Low Number Of Samples Research Articles

Meth-SemiCancer: a cancer subtype classification framework via semi-supervised learning utilizing DNA methylation profiles

BackgroundIdentification of the cancer subtype plays a crucial role to provide an accurate diagnosis and proper treatment to improve the clinical outcomes of patients. Recent studies have shown that DNA methylation is one of the key factors for tumorigenesis and tumor growth, where the DNA methylation signatures have the potential to be utilized as cancer subtype-specific markers. However, due to the high dimensionality and the low number of DNA methylome cancer samples with the subtype information, still, to date, a cancer subtype classification method utilizing DNA methylome datasets has not been proposed.ResultsIn this paper, we present meth-SemiCancer, a semi-supervised cancer subtype classification framework based on DNA methylation profiles. The proposed model was first pre-trained based on the methylation datasets with the cancer subtype labels. After that, meth-SemiCancer generated the pseudo-subtypes for the cancer datasets without subtype information based on the model’s prediction. Finally, fine-tuning was performed utilizing both the labeled and unlabeled datasets.ConclusionsFrom the performance comparison with the standard machine learning-based classifiers, meth-SemiCancer achieved the highest average F1-score and Matthews correlation coefficient, outperforming other methods. Fine-tuning the model with the unlabeled patient samples by providing the proper pseudo-subtypes, encouraged meth-SemiCancer to generalize better than the supervised neural network-based subtype classification method. meth-SemiCancer is publicly available at https://github.com/cbi-bioinfo/meth-SemiCancer.

ANKLE LIGAMENT REPAIR FOR RESTORATION OF FUNCTION: EXPLORING THE APPLICATION OF DECELLULARIZED GRAFTS

Ligament integrity is directly associated with ankle stability. Nearly 40% of ankle sprains result in chronic ankle instability, affecting biomechanics and potentially causing osteoarthritis. Ligament replacement could restore stability and avoid this degenerative pathway, but a greater understanding of ankle ligament behaviour is required. Additionally, autograft or allograft use is limited by donor-site morbidity and inflammatory responses respectively. Decellularised porcine grafts could address this, by removing cellular material to prevent acute immune responses, while preserving mechanical properties.This project will characterise commonly injured ankle ligaments and damage mechanisms, identify ligament reconstruction requirements, and investigate the potential of decellularised porcine grafts as a replacement material.Several porcine tendons were evaluated to identify suitable candidates for decellularisation. The viscoelastic properties of native tissues were assessed using dynamic mechanical analysis (DMA), followed by ramp to ‘sub-rupture’ at 1% strain/s, and further DMA. Multiple samples (n=5) were taken along the graft to assess variation along the tendon.When identifying suitable porcine tendons, a lack of literature on human ankle ligaments was identified. Inconsistencies in measurement methods and properties reported makes comparison between studies difficult.Preliminary testing on porcine tendons suggested there is little variation in viscoelastic properties along the length of tendon. Testing also suggested strain rates of 1%/s sub-rupture was not large enough to affect viscoelastic properties (no changes in storage or loss moduli or tanẟ). Further testing is underway to improve upon low initial sample numbers and confirm these results, with varying strain rates to identify suitable sub-rupture sprain conditions.This work highlights need for new data on human ankle ligaments to address knowledge gaps and identify suitable replacement materials. Future work will generate this data and decellularise porcine tendons of similar dimensions. Collagen damage will be investigated using histology and lightsheet microscopy, and viscoelastic changes through DMA.

