RESUMEN
PURPOSE: This study aimed to develop a classifier using supervised machine learning to effectively assess the impact of clinical, demographical, and biochemical factors in accurately predicting the antiseizure medications (ASMs) treatment response in people with epilepsy (PWE). METHODS: Data was collected from 786 PWE at the Outpatient Department of Neurology, Institute of Human Behavior and Allied Sciences (IHBAS), New Delhi, India from 2005 to 2015. Patients were followed up at the 2nd, 4th, 8th, and 12th month over the span of 1 year for the drugs being administered and their dosage, the serum drug levels, the frequency of seizure control, drug efficacy, the adverse drug reactions (ADRs), and their compliance to ASMs. Several features, including demographic details, medical history, and auxiliary examinations electroencephalogram (EEG) or Computed Tomography (CT) were chosen to discern between patients with distinct remission outcomes. Remission outcomes were categorized into 'good responder (GR)' and 'poor responder (PR)' based on the number of seizures experienced by the patients over the study duration. Our dataset was utilized to train seven classical machine learning algorithms i.e Extreme Gradient Boost (XGB), K-Nearest Neighbor (KNN), Support Vector Classifier (SVC), Decision Tree (DT), Random Forest (RF), Naïve Bayes (NB) and Logistic Regression (LR) to construct classification models. RESULTS: Our research findings indicate that 1) among the seven algorithms examined, XGB and SVC demonstrated superior predictive performances of ASM treatment outcomes with an accuracy of 0.66 each and ROC-AUC scores of 0.67 (XGB) and 0.66 (SVC) in distinguishing between PR and GR patients. 2) The most influential factor in discerning PR to GR patients is a family history of seizures (no), education (literate) and multitherapy with Chi-square (χ2) values of 12.1539, 8.7232 and 13.620 respectively and odds ratio (OR) of 2.2671, 0.4467, and 1.9453 each. 3). Furthermore, our surrogate analysis revealed that the null hypothesis for both XGB and SVC was rejected at a 100â¯% confidence level, underscoring the significance of their predictive performance. These findings underscore the robustness and reliability of XGB and SVC in our predictive modelling framework. SIGNIFICANCE: Utilizing XG Boost and SVC-based machine learning classifier, we successfully forecasted the likelihood of a patient's response to ASM treatment, categorizing them as either PR or GR, post-completion of standard epilepsy examinations. The classifier's predictions were found to be statistically significant, suggesting their potential utility in improving treatment strategies, particularly in the personalized selection of ASM regimens for individual epilepsy patients.
RESUMEN
BACKGROUND: The clinical heterogeneity in major depressive disorder (MDD), variable treatment response, and conflicting findings limit the ability of genomics toward the discovery of evidence-based diagnosis and treatment regimen. This study attempts to curate all genetic association findings to evaluate potential variants for clinical translation. METHODS: We systematically reviewed all candidates and genome-wide association studies for both MDD susceptibility and antidepressant response, independently, using MEDLINE, particularly to identify replicated findings. These variants were evaluated for functional consequences using different in silico tools and further estimated their diagnostic predictability by calculating positive predictive values. RESULTS: A total of 217 significantly associated studies comprising 1200 variants across 545 genes and 128 studies including 921 variants across 412 genes were included with MDD susceptibility and antidepressant response, respectively. Although the majority of associations were confirmed by a single study, we identified 31 and 18 replicated variants (in at least 2 studies) for MDD and antidepressant response. Functional annotation of these 31 variants predicted 20% coding variants as deleterious/damaging and 80.6% variants with regulatory effect. Similarly, the response-related 18 variants revealed 25% coding variant as damaging and 88.2% with substantial regulatory potential. Finally, we could calculate the diagnostic predictability of 19 and 5 variants whose positive predictive values ranges from 0.49 to 0.66 for MDD and 0.36 to 0.66 for response. CONCLUSIONS: The replicated variants presented in our data are promising for disease diagnosis and improved response outcomes. Although these quantitative assessment measures are solely directive of available observational evidence, robust homogenous validation studies are required to strengthen these variants for molecular diagnostic application.
Asunto(s)
Trastorno Depresivo Mayor , Humanos , Trastorno Depresivo Mayor/diagnóstico , Trastorno Depresivo Mayor/tratamiento farmacológico , Trastorno Depresivo Mayor/genética , Estudio de Asociación del Genoma Completo , Antidepresivos/uso terapéuticoRESUMEN
The humancytochrome P450 1A (CYP1A) subfamily genes, CYP1A1 and CYP1A2, encoding monooxygenases are critically involved in biotransformation of key endogenous substrates (estradiol, arachidonic acid, cholesterol) and exogenous compounds (smoke constituents, carcinogens, caffeine, therapeutic drugs). This suggests their significant involvement in multiple biological pathways with a primary role of maintaining endogenous homeostasis and xenobiotic detoxification. Large interindividual variability exist in CYP1A gene expression and/or catalytic activity of the enzyme, which is primarily due to the existence of polymorphic alleles which encode them. These polymorphisms (mainly single nucleotide polymorphisms, SNPs) have been extensively studied as susceptibility factors in a spectrum of clinical phenotypes. An in-depth understanding of the effects of polymorphic CYP1A genes on the differential metabolic activity and the resulting biological pathways is needed to explain the clinical implications of CYP1A polymorphisms. The present review is intended to provide an integrated understanding of CYP1A metabolic activity with unique substrate specificity and their involvement in physiological and pathophysiological roles. The article further emphasizes on the impact of widely studied CYP1A1 and CYP1A2 SNPs and their complex interaction with non-genetic factors like smoking and caffeine intake on multiple clinical phenotypes. Finally, we attempted to discuss the alterations in metabolism/physiology concerning the polymorphic CYP1A genes, which may underlie the reported clinical associations. This knowledge may provide insights into the disease pathogenesis, risk stratification, response to therapy and potential drug targets for individuals with certain CYP1A genotypes.
