The De Ritis ratio as prognostic biomarker of in-hospital mortality in COVID-19 patients.

Increased concentrations of serum aspartate transaminase (AST) and alanine transaminase (ALT) are common in COVID-19 patients. However, their capacity to predict mortality, particularly the AST/ALT ratio, commonly referred to as the De Ritis ratio, is unknown. We investigated the association between the De Ritis ratio on admission and in-hospital mortality in 105 consecutive patients with coronavirus disease of 2019 (COVID-19) admitted to three COVID-19 referral centres in Sardinia, Italy. The De Ritis ratio was significantly lower in survivors than nonsurvivors (median: 1.25; IQR: 0.91-1.64 vs 1.67; IQR: 1.38-1.97, P = .002) whilst there were no significant between-group differences in ALT and AST concentrations. In ROC curve analysis, the AUC value of the De Ritis ratio was 0.701 (95% CI 0.603-0.787, P = .0006) with sensitivity and specificity of 74% and 70%, respectively. Kaplan-Meier survival curves showed a significant association between the De Ritis ratio and mortality (logrank test P = .014). By contrast, no associations were observed between the ALT and AST concentrations and mortality (logrank test P = .83 and P = .62, respectively). In multivariate Cox regression analysis, the HR in patients with De Ritis ratios ≥1.63 (upper tertile of this parameter) remained significant after adjusting for age, gender, smoking status, cardiovascular disease, intensity of care, diabetes, respiratory diseases, malignancies and kidney disease (HR: 2.46, 95% CI 1.05-5.73, P = .037). Therefore, the De Ritis ratio on admission was significantly associated with in-hospital mortality in COVID-19 patients. Larger studies are required to confirm the capacity of this parameter to independently predict mortality in this group.

Deep-Framework: A Distributed, Scalable, and Edge-Oriented Framework for Real-Time Analysis of Video Streams.

Edge computing is the best approach for meeting the exponential demand and the real-time requirements of many video analytics applications. Since most of the recent advances regarding the extraction of information from images and video rely on computation heavy deep learning algorithms, there is a growing need for solutions that allow the deployment and use of new models on scalable and flexible edge architectures. In this work, we present Deep-Framework, a novel open source framework for developing edge-oriented real-time video analytics applications based on deep learning. Deep-Framework has a scalable multi-stream architecture based on Docker and abstracts away from the user the complexity of cluster configuration, orchestration of services, and GPU resources allocation. It provides Python interfaces for integrating deep learning models developed with the most popular frameworks and also provides high-level APIs based on standard HTTP and WebRTC interfaces for consuming the extracted video data on clients running on browsers or any other web-based platform.

Application of a new method for GWAS in a related case/control sample with known pedigree structure: identification of new loci for nephrolithiasis.

In contrast to large GWA studies based on thousands of individuals and large meta-analyses combining GWAS results, we analyzed a small case/control sample for uric acid nephrolithiasis. Our cohort of closely related individuals is derived from a small, genetically isolated village in Sardinia, with well-characterized genealogical data linking the extant population up to the 16(th) century. It is expected that the number of risk alleles involved in complex disorders is smaller in isolated founder populations than in more diverse populations, and the power to detect association with complex traits may be increased when related, homogeneous affected individuals are selected, as they are more likely to be enriched with and share specific risk variants than are unrelated, affected individuals from the general population. When related individuals are included in an association study, correlations among relatives must be accurately taken into account to ensure validity of the results. A recently proposed association method uses an empirical genotypic covariance matrix estimated from genome-screen data to allow for additional population structure and cryptic relatedness that may not be captured by the genealogical data. We apply the method to our data, and we also investigate the properties of the method, as well as other association methods, in our highly inbred population, as previous applications were to outbred samples. The more promising regions identified in our initial study in the genetic isolate were then further investigated in an independent sample collected from the Italian population. Among the loci that showed association in this study, we observed evidence of a possible involvement of the region encompassing the gene LRRC16A, already associated to serum uric acid levels in a large meta-analysis of 14 GWAS, suggesting that this locus might lead a pathway for uric acid metabolism that may be involved in gout as well as in nephrolithiasis.

Fine mapping of five loci associated with low-density lipoprotein cholesterol detects variants that double the explained heritability.