Abstract 2056: Large-scale whole genome sequencing of hepatocellular carcinoma reveals stepwise cancer evolution

Abstract Liver cancer is a leading cause of death worldwide where somatic mutation plays a key role in tumorigenesis. As such, many recent researches have been conducted using DNA sequencing methods to study somatic mutations in liver cancer, such as whole-exome sequencing (WES), panel sequencing, and whole-genome sequencing (WGS). However, our understanding of genomic landscape of liver tumorigenesis has been limited by a low number of quality samples from diverse clinical contexts. Here, we explore comprehensive genomic landscape of liver cancer throughout its tumorigenesis using large-scale whole-genome sequencing of hepatocellular carcinoma (HCC) in various cancer stages and liver conditions. Total 500 patients of HCCs were enrolled in this study and whole-genome sequenced. 462 samples that passed quality control were included for final analysis. Most of the samples (427; 92%) were obtained from surgical resection and the rest (35; 8%) from liver transplantation. We used a standard WGS analysis pipeline from read mapping to variant calling of single-base substitution (SBS), indel, structural variation (SV), and copy number variation (CNV). Downstream data analyses were performed using in-house scripts with a particular focus on mutational signatures. To study how genome changes as HCC progresses, we classified samples into three groups by T stage: T1, T2, and T3/4. The number of samples for each group was 205 (44%), 222 (48%), and 35 (8%), respectively. There was no significant difference in the number of SBSs and indels, but the number of SVs was significantly higher in T3/4 (the mean of T1/2 vs. T3/4 = 51 vs. 93; p=0.001). Major mutational signatures were SBS5, 40 (both clock-like), 12 (unknown), and 22 (aristolochic acid) for SBSs and ID1, 2, 5 (all clock-like), and 3 (tobacco smoking) for indels. The proportion of ID2, characterized by 1bp deletion at homopolymer site, was increased in the samples of higher T stage (the mean proportion of T1/2 vs. T3/4=11% vs. 16%; p=0.01). The higher-stage samples also harbored more frequent whole-genome duplication (WGD) and loss of heterozygosity (LOH) (the frequency of WGD in T1/2 vs. T3/4=42% vs. 68%, p<0.001; the mean LOH fraction of T1/2 vs. T3/4=0.12 vs. 0.19, p<0.001). There was a significant enrichment of driver mutations of multiple genes (e.g., TP53, MTCP1, CCND3, etc.) in T3/4 group. Interestingly, in surgically resected T1 samples, ID2 proportion was the only significant contributor to early recurrence over tumor size, N stage, mean ploidy, and LOH fraction (n=199, p=0.05; Cox regression). Overall, we found that HCCs at different T stages exhibit some common and yet apparently distinguishing genomic features. We believe that the comprehensive genomic profile of HCC across cancer stages will provide deep insights into the genomic evolution of HCC and one day lend itself to clinicogenomic bases for better HCC management in the future. Citation Format: Seongyeol Park, Suk Kyun Hong, Joonoh Lim, YoungRok Choi, Boram Yi, Jaemo Koo, June-Young Koh, Ryul Kim, Nam-Joon Yi, Jeong Seok Lee, Kyung-Suk Suh, Kwang-Woong Lee, Young Seok Ju. Large-scale whole genome sequencing of hepatocellular carcinoma reveals stepwise cancer evolution [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2056.

Overexpression of CLDN16 in ovarian cancer is modulated by PI3K and PKC pathways

Epithelial ovarian cancer (EOC) is the gynecological malignant tumor of poorest prognosis and higher mortality rate. Chemotherapy is the base of high-grade serous ovarian cancer (HGSOC) treatment; however, it favors the emergence of chemoresistance and metastasis. Thus, there is an urge to search for new therapeutic targets, such as proteins related to cellular proliferation and invasion. Herein, we investigated the expression profile of claudin-16 (CLDN16 protein and CLDN16 transcript) and its possible functions in EOC. In silico analysis of CLDN16 expression profile was performed using data extracted from GENT2 and GEPIA2 platforms. A retrospective study was carried out with 55 patients to evaluate the expression of CLDN16. The samples were evaluated by immunohistochemistry, immunofluorescence, qRT-PCR, molecular docking, sequencing, and immunoblotting assays. Statistical analyzes were performed using Kaplan-Meier curves, one-way ANOVA, Turkey posttest. Data were analyzed using GraphPad Prism 8.0. In silico experiments showed that CLDN16 is overexpressed in EOC. 80.0% of all EOC types overexpressed CLDN16, of which in 87% of the cases the protein is restricted to cellular cytoplasm. CLDN16 expression was not related to tumor stage, tumor cells differentiation status, tumor responsiveness to cisplatin, or patients' survival rate. When compared to data obtained from in silico analysis regarding EOC stage and degree of differentiation, differences were found in the former but not in the later, neither in survival curves. CLDN16 expression in HGSOC OVCAR-3 cells increased by 1.95-fold (p < 0.001), 2.32-fold (p < 0.001), and 6.57-fold (p < 0.001) via PKC, PI3K, and estrogen pathways, respectively. Altogether, our results suggest that despite the low number of samples included in our in vitro studies, adding to the expression profile findings, we provided a comprehensive study of CLDN16 expression in EOC. Therefore, we hypothesize that CLDN16 is a potential target in the diagnosis and treatment of the disease.