Asunto(s)
Citocromo P-450 CYP1A1 , Citocromo P-450 CYP1A2 , Cafeína , Citocromo P-450 CYP1A1/genética , Citocromo P-450 CYP1A2/genética , Citocromo P-450 CYP1A2/metabolismo , Oxigenasas de Función Mixta/genética , Polimorfismo de Nucleótido Simple , HumanosRESUMEN
Pneumonia, an acute respiratory tract infection, is one of the major causes of mortality worldwide. Depending on the site of acquisition, pneumonia can be community acquired pneumonia (CAP) or nosocomial pneumonia (NP). The risk of pneumonia, is partially driven by host genetics. CYP1A1 is a widely studied pulmonary CYP family gene primarily expressed in peripheral airway epithelium. The CYP1A1 genetic variants, included in this study, alter the gene activity and are known to contribute in lung inflammation, which may cause pneumonia pathogenesis. In this study, we performed a meta-analysis to establish the possible contribution of CYP1A1 gene, and its three variants (rs2606345, rs1048943 and rs4646903) towards the genetic etiology of pneumonia risk. Using PRISMA guidelines, we systematically reviewed and meta-analysed case-control studies, evaluating risk of pneumonia in patients carrying the risk alleles of CYP1A1 variants. Heterogeneity across the studies was evaluated using I2 statistics. Based on heterogeneity, a random-effect (using maximum likelihood) or fixed-effect (using inverse variance) model was applied to estimate the effect size. Pooled odds ratio (OR) was calculated to estimate the overall effect of the risk allele association with pneumonia susceptibility. Egger's regression test and funnel plot were used to assess publication bias. Subgroup analysis was performed based on pneumonia type (CAP and NP), population, as well as age group. A total of ten articles were identified as eligible studies, which included 3049 cases and 2249 healthy controls. The meta-analysis findings revealed CYP1A1 variants, rs2606345 [T vs G; OR = 1.12 (0.75-1.50); p = 0.02; I2 = 84.89%], and rs1048943 [G vs T; OR = 1.19 (0.76-1.61); p = 0.02; I2 = 0.00%] as risk markers whereas rs4646903 showed no statistical significance for susceptibility to pneumonia. On subgroup analysis, both the genetic variants showed significant association with CAP but not with NP. We additionally performed a spatial analysis to identify the key factors possibly explaining the variability across countries in the prevalence of the coronavirus disease 2019 (COVID-19), a viral pneumonia. We observed a significant association between the risk allele of rs2606345 and rs1048943, with a higher COVID-19 prevalence worldwide, providing us important links in understanding the variability in COVID-19 prevalence.
Asunto(s)
COVID-19 , Infecciones Comunitarias Adquiridas , Neumonía , COVID-19/genética , Señales (Psicología) , Citocromo P-450 CYP1A1/genética , Predisposición Genética a la Enfermedad , Genética Humana , Humanos , Neumonía/genética , Polimorfismo de Nucleótido Simple , Factores de RiesgoRESUMEN
Epilepsy, a neurological disease characterized by recurrent seizures, is highly heterogeneous in nature. Based on the prevalence, epilepsy is classified into two types: common and rare epilepsies. Common epilepsies affecting nearly 95% people with epilepsy, comprise generalized epilepsy which encompass idiopathic generalized epilepsy like childhood absence epilepsy, juvenile myoclonic epilepsy, juvenile absence epilepsy and epilepsy with generalized tonic-clonic seizure on awakening and focal epilepsy like temporal lobe epilepsy and cryptogenic focal epilepsy. In 70% of the epilepsy cases, genetic factors are responsible either as single genetic variant in rare epilepsies or multiple genetic variants acting along with different environmental factors as in common epilepsies. Genetic testing and precision treatment have been developed for a few rare epilepsies and is lacking for common epilepsies due to their complex nature of inheritance. Precision medicine for common epilepsies require a panoramic approach that incorporates polygenic background and other non-genetic factors like microbiome, diet, age at disease onset, optimal time for treatment and other lifestyle factors which influence seizure threshold. This review aims to comprehensively present a state-of-art review of all the genes and their genetic variants that are associated with all common epilepsy subtypes. It also encompasses the basis of these genes in the epileptogenesis. Here, we discussed the current status of the common epilepsy genetics and address the clinical application so far on evidence-based markers in prognosis, diagnosis, and treatment management. In addition, we assessed the diagnostic predictability of a few genetic markers used for disease risk prediction in individuals. A combination of deeper endo-phenotyping including pharmaco-response data, electro-clinical imaging, and other clinical measurements along with genetics may be used to diagnose common epilepsies and this marks a step ahead in precision medicine in common epilepsies management.