Complex trait genome-wide association studies (GWAS) provide an efficient strategy for evaluating large numbers of common variants in large numbers of individuals and for identifying trait-associated variants. Nevertheless, GWAS often leave much of the trait heritability unexplained. We hypothesized that some of this unexplained heritability might be due to common and rare variants that reside in GWAS identified loci but lack appropriate proxies in modern genotyping arrays. To assess this hypothesis, we re-examined 7 genes (APOE, APOC1, APOC2, SORT1, LDLR, APOB, and PCSK9) in 5 loci associated with low-density lipoprotein cholesterol (LDL-C) in multiple GWAS. For each gene, we first catalogued genetic variation by re-sequencing 256 Sardinian individuals with extreme LDL-C values. Next, we genotyped variants identified by us and by the 1000 Genomes Project (totaling 3,277 SNPs) in 5,524 volunteers. We found that in one locus (PCSK9) the GWAS signal could be explained by a previously described low-frequency variant and that in three loci (PCSK9, APOE, and LDLR) there were additional variants independently associated with LDL-C, including a novel and rare LDLR variant that seems specific to Sardinians. Overall, this more detailed assessment of SNP variation in these loci increased estimates of the heritability of LDL-C accounted for by these genes from 3.1% to 6.5%. All association signals and the heritability estimates were successfully confirmed in a sample of ∼10,000 Finnish and Norwegian individuals. Our results thus suggest that focusing on variants accessible via GWAS can lead to clear underestimates of the trait heritability explained by a set of loci. Further, our results suggest that, as prelude to large-scale sequencing efforts, targeted re-sequencing efforts paired with large-scale genotyping will increase estimates of complex trait heritability explained by known loci.

Pneumomediastinum in COVID-19: Risk factors and outcomes from a multicentre case-control study.

BACKGROUND: An increased incidence of pneumomediastinum has been observed among patients hospitalized with coronavirus disease 2019 (COVID-19) pneumonia. The study aimed to identify risk factors for COVID-19-associated pneumomediastinum and investigate the impact of pneumomediastinum on clinical outcomes. METHODS: In this multicentre retrospective case-control study, we included consecutive patients with COVID-19 pneumonia and pneumomediastinum hospitalized from March 2020 to July 2020 at ten centres; then, we identified a similarly sized control group of consecutive patients hospitalized with COVID-19 pneumonia and respiratory failure who did not develop pneumomediastinum during the same period. Clinical, laboratory, and radiological characteristics, as well as respiratory support and outcomes, were collected and compared between the two groups. Risk factors of pneumomediastinum were assessed by multivariable logistic analysis. RESULTS: Overall 139 patients with pneumomediastinum and 153 without pneumomediastinum were analysed. Lung involvement ≥75 %, consolidations, body mass index (BMI) < 22 kg/m2, C-reactive protein (CRP) > 150 mg/L, D-dimer >3000 ng/mL FEUs, and smoking exposure >20 pack-year were all independently correlated with the occurrence of pneumomediastinum. Patients with pneumomediastinum had a longer hospital stay (mean ± SD 31.2 ± 20.2 days vs 19.6 ± 14.2, p < 0.001), higher intubation rate (73/139, 52.5 % vs 27/153, 17.6 %, p < 0.001), and in-hospital mortality (68/139, 48.9 % vs 36/153, 23.5 %, p < 0.001) compared to controls. CONCLUSIONS: Extensive lung parenchyma involvement, consolidations, low BMI, high inflammatory markers, and tobacco exposure are associated with a greater risk of pneumomediastinum in COVID-19 pneumonia. This complication significantly worsens the outcomes.

A strategy analysis for genetic association studies with known inbreeding.

BACKGROUND: Association studies consist in identifying the genetic variants which are related to a specific disease through the use of statistical multiple hypothesis testing or segregation analysis in pedigrees. This type of studies has been very successful in the case of Mendelian monogenic disorders while it has been less successful in identifying genetic variants related to complex diseases where the insurgence depends on the interactions between different genes and the environment. The current technology allows to genotype more than a million of markers and this number has been rapidly increasing in the last years with the imputation based on templates sets and whole genome sequencing. This type of data introduces a great amount of noise in the statistical analysis and usually requires a great number of samples. Current methods seldom take into account gene-gene and gene-environment interactions which are fundamental especially in complex diseases. In this paper we propose to use a non-parametric additive model to detect the genetic variants related to diseases which accounts for interactions of unknown order. Although this is not new to the current literature, we show that in an isolated population, where the most related subjects share also most of their genetic code, the use of additive models may be improved if the available genealogical tree is taken into account. Specifically, we form a sample of cases and controls with the highest inbreeding by means of the Hungarian method, and estimate the set of genes/environmental variables, associated with the disease, by means of Random Forest. RESULTS: We have evidence, from statistical theory, simulations and two applications, that we build a suitable procedure to eliminate stratification between cases and controls and that it also has enough precision in identifying genetic variants responsible for a disease. This procedure has been successfully used for the beta-thalassemia, which is a well known Mendelian disease, and also to the common asthma where we have identified candidate genes that underlie to the susceptibility of the asthma. Some of such candidate genes have been also found related to common asthma in the current literature. CONCLUSIONS: The data analysis approach, based on selecting the most related cases and controls along with the Random Forest model, is a powerful tool for detecting genetic variants associated to a disease in isolated populations. Moreover, this method provides also a prediction model that has accuracy in estimating the unknown disease status and that can be generally used to build kit tests for a wide class of Mendelian diseases.