Challenges of Integrating APOL1 Genetic Test Results into the Electronic Health Record.

Integrating genetic test results into the electronic health record (EHR) is essential for integrating genetic testing into clinical practice. This article describes the organizational challenges of integrating discrete apolipoprotein L1 (APOL1) genetic test results into the EHR for a research study on culturally sensitive genetic counseling for living kidney donors. We convened a multidisciplinary team across three institutions (Northwestern University, Northwestern Memorial HealthCare [NMHC], and OHSU Knight Diagnostic Laboratories [KDL]), including researchers, physicians, clinical information technology, and project management. Through a series of meetings over a year between the team and the genetic testing laboratory, we explored and adjusted our EHR integration plan based on regulatory and budgetary constraints. Our original proposal was to transmit results from KDL to NMHC as structured data sent via Health Level Seven (HL7) v2 message. This was ultimately deemed infeasible given the time and resources required to establish the interface, and the low number of samples to be processed for the study (n = 316). We next explored the use of Epic's Care Everywhere interoperability platform, but learned it was not possible as a laboratory test ordered for a research study; even though our intent was to study the APOL1 genetic test result's clinical use and impact, test results were still considered "research results." Faced with two remaining options-downloading a PDF from the KDL laboratory portal or scanning a faxed result from KDL-only a PDF of the APOL1 test result could be integrated into the EHR, reinforcing the status quo. Even with early and ongoing stakeholder engagement, dedicated project management, and funding, unanticipated implementation challenges-especially for research projects-can result in drastic design tradeoffs.

SARS-CoV-2 in outdoor air following the third wave lockdown release, Portugal, 2021.

Aiming to contribute with more data on the presence of SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) in outdoor environments, we performed air sampling in outdoor terraces from restaurants in three major cities of Portugal in April 2021, following the third wave lockdown release in the country. Air samples (n=19) were collected in 19 restaurant terraces during lunch time. Each air sample was collected using a Coriolis Compact air sampler, followed by RNA extraction and real-time quantitative PCR for the detection of viral RNA. Viral viability was also assessed through RNAse pre-treatment of samples. Only one of the 19 air samples was positive for SARS-CoV-2 RNA, with 7337 gene copies m-3 for the genomic region N2, with no viable virus in this sample. The low number of positive samples found in this study is not surprising, as sampling took place in outdoor settings where air circulation is optimal, and aerosols are rapidly dispersed by the air currents. These results are consistent with previous reports stating that transmission of SARS-CoV-2 in outdoor spaces is low, although current evidence shows an association of exposures in settings where drinking and eating is possible on-site with an increased risk in acquiring SARS-CoV-2 infection. Moreover, the minimal infectious dose for SARS-CoV-2 still needs to be determined so that the real risk of infection in different environments can be accurately established.

