Machine Learning to Early Prediction of Chronic Kidney Disease: Using Imbalanced and Limited Size Data Sets (Preprint)

BACKGROUND Chronic kidney disease (CKD) is a worldwide public health problem, usually diagnosed in the late stages of the disease, increasing public health costs and mortality rates. The late diagnosis is even more critical in low- and middle-income countries due to the high poverty levels, many hard-to-reach locations, and sometimes lack/precarious primary care. Therefore, to alleviate these issues, investment in early prediction is necessary. OBJECTIVE The purpose of this study is to assist the early prediction of CKD, addressing problems related to imbalanced and limited-size data sets. METHODS To address our multi-class problem (low risk, medium risk, high risk, and very high risk), we used data from medical records of 60 Brazilians with or without a diagnosis of CKD, containing the following attributes: hypertension, diabetes mellitus, creatinine, urea, albuminuria, age, gender, and glomerular filtration rate. We used two approaches for oversampling: (1) manual augmentation with data validated by an experienced nephrologist and (2) automated augmentation with the synthetic minority oversampling technique (SMOTE), borderline-SMOTE, and Borderline-SMOTE support vector machine. We implemented classification models based on such data sets and the algorithms: decision tree (DT), random forest, and multi-class AdaBoosted DTs. We also applied the overall local accuracy and local class accuracy methods for dynamic classifier selection; and the k-nearest oracles-union, k-nearest oracles-eliminate, and META-DES for dynamic ensemble selection. We analyzed the models' performances using the hold-out validation, multiple stratified cross-validation (CV), and nested CV. We also computed the importance of features using feature selection methods. RESULTS The best performance was achieved using the DT and multi-class AdaBoosted DTs classification models, oversampled with SMOTE, and validated with the multiple stratified CV and nested CV methods. The DT model presented the highest accuracy score (98.99%) for both multiple stratified CV and nested CV, followed by multi-class AdaBoosted DTs (97.99% and 98.00%), respectively. CONCLUSIONS The SMOTE and multiple stratified CV or nested CV methods provided reliable results for such an imbalanced and limited size data set. During CKD monitoring, based on the DT model, assuming the previous DM evaluation, the user only needs to perform two blood tests: creatinine and urea. Thus, the DT model can assist in designing systems for the early prediction of CKD, providing easy interpretation and cost reduction.

Hiatal Hernia After Esophagectomy: A Single-Center Retrospective Analysis of a Rare Postoperative Phenomenon

Development of a post-esophagectomy hiatal hernia (PEHH) is a rare, but problematic, sequela with the current reported prevalence ranging up to 20%. To determine the incidence rate of PEHH at our institution, a retrospective review of all transhiatal esophagectomies performed from 2012 to 2020 was conducted. Demographic, operative, and oncologic data were collected, rates of PEHH were calculated, and characteristics of subsequent repair were reviewed and analyzed. A total of 160 transhiatal esophagectomies were included, of which four patients (2.5%) developed a PEHH at a mean of 12 months postoperatively (range: 3-28 months) with symptomatology driving the diagnosis for three patients. The limited size of our study does not allow for statistically significant determinations regarding risk factors or method of repair. The true prevalence of a hiatal defect is likely higher than reported, as clinically asymptomatic patients are not captured in our current literature.

A Spider Toxin Exemplifies the Promises and Pitfalls of Cell-Free Protein Production for Venom Biodiscovery

Arthropod venoms offer a promising resource for the discovery of novel bioactive peptides and proteins, but the limited size of most species translates into minuscule venom yields. Bioactivity studies based on traditional fractionation are therefore challenging, so alternative strategies are needed. Cell-free synthesis based on synthetic gene fragments is one of the most promising emerging technologies, theoretically allowing the rapid, laboratory-scale production of specific venom components, but this approach has yet to be applied in venom biodiscovery. Here, we tested the ability of three commercially available cell-free protein expression systems to produce venom components from small arthropods, using U2-sicaritoxin-Sdo1a from the six-eyed sand spider Hexophtalma dolichocephala as a case study. We found that only one of the systems was able to produce an active product in low amounts, as demonstrated by SDS-PAGE, mass spectrometry, and bioactivity screening on murine neuroblasts. We discuss our findings in relation to the promises and limitations of cell-free synthesis for venom biodiscovery programs in smaller invertebrates.

In-plane crystalline anisotropy of bulk β-Ga2O3

The anisotropy of X-ray diffraction scanning of (201) β-Ga2O3 bulk material has been investigated. Symmetric rocking curves (RCs) exhibit distinctly different broadening along different azimuths, with a maximum along [102] and a minimum along a direction rotated by 30° from [010]. Williamson–Hall analysis was applied to study possible factors causing the broadening in these RCs, including instrumental factors, mosaic tilt and coherent scattering. It was found that the RC broadening is determined by both isotropic mosaic tilt and anisotropy in the length over which the crystal structure is not disrupted by limiting factors such as grain boundaries or stacking faults, which we term the `lateral limited size'. In this case, the lateral limited size is governed by {200} stacking faults along the [102] direction and grain boundaries along the [010] direction. The result presents a new anisotropy characteristic of (201) β-Ga2O3.

