Development and external validation of a logistic and a penalized logistic model using machine-learning techniques to predict suicide attempts: A multicenter prospective cohort study in Korea.

Despite previous efforts to build statistical models for predicting the risk of suicidal behavior using machine-learning analysis, a high-accuracy model can lead to overfitting. Furthermore, internal validation cannot completely address this problem. In this study, we created models for predicting the occurrence of suicide attempts among Koreans at high risk of suicide, and we verified these models in an independent cohort. We performed logistic and penalized regression for suicide attempts within 6 months among suicidal ideators and attempters in The Korean Cohort for the Model Predicting a Suicide and Suicide-related Behavior (K-COMPASS). We then validated the models in a test cohort. Our findings indicated that several factors significantly predicted suicide attempts in the models, including young age, suicidal ideation, previous suicidal attempts, anxiety, alcohol abuse, stress, and impulsivity. The area under the curve and positive predictive values were 0.941 and 0.484 after variable selection and 0.751 and 0.084 in the test cohort. The corresponding values for the penalized regression model were 0.943 and 0.524 in the original training cohort and 0.794 and 0.115 in the test cohort. The prediction model constructed through a prospective cohort study of the suicide high-risk group showed satisfactory accuracy even in the test cohort. The accuracy with penalized regression was greater than that with the "classical" logistic model.

DeepPIG: deep neural network architecture with pairwise connected layers and stochastic gates using knockoff frameworks for feature selection.

Selecting relevant feature subsets is essential for machine learning applications. Among the feature selection techniques, the knockoff filter procedure proposes a unique framework that minimizes false discovery rates (FDR). However, employing a deep neural network architecture for a knockoff filter framework requires higher detection power. Using the knockoff filter framework, we present a Deep neural network with PaIrwise connected layers integrated with stochastic Gates (DeepPIG) for the feature selection model. DeepPIG exhibited better detection power in synthetic data than the baseline and recent models such as Deep feature selection using Paired-Input Nonlinear Knockoffs (DeepPINK), Stochastic Gates (STG), and SHapley Additive exPlanations (SHAP) while not violating the preselected FDR level, especially when the signal of the features were weak. The selected features determined by DeepPIG demonstrated superior classification performance compared with the baseline model in real-world data analyses, including the prediction of certain cancer prognosis and classification tasks using microbiome and single-cell datasets. In conclusion, DeepPIG is a robust feature selection approach even when the signals of features are weak. Source code is available at https://github.com/DMCB-GIST/DeepPIG .

Establishing a Framework for Evaluating the Effectiveness of Vaccines Targeting National Vaccination Programs.

BACKGROUND: The increasing number of vaccines and the complexity of immunization programs, along with continuous changes in the epidemiology of infectious diseases, necessitate a systematic approach to vaccine effectiveness (VE) evaluation. This study presents a preliminary survey to establish a VE evaluation framework in Korea, focusing on the National Immunization Program. METHODS: Experts' opinions were collected through a two-round online survey targeting key stakeholders. The first round consisted of two multiple-choice questions and two open-ended questions. The second round was a quantitative survey with 17 questionnaires based on five domains derived by analyzing the results of the first-round survey. RESULTS: The results emphasize the necessity and urgency of a government-led VE evaluation system and the establishment of a multidisciplinary evaluation organization. Key considerations include personnel, budget, data integration, legal standards, and surveillance system enhancements. CONCLUSION: These findings provide valuable insights for policymakers, emphasizing the need for collaboration, financial support, and robust data management in developing evidence-based vaccination policies.

One-year mortality and morbidities of severe fever with thrombocytopenia syndrome compared with other diseases: A nationwide cohort study in South Korea.