A Study of Milk Composition and Coagulation Properties of Holstein-Friesian, Jersey, and Their Cross Milked Once or Twice a Day

The objective of the study was to explore the effect of breed on the composition and coagulation properties (rennet coagulation time (min), curd firming rate (min), and curd firmness (mm)) of milk from cows milked once a day or twice a day in the morning and afternoon, using a Formagraph. Thirty cows (11 Holstein-Friesian, 8 Holstein-Friesian × Jersey cross, and 11 Jersey) from a once-a-day milking herd and thirty cows (16 Holstein-Friesian, 10 Holstein-Friesian × Jersey cross, and 4 Jersey) from a twice-a-day milking herd were sampled in late lactation. The milk composition and coagulation properties were analysed for each milk sample. Jersey cows had better milk coagulation properties at each milking frequency-milking time compared to Holstein-Friesian cows. Curd firmness 30 min after the addition of rennet was positively (p &lt; 0.05) correlated with the protein concentration. However, the correlations were inconsistent between milking frequencies and milking times, resulting in poor prediction of the changes in cheese-making potential. This study indicated that milk composition and coagulation properties were affected by breed and milking frequency. The effect of the breed could be due to the variation in the composition of the milk, but firm recommendations were hampered by a low number of samples. Further research with larger cow numbers is justified.

Diagnostic Methods in Forensic Pathology: A New Sign in Death from Hanging.

To evaluate the usefulness of studying vital injuries at the sternal head insertion of the sternocleidomastoid muscle in the medico-legal assessment of death by hanging. Study material was obtained from eight bodies of people who died from hanging. The control group included as many specimens collected from people who died from traumatic causes other than hanging (precipitation from medium to large heights and traffic accidents). The structures under study were examined histologically with a BX-51 light microscope (Olympus). An analysis of the extravasated erythrocytes was performed by counting the number per mm2 in the histologic section on 10 HPF (400×), and Student's t-test for a comparison of the averages was applied for all parametric values. The authors noted that the key finding, indicative of the subject's viability at the time of discontinuation, was the presence of recent hemorrhagic infiltrate (in the absence of hemosiderin) at the tendon insertion of the sternocleidomastoid muscle and the proximal part of the muscle itself. All specimens tested were positive for the presence of hemorrhagic infiltrate at the portions tested in a statistically significant manner. In contrast, in the control cases there was no or, where present, no statistically significant (p < 0.05) presence of recent hemorrhagic infiltrate. The limitation of the study is the low number of samples examined. In any case, the results obtained are strongly indicative of the possibility of using this type of forensic pathological investigation in cases where there is a doubt in terms of a differential diagnosis between hanging (suicidal type) and suspension of a corpse in a simulation of hanging.

Gdaphen, R pipeline to identify the most important qualitative and quantitative predictor variables from phenotypic data

BackgroundIn individuals or animals suffering from genetic or acquired diseases, it is important to identify which clinical or phenotypic variables can be used to discriminate between disease and non-disease states, the response to treatments or sexual dimorphism. However, the data often suffers from low number of samples, high number of variables or unbalanced experimental designs. Moreover, several parameters can be recorded in the same test. Thus, correlations should be assessed, and a more complex statistical framework is necessary for the analysis. Packages already exist that provide analysis tools, but they are not found together, rendering the decision method and implementation difficult for non-statisticians.ResultWe present Gdaphen, a fast joint-pipeline allowing the identification of most important qualitative and quantitative predictor variables to discriminate between genotypes, treatments, or sex. Gdaphen takes as input behavioral/clinical data and uses a Multiple Factor Analysis (MFA) to deal with groups of variables recorded from the same individuals or anonymize genotype-based recordings. Gdaphen uses as optimized input the non-correlated variables with 30% correlation or higher on the MFA-Principal Component Analysis (PCA), increasing the discriminative power and the classifier’s predictive model efficiency. Gdaphen can determine the strongest variables that predict gene dosage effects thanks to the General Linear Model (GLM)-based classifiers or determine the most discriminative not linear distributed variables thanks to Random Forest (RF) implementation. Moreover, Gdaphen provides the efficacy of each classifier and several visualization options to fully understand and support the results as easily readable plots ready to be included in publications. We demonstrate Gdaphen capabilities on several datasets and provide easily followable vignettes.ConclusionsGdaphen makes the analysis of phenotypic data much easier for medical or preclinical behavioral researchers, providing an integrated framework to perform: (1) pre-processing steps as data imputation or anonymization; (2) a full statistical assessment to identify which variables are the most important discriminators; and (3) state of the art visualizations ready for publication to support the conclusions of the analyses. Gdaphen is open-source and freely available at https://github.com/munizmom/gdaphen, together with vignettes, documentation for the functions and examples to guide you in each own implementation.