Gender identity in interaction: overcoming heteronormativity

The present article reviews theoretical concepts that can contribute to the analysis of the construction of gender identity in interaction, moving on from heteronormativity, understood as the normalization of heterosexuality as the only, or more, legitimate form of sexuality. Identity is discussed together with the concepts of face, rapport and (im)politeness from a discursive approach (van der Bom & Mills, 2015). It is argued that gender identity face builds on attributes of both respectability and identity faces with differing strengths and saliency depending on the individuals and the context. Analysis is limited to the construction of hetero and gay male gender identities in interaction with women in academic contexts and draws on data from a corpus of naturally occurring interactions compiled by the author. Gay males seem to differ from hetero males in in their choice of resources for doing face-enhancing positive politeness and rapport with their female colleagues. Despite the limited size of the sample, the study hopes to contribute to a better understanding of the construction of gender identity from a discursive approach.

APROXIMACIÓN AL TURISMO RELACIONADO CON LAS ZONAS DE BAÑO EN EXTREMADURA: UN META-ANÁLISIS A PARTIR DE TRABAJOS ACADÉMICOS INÉDITOS

Más de un centenar de trabajos académicos son presentados anualmente en las facultades españolas con el objetivo de obtener un título de Graduado en Turismo. Muchos de ellos, se mantienen inéditos, a pesar de la utilidad de los datos que recopilan, bien por aspectos formales o por la escasa envergadura que presentan en sí de manera aislada. En este trabajo, por tanto, se han recopilado y analizado los datos de una serie de estudios que han sido llevados a cabo por estudiantes de turismo, y cuyo objetivo era analizar la oferta y la demanda turística de zonas de baño de Extremadura con diferentes características: embalses, gargantas, balnearios y una antigua cantera. Todas ellas comparten fortalezas como la pertenencia a espacios protegidos y la cercanía a numerosos recursos turísticos pero se observan diferentes aprovechamientos en términos de oferta hotelera y de restauración. Respecto a la demanda, predomina el turismo nacional, con vínculos a las localidades cercanas y con una mayor concentración en los fines de semana. More than a hundred academic works are presented annually in Spanish faculties with the aim of obtaining a degree in Tourism. Many of them remain unpublished, despite the usefulness of their data, either because formal aspects or the limited size of their studies. In this work, we have collected and analyzed a sort of academic works in which tourist supply and demand of bathing areas of Extremadura (reservoirs, gorges and an old quarry) have been the core of the study. All of them share strengths such as belonging to protected areas and the proximity to numerous tourist resources but different uses are observed in terms of hotel and restaurant offer. Regarding demand, national tourism linked to nearby towns with a greater concentration on weekends is predominant.

Long non-coding RNA Xist regulates oocyte loss via suppressing miR-23b-3p/miR-29a-3p maturation and upregulating STX17 in perinatal mouse ovaries

The fecundity of female mammals is resolved by the limited size of the primordial follicle (PF) pool formed perinatally. The establishment of PF pool is accompanied by a significant programmed oocyte death. Long non-coding RNAs (lncRNA) are central modulators in regulating cell apoptosis or autophagy in multiple diseases, however, the significance of lncRNAs governing perinatal oocyte loss remains unknown. Here we find that Yin-Yang 1 (YY1) directly binds to the lncRNA X-inactive-specific transcript (Xist) promoter and facilitates Xist expression in the perinatal mouse ovaries. Xist is highly expressed in fetal ovaries and sharply downregulated along with the establishment of PF pool after birth. Gain or loss of function analysis reveals that Xist accelerates oocyte autophagy, mainly through binding to pre-miR-23b or pre-miR-29a in the nucleus and preventing the export of pre-miR-23b/pre-miR-29a to the cytoplasm, thus resulting in decreased mature of miR-23b-3p/miR-29a-3p expression and upregulation miR-23b-3p/miR-29a-3p co-target, STX17, which is essential for timely control of the degree of oocyte death in prenatal mouse ovaries. Overall, these findings identify Xist as a key non-protein factor that can control the biogenesis of miR-23b-3p/miR-29a-3p, and this YY1-Xist-miR-23b-3p/miR-29a-3p-STX17 regulatory axis is responsible for perinatal oocyte loss through autophagy.

Open and Anisotropic Soft Regions in a Model Polymer Glass

The vibrational dynamics of a model polymer glass is studied by Molecular Dynamics simulations. The focus is on the “soft” monomers with high participation to the lower-frequency vibrational modes contributing to the thermodynamic anomalies of glasses. To better evidence their role, the threshold to qualify monomers as soft is made severe, allowing for the use of systems with limited size. A marked tendency of soft monomers to form quasi-local clusters involving up to 15 monomers is evidenced. Each chain contributes to a cluster up to about three monomers and a single cluster involves a monomer belonging to about 2–3 chains. Clusters with monomers belonging to a single chain are rare. The open and tenuous character of the clusters is revealed by their fractal dimension df<2. The inertia tensor of the soft clusters evidences their strong anisotropy in shape and remarkable linear correlation of the two largest eigenvalues. Owing to the limited size of the system, finite-size effects, as well as dependence of the results on the adopted polymer length, cannot be ruled out.