Microsatellites and SNPs linkage analysis in a Sardinian genetic isolate confirms several essential hypertension loci previously identified in different populations.

BACKGROUND: A multiplicity of study designs such as gene candidate analysis, genome wide search (GWS) and, recently, whole genome association studies have been employed for the identification of the genetic components of essential hypertension (EH). Several genome-wide linkage studies of EH and blood pressure-related phenotypes demonstrate that there is no single locus with a major effect while several genomic regions likely to contain EH-susceptibility loci were validated by multiple studies. METHODS: We carried out the clinical assessment of the entire adult population in a Sardinian village (Talana) and we analyzed 16 selected families with 62 hypertensive subjects out of 267 individuals. We carried out a double GWS using a set of 902 uniformly spaced microsatellites and a high-density SNPs map on the same group of families. RESULTS: Three loci were identified by both microsatellites and SNP scans and the obtained linkage results showed a remarkable degree of similarity. These loci were identified on chromosome 2q24, 11q23.1-25 and 13q14.11-21.33. Further support to these findings is their broad description present in literature associated to EH or related phenotypes. Bioinformatic investigation of these loci shows several potential EH candidate genes, several of whom already associated to blood pressure regulation pathways. CONCLUSION: Our search for major susceptibility EH genetic factors evidences that EH in the genetic isolate of Talana is due to the contribution of several genes contained in loci identified and replicated by earlier findings in different human populations.

Methylenetetrahydrofolate reductase gene polymorphisms in Burkina Faso: impact on plasma fasting homocysteine and after methionine loading test.

In Burkina Faso the levels of plasma homocysteine (Hcy) are lower and the methionine loading tests suggest a more effective Hcy metabolism. The polymorphisms of methylenetetrahydrofolate reductase (MTHFR) showed a relevant difference in the allele frequencies of T MTHFR-677 in young and in old subjects, while the allele frequency of C MTHFR-1298 was comparable in young and old subjects. The aim of this paper was to study the impact of the MTHFR polymorphisms on plasma fasting Hcy and after methionine loading in Burkina Faso. The young subjects with CC MTHFR-677 genotype had levels of Hcy significantly lower than CT and TT subjects. The level of Hcy in subjects who had AA, AC and CC MTHFR-1298 genotypes were comparable. The levels of Hcy after the methionine loading test were significantly higher in CT and TT MTHFR-677 genotype. These results suggest that the genetic situation in Burkina Faso is different from that of other Western countries and this guarantees the maintenance of lower plasma levels of Hcy in young and old Africans. The elevated levels of plasma Hcy in old subjects compared to young subjects, against the low prevalence of the T allele in elderly subjects, is discussed.

High differentiation among eight villages in a secluded area of Sardinia revealed by genome-wide high density SNPs analysis.

To better design association studies for complex traits in isolated populations it's important to understand how history and isolation moulded the genetic features of different communities. Population isolates should not "a priori" be considered homogeneous, even if the communities are not distant and part of a small region. We studied a particular area of Sardinia called Ogliastra, characterized by the presence of several distinct villages that display different history, immigration events and population size. Cultural and geographic isolation characterized the history of these communities. We determined LD parameters in 8 villages and defined population structure through high density SNPs (about 360 K) on 360 unrelated people (45 selected samples from each village). These isolates showed differences in LD values and LD map length. Five of these villages show high LD values probably due to their reduced population size and extreme isolation. High genetic differentiation among villages was detected. Moreover population structure analysis revealed a high correlation between genetic and geographic distances. Our study indicates that history, geography and biodemography have influenced the genetic features of Ogliastra communities producing differences in LD and population structure. All these data demonstrate that we can consider each village an isolate with specific characteristics. We suggest that, in order to optimize the study design of complex traits, a thorough characterization of genetic features is useful to identify the presence of sub-populations and stratification within genetic isolates.

EDA2R is associated with androgenetic alopecia.

Androgenetic alopecia (AGA) is a common heritable polygenic disorder whose genetics is not fully understood, even though it seems to be X-linked. We carried out an epidemiological survey for AGA on 9,000 people from 8 isolated villages of a secluded region of Sardinia (Ogliastra), and identified a large cohort of affected individuals. We genotyped 200 cases and 200 controls (mean kinship 0.001) with the 500k chip array and conducted case-control association analysis on the X chromosome. We identified Xq11-q12 as strongly associated with AGA. In particular, we found that rs1352015 located 8 kb from the EDA2R gene showed the best result (P=7.77e(-7)). This region also contains the AR gene, hence we tested both genes in 492 cases and 492 controls. We found that the non-synonymous SNP rs1385699 on EDA2R gave the best result (P=3.9e(-19)) whereas rs6152 on the AR gene is less significant (P=4.17e(-12)). Further statistical analysis carried out by conditioning each gene to the presence of the other showed that the association with EDA2R is independent while the association with AR seems to be the result of linkage disequilibrium. These results give insight into the pathways involved in AGA etiology.