An Efficient Compressive Sensed Video Codec with Inter-Frame Decoding and Low-Complexity Intra-Frame Encoding.

This paper is dedicated to video coding based on a compressive sensing (CS) framework. In CS, it is assumed that if a video sequence is sparse in some transform domain, then it could be reconstructed from a much lower number of samples (called measurements) than the Nyquist-Shannon theorem requires. Here, the performance of such a codec depends on how the measurements are acquired (or sensed) and compressed and how the video is reconstructed from the decoded measurements. Here, such a codec potentially could provide significantly faster encoding compared with traditional block-based intra-frame encoding via Motion JPEG (MJPEG), H.264/AVC or H.265/HEVC standards. However, existing video codecs based on CS are inferior to the traditional codecs in rate distortion performance, which makes them useless in practical scenarios. In this paper, we present a video codec based on CS called CS-JPEG. To the author's knowledge, CS-JPEG is the first codec based on CS, combining fast encoding and high rate distortion results. Our performance evaluation shows that, compared with the optimized software implementations of MJPEG, H.264/AVC, and H.265/HEVC, the proposed CS-JPEG encoding is 2.2, 1.9, and 30.5 times faster, providing 2.33, 0.79, and 1.45 dB improvements in the peak signal-to-noise ratio, respectively. Therefore, it could be more attractive for video applications having critical limitations in computational resources or a battery lifetime of an upstreaming device.

Reconstruction of incomplete wildfire data using deep generative models

We present our submission to the Extreme Value Analysis 2021 Data Challenge in which teams were asked to accurately predict distributions of wildfire frequency and size within spatio-temporal regions of missing data. For this competition, we developed a variant of the powerful variational autoencoder models, which we call Conditional Missing data Importance-Weighted Autoencoder (CMIWAE). Our deep latent variable generative model requires little to no feature engineering and does not necessarily rely on the specifics of scoring in the Data Challenge. It is fully trained on incomplete data, with the single objective to maximize log-likelihood of the observed wildfire information. We mitigate the effects of the relatively low number of training samples by stochastic sampling from a variational latent variable distribution, as well as by ensembling a set of CMIWAE models trained and validated on different splits of the provided data.

Metabolic Signature of Energy Metabolism Alterations and Excess Nitric Oxide Production in Culture Media Correlate with Low Human Embryo Quality and Unsuccessful Pregnancy.

Notwithstanding the great improvement of ART, the overall rate of successful pregnancies from implanted human embryos is definitely low. The current routine embryo quality assessment is performed only through morphological criteria, which has poor predictive capacity since only a minor percentage of those in the highest class give rise to successful pregnancy. Previous studies highlighted the potentiality of the analysis of metabolites in human embryo culture media, useful for the selection of embryos for implantation. In the present study, we analyzed in blind 66 human embryo culture media at 5 days after in vitro fertilization with the aim of quantifying compounds released by cell metabolism that were not present as normal constituents of the human embryo growth media, including purines, pyrimidines, nitrite, and nitrate. Only some purines were detectable (hypoxanthine and uric acid) in the majority of samples, while nitrite and nitrate were always detectable. When matching biochemical results with morphological evaluation, it was found that low grade embryos (n = 12) had significantly higher levels of all the compounds of interest. Moreover, when matching biochemical results according to successful (n = 17) or unsuccessful (n = 25) pregnancy, it was found that human embryos from the latter group released higher concentrations of hypoxanthine, uric acid, nitrite, and nitrate in the culture media. Additionally, those embryos that developed into successful pregnancies were all associated with the birth of healthy newborns. These results, although carried out on a relatively low number of samples, indicate that the analysis of the aforementioned compounds in the culture media of human embryos is a potentially useful tool for the selection of embryos for implantation, possibly leading to an increase in the overall rate of ART.

Weakly supervised anomaly detection coupled with Fourier transform infrared (FT-IR) spectroscopy for the identification of non-normal tissue.

