PCA fails to reconstruct. Autoencoder was found to be more powerful than PCA in reconstructing activities – with 1% of original dimensions, it reconstructed the activities with an error rate of 1.24%. Moreover, when implemented as a variational autoencoder, autoencoder generated realistic-looking samples of driver activities. The top three most distinctive activity patterns of Singapore truck drivers are reported using PCA.

Principal Component Analysis to Evaluate Relationships in Reactivity to Commercial Extracts and Uncharacterized Fungi among Asthmatic and Allergic Rhinitis Subjects

Abstract Rationale Identifying patterns in reactivity to allergens can uncover shared respiratory health risks among subjects suffering from chronic respiratory diseases. Therefore, there is a need to identify statistical approaches that can expose inter-allergen relationships. Methods : We implemented spearman, exponential, and logistic principal component analysis on specific immunoglobulin-E (sIgE) levels, reactivity to commercial extracts, basidiomycetes (Ganoderma sp, Chlorophyllum sp, Pleurotus sp) crude extracts, and to previously uncharacterized fungal airborne particulate among asthmatics, allergic and non-allergic rhinitis subjects. Results : Positive correlations (p = 0.03 to p &amp;lt; 0.001) were identified between reactivity of two commercial fungal extracts (Aspergillus sp, Alternaria sp), and among all three basidiomycete crude extracts (Ganoderma sp., Chlorophyllum, sp, Pleurotus sp.; p = 0.01 to p &amp;lt; 0.001). sIgE levels correlated with all five categories of airborne fungal particulates (un-identified particulates, ascospores, basidiospores, fungal fragments, and mitosporic fungi), but only with three extracts (Aspergillus sp., Alternaria sp, and Chlorophyllum sp). Based on reactivity to fungal airborne particulate, we identified three groups: subjects reactive to ascospores, reactive to basidiospores and unidentified particulate, and reactive to mitosporic fungi and fungal fragments. Conclusions : Our finding suggests that PCA of sIgE, together with reactivity to commercial and previously uncharacterized allergens present a suitable strategy to identify respiratory health risks among individuals suffering from chronic respiratory diseases, such as asthma and allergic rhinitis.

Principal component analysis of binary genomics data.

Genome-wide measurements of genetic and epigenetic alterations are generating more and more high-dimensional binary data. The special mathematical characteristics of binary data make the direct use of the classical principal component analysis (PCA) model to explore low-dimensional structures less obvious. Although there are several PCA alternatives for binary data in the psychometric, data analysis and machine learning literature, they are not well known to the bioinformatics community. Results: In this article, we introduce the motivation and rationale of some parametric and nonparametric versions of PCA specifically geared for binary data. Using both realistic simulations of binary data as well as mutation, CNA and methylation data of the Genomic Determinants of Sensitivity in Cancer 1000 (GDSC1000), the methods were explored for their performance with respect to finding the correct number of components, overfit, finding back the correct low-dimensional structure, variable importance, etc. The results show that if a low-dimensional structure exists in the data, that most of the methods can find it. When assuming a probabilistic generating process is underlying the data, we recommend to use the parametric logistic PCA model, while when such an assumption is not valid and the data are considered as given, the nonparametric Gifi model is recommended. The codes to reproduce the results in this article are available at the homepage of the Biosystems Data Analysis group (www.bdagroup.nl).

PATIENTS’ ASSESSMENT OF IN-PATIENT MEDICAL CARE QUALITY IN KAZAKHSTAN

In this cross-sectional study we assessed self-perceived quality of inpatient healthcare in three Kazakhstani cities. Altogether, 923 patients admitted to the city hospitals in Almaty, Aktobe and Semey filled out an anonymous questionnaire. 63 % of them were emergency patients. Binary logistic regression and principal component analysis (PCA) were used. Adjustments for social and demographic patients' characteristics were made. Altogether, 82.9 % (95 % CI 80.3-85.2) of patients rated the quality of healthcare as good or excellent. Among emergency patients, the key factors for being unsatisfied by the quality of services were insufficient quality of hospital food (OR 4.2; 95 % CI 1.70, 10.4), poor communication between doctors and nurses (OR 6.3; 95 % CI 2.3; 17.3) and the lack of adequate explanations regarding procedures and medication (OR 3.8; 95 % CI 1.5, 9.6). Among non-emergency patients, the key factors for being unsatisfied by the quality of services were insufficient quality of hospital food (OR 2.3; 95 % CI 1.2, 4.6), poor communication between doctors and nurses (OR 6.3; 95 % CI 2.9, 13.5), situation if the physician did not pay attention to whether a patient understood information about his/her condition or treatment (OR 2.4; 95 % CI 1.0, 5.2), when the patient has not had the opportunity to discuss his/her condition with medical staff (OR 2.7; 95 % CI 1.2, 5.8), if the patient's relatives was not given a possibility to communicate with a doctor (OR 4.3; 95 % CI 1.7, 11.0), or such this possibility was limited (OR 4.8; 95 % CI 1.9; 12.2).

Possible factors affecting internet addiction: a case study of higher education students of Qatar

The sea change brought by the internet and technology to human life came along with all possible hazards, if not adopted in the correct manner. In this regard, the current study attempted an effort to capture the opinion of the students about the cause and effect of the factors in internet usage and specifically to internet addiction. Data is collected from 357 students of Qatar using questionnaire. The collected data is validated and analysed using binary logistic regression and principal component analysis. Though the construct 'culture' is introduced in the model, no significant contribution and impact is found on the internet usage and addiction. However, the constructs introduced are found to have association with internet addiction but in logistic regression analysis, the education level and time management have actual impact on the internet addiction. But unlike the above results, the principal component analysis showed that demographic variables like age, gender, income level and others are able to explain much variation in internet usage.

Attributes of basking sites of the European pond turtle (Emys orbicularis) in central Italy

Attributes of basking sites are important elements to study in management plans of threatened freshwater turtles. Here, we analyzed the basking-site characteristics of European pond turtle (Emys orbicularis) populations in a Mediterranean territory of central Italy (Tolfa Mountains, Latium). We used logistic regression and Principal Components Analysis to characterize 29 presence sites versus 61 random sites, through 16 descriptive variables recorded within a 5 m radius from the sighting/target spot. Our analyses revealed that some variables (i.e. water turbidity, presence of small coves, submerged vegetation, and emergent tree-trunks) were those that influenced most strongly the presence of turtles on potential basking sites. Maintenance of deadwood in water and preservation of submerged aquatic vegetation should be included in the management planning for this turtle species in central Italy.