The detection and classification of histopathological abnormal tissue constituents using machine learning (ML) techniques generally requires example data for each tissue or cell type of interest. This creates problems for studies on tissue that will have few regions of interest, or for those looking to identify and classify diseases of rarity, resulting in inadequate sample sizes from which to build multivariate and ML models. Regarding the impact on vibrational spectroscopy, specifically infrared (IR) spectroscopy, low numbers of samples may result in ineffective modelling of the chemical composition of sample groups, resulting in detection and classification errors. Anomaly detection may be a solution to this problem, enabling users to effectively model tissue constituents considered to represent normal tissue to capture any abnormal tissue and identify instances of non-normal tissue, be it disease or spectral artefacts. This work illustrates how a novel approach using a weakly supervised anomaly detection algorithm paired with IR microscopy can detect non-normal tissue spectra. In addition to incidental interferents such as hair, dust, and tissue scratches, the algorithm can also detect regions of diseased tissue. The model is never introduced to instances of these groups, training solely on healthy control data using only the IR spectral fingerprint region. This approach is demonstrated using liver tissue data from an agrochemical exposure mouse study.

Makine Öğrenmesinde Dengesiz Sınıf Problemiyle Başa Çıkmak İçin Sınıf Ağırlıklandırma Tekniği Kullanımı: Metodolojik Araştırma

Objective: Machine learning algorithms are based upon the assumption that data are balanced and so they do not provide good results in imbalanced datasets. This study aimed to explain the methods to be used for fitting a highly accurate model which better classifies the class of interest in imbalanced datasets with the class having a lower number of samples. Material and Methods: The study was planned as a methodological research. There are several weighting methods to calculate the class weight. This study included 4 most frequently used weighting methods. These are inverse of number of samples, inverse of square root of number of samples, effective number of examples and sample based class weight methods. In our study, 4 different class weighting methods were used on random forest and support vector machine, and it was explained how those methods affected class-based performances and the overall performance. Results: In simulated datasets, the best performance was achieved using the using the inverse of square root of number of samples class weighting method both on random forest and support vector machine. In real dataset, the best performance was achieved using the sample based class weight class weighting method on support vector machine. Conclusion: It was seen that all of the class weighting methods used in both machine learning methods were found to increase the performance of the class where recurrence was seen, therefore increasing the overall performance. It has been seen how effective the class weighting method is in dealing with the class imbalance problem.

Recent Trends of Dengue Virus Infection in Dhaka, Bangladesh

The disease caused by the Dengue virus (DENV), belongs to the Flaviviridae family, which is the most common vector for transmission of the virus between humans and arthropods Aedes aegypti mosquitoes. In recent years, the number of dengue cases has increased drastically throughout the world, according to a survey around 390 million dengue infection cases are recorded annually, among which one-fourth of cases show clinical manifestations. In Bangladesh, dengue outbreaks are on the increasing trend despite the efforts to control this disease A Hospital-based study of dengue patients in Dhaka city was conducted to check the recent trends of dengue infection among the inhabitants of Dhaka, Bangladesh. 89 blood samples from different patients with appropriate symptoms were tested using the Dengue NS1 Quanticard kit. Among samples, 60 (68%) samples were tested as positive whereas 22 (25%) showed negative results and 6 (7%) showed invalid results in tests. Despite the testing of a low number of samples, it can be clearly stated that a large number of positive results was observed with specific symptoms commonly found in Dengue infection. Among all the positive results some (7%) of the patients did not show any signs and symptoms, but after testing the results were positive against the viral antigen. Attempt to increase awareness about dengue fever through effective management measures from the government and NGOs are very much needed. Stamford Journal of Microbiology, Vol.12 (1) 2022: 43-46

Australian Rotavirus Surveillance Program: Annual Report, 2021.

This report from the Australian Rotavirus Surveillance Program describes the circulating rotavirus genotypes identified in children and adults during the period 1 January to 31 December 2021. During this period, 521 faecal specimens had been referred for rotavirus G- and P- genotype analysis, of which 474 were confirmed as rotavirus positive. Of these, 336/474 were wildtype rotavirus strains and 138/474 were identified as vaccine-like. Of the 336 wildtype samples, 87.5% (n = 294/336) were identified as G8P[8], and were detected in five of the six jurisdictions that provided samples for the reporting period. Two rotavirus outbreaks, located in the Northern Territory and Western Australia, were also attributed to G8P[8]. As with the 2020 reporting period, a low number of stool samples were received for this reporting period as a result of the COVID-19 pandemic. However, an unexpectedly high proportion of samples with unusual genotypes were identified which were potentially zoonotic in nature, including feline G3, P[9], bovine-like G8, P[14], and porcine-like G4, G6, P[1], and P[6]. Ongoing rotavirus surveillance is crucial to identify changes in genotypic patterns and to provide diagnostic laboratories with quality assurance by reporting incidences of wildtype, vaccine-like, or false positive rotavirus results.

Microbiological Safety of Cut Melons Sold in Portuguese Retail Markets: A Pilot Study.

Due to the increasing consciousness of a healthy diet and pursuit of convenience among consumers, the market for fresh fruit is on the rise, and the melon is among the most welcome of fruits for its sensory attributes and nutritional properties. Consumption safety of cut fruit remains an issue of concern that may affect public health. This study aimed to perform the microbiological characterisation of a melon, Cucumis melo L. var. “Piel de Sapo”, cut by retailers, wrapped in plastic cling film and kept at room temperature in local fruit shops. In addition, the possible transfer of relevant foodborne pathogens, during slicing, from the peel to the interior of the melon, and bacterial growth, were also evaluated when the melon slices were stored at abusive temperatures for 2 days. In this pilot study, a low number of samples were characterised microbiologically (26 cut melons), and some isolates were identified by 16S rRNA sequencing. No Listeria spp. or Salmonella spp. were detected in any of the samples, while Escherichia coli and Staphylococcus aureus were present in four and six out of twenty-six samples, respectively. Following artificial contamination of melons with cocktails of Salmonella spp., E. coli and Listeria monocytogenes, it was observed that, despite the smaller number of L. monocytogenes recovered, all the pathogens were transferred from the contaminated peels to the interior of the melons. Furthermore, over storage time, significant differences were observed (p < 0.05) between the counts obtained from melon slices immediately after cutting (0 h), and after 24 and 48 h at 20 °C, with an increase of about 4 log CFU/g in all the pathogens. In conclusion, some cut melons classified as microbiologically unacceptable or unsatisfactory are being sold in local fruit shops in the Porto Metropolitan Area, Portugal. Although absent in the samples analysed, Salmonella spp. and L. monocytogenes, if present, can be transferred from the outside to the inside of the fruit by the cutting blade and, if not consumed immediately and stored at abusive temperatures, this ready-to-eat product poses a risk of infection. This pilot study, performed for the first time in Portugal under these conditions, clearly demonstrates the need for education campaigns to alert local sellers and consumers of the risk posed by cut melons.

Construction of a nomogram for predicting COVID-19 in-hospital mortality: A machine learning analysis

Background and objectivesWe aim to verify the use of ML algorithms to predict patient outcome using a relatively small dataset and to create a nomogram to assess in-hospital mortality of patients with COVID-19. MethodsA database of 200 COVID-19 patients admitted to the Clinical Hospital of State University of Campinas (UNICAMP) was used in this analysis. Patient features were divided into three categories: clinical, chest abnormalities, and body composition characteristics acquired by computerized tomography. These features were evaluated independently and combined to predict patient outcomes. To minimize performance fluctuations due to low sample number, reduce possible bias related to outliers, and evaluate the uncertainties generated by the small dataset, we developed a shuffling technique, a modified version of the Monte Carlo Cross Validation, creating several subgroups for training the algorithm and complementary testing subgroups. The following ML algorithms were tested: random forest, boosted decision trees, logistic regression, support vector machines, and neural networks. Performance was evaluated by analyzing Receiver operating characteristic (ROC) curves. The importance of each feature in the determination of the outcome predictability was also studied and a nomogram was created based on the most important features selected by the exclusion test. ResultsAmong the different sets of features, clinical variables age, lymphocyte number and weight were the most valuable features for prognosis prediction. However, we observed that skeletal muscle radiodensity and presence of pleural effusion were also important for outcome determination. Integrating these independent predictors was successfully developed to accurately predict mortality in COVID-19 in hospital patients. A nomogram based on these five features was created to predict COVID-19 mortality in hospitalized patients. The area under the ROC curve was 0.86 ± 0.04. ConclusionML algorithms can be reliable for the prediction of COVID-19-related in-hospital mortality, even when using a relatively small dataset. The success of ML techniques in smaller datasets broadens the applicability of these methods in several problems in the medical area. In addition, feature importance analysis allowed us to determine the most important variables for the prediction tasks resulting in a nomogram with good accuracy and clinical utility in predicting COVID-19 in-hospital mortality.

Unified Theory of Characteristic Modes—Part II: Tracking, Losses, and FEM Evaluation

This is the second component of a two-part article dealing with a unification of characteristic mode decomposition. This second part addresses modal tracking, interpolation, and the role of ohmic losses, and presents several numerical examples for surface-based method-of-moment (MoM) formulations. A new tracking algorithm based on the algebraic properties of the transition matrix is developed, achieving excellent precision and requiring a very low number of frequency samples when compared with procedures previously reported in the literature. The transition matrix is further used to show that characteristic mode decomposition of lossy objects fails to deliver orthogonal far-fields and to demonstrate how characteristic modes can be evaluated using the finite element method (FEM).

TV‐L1 Ordinal Logistic Regression Reveals New Morphometric Patterns Related to Parkinsonian Symptom Severity: An ENIGMA‐PD study

AbstractBackgroundParkinson’s disease (PD) is heterogeneous, both phenotypically and in terms of temporal progression. The Hoehn and Yahr (HY) scale is a well‐established PD staging approach, and identifies 5 stages of the disease. Morphometric effects in deep gray matter regions of the brain associated with HY stages are complex; a recent large‐scale ENIGMA‐PD study showed higher local subcortical volumes in early HY stages relative to controls, followed by a precipitous decrease after stage 2 [1]. This finding motivates a closer look at fine‐level morphometry beyond gross volume measures. Here, we developed and applied a novel machine learning algorithm to reveal the subcortical shape signatures of HY staging.MethodWe computed shape features in 7 bilateral subcortical regions [2] based on T1‐weighted MRI data from 2,322 PD subjects and 1,207 controls from 20 ENIGMA‐PD cohorts (HY stages in Table 1). We developed a sparse, spatially coherent (total variation/TV‐L1) ordinal linear logistic classifier [3] to predict HY stages with a single linear model. We applied the model to vertex‐wise medial thickness features. We optimized regularization parameters for balanced recall (sensitivity) and precision using a 4‐fold cross‐validation grid search. Very low numbers of HY4 and HY5 samples necessitated merging stages 3‐5 into one category. For comparison, we also trained 4 binary TV‐L1 logit models on the same features [4], discriminating (1) PD‐Control; (2) HY1‐HY2; (3) HY1‐HY345; (4) HY2‐HY345, using ROC area‐under‐the‐curve (AUC) evaluation.ResultAcross‐stage mean out‐of‐sample precision and recall were 0.43, and 0.393, respectively (chance=0.33). Table 2 shows the confusion matrix and precision/recall for each HY stage. All models’ linear coefficient maps are displayed in Figures 1,2. Binary classification ROC‐AUC was 0.66 for PD‐Control, and ranged from 0.62 to 0.73 for HY prediction (Figure 2).ConclusionWe developed an ordit machine learning model for morphometric shape‐based ordinal classification of disease stages, training it for Parkinson’s Disease Hoehn and Yahr stage prediction on a large MRI collection. Performance was substantially above chance. Model weight maps indicate early increased thalamic thickness, followed by a complex thinning pattern associated with later HY stages.