BACKGROUND: The long-term mortality and morbidity of patients with severe fever with thrombocytopenia syndrome (SFTS) remain unclear. METHODS: This retrospective cohort study was conducted using the National Health Insurance Service dataset on hospitalized patients with SFTS aged ≥20 years between 2016 and 2021 (n = 1,217). Each SFTS case was matched with three controls hospitalized for non-SFTS-related diseases using propensity score matching. The all-cause mortality of patients with SFTS was evaluated during the one-year follow-up and compared with that of controls. Post-discharge events were investigated to determine the effects of SFTS on post-acute sequelae. RESULTS: Finally, 1,105 patients with SFTS and 3,315 controls were included. Patients with SFTS had a higher risk of death during the one-year follow-up than that of controls (hazard ratio [HR], 2·26; 95% confidence interval [CI], 1·82-2·81). Thirty-day mortality was significantly higher in the SFTS group (HR, 3·99; 95% CI, 3·07-5·19) than in the control group. An increased risk of death after 31-365 days was observed among controls, though this difference was significant only among patients in their 80s (HR, 0·18; 95% CI, 0·06-0·57). For post-discharge events, patients in the SFTS group exhibited a higher risk of readmission (HR, 1·17; 95% CI, 1·04-1·32) and emergency room visit (HR, 2·32; 95% CI, 1·96-2·76) than those in the control group. CONCLUSION: SFTS induces a higher risk of short-term mortality and post-acute sequelae in hospitalized patients during a one-year follow-up than non-SFTS-related diseases. Our results provide guidance for the management of SFTS.

Once upon a Time, There Was a Piece of Wood: Present Knowledge and Future Perspectives in Fungal Deterioration of Wooden Cultural Heritage in Terrestrial Ecosystems and Diagnostic Tools.

Wooden Cultural Heritage (WCH) represents a significant portion of the world's historical and artistic heritage, consisting of immovable and movable artefacts. Despite the expertise developed since ancient times to enhance its durability, wooden artefacts are inevitably prone to degradation. Fungi play a pivotal role in the deterioration of WCH in terrestrial ecosystems, accelerating its decay and leading to alterations in color and strength. Reviewing the literature of the last 25 years, we aimed to provide a comprehensive overview of fungal diversity affecting WCH, the biochemical processes involved in wood decay, and the diagnostic tools available for fungal identification and damage evaluation. Climatic conditions influence the occurrence of fungal species in threatened WCH, characterized by a prevalence of wood-rot fungi (e.g., Serpula lacrymans, Coniophora puteana) in architectural heritage in temperate and continental climates and Ascomycota in indoor and harsh environments. More efforts are needed to address the knowledge fragmentation concerning biodiversity, the biology of the fungi involved, and succession in the degradative process, which is frequently centered solely on the main actors. Multidisciplinary collaboration among engineers, restorers, and life sciences scientists is vital for tackling the challenges posed by climate change with increased awareness. Traditional microbiology and culture collections are fundamental in laying solid foundations for a more comprehensive interpretation of big data.

Integrated proteomic and genomic analysis to identify predictive biomarkers for valproate response in bipolar disorder: a 6-month follow-up study.

BACKGROUND: Several genetic studies have been undertaken to elucidate the intricate interplay between genetics and drug responses in bipolar disorder (BD). However, there has been notably limited research on biomarkers specifically linked to valproate, with only a few studies investigating integrated proteomic and genomic factors in response to valproate treatment. Therefore, this study aimed to identify biological markers for the therapeutic response to valproate treatment in BD. Patients with BD in remission were assessed only at baseline, whereas those experiencing acute mood episodes were evaluated at three points (baseline, 8 ± 2 weeks, and 6 ± 1 months). The response to valproate treatment was measured using the Alda scale, with individuals scoring an Alda A score ≥ 5 categorized into the acute-valproate responder (acute-VPAR) group. We analyzed 158 peptides (92 proteins) from peripheral blood samples using multiple reaction monitoring mass spectrometry, and proteomic result-guided candidate gene association analyses, with 1,627 single nucleotide variants (SNVs), were performed using the Korean chip. RESULTS: The markers of 37 peptides (27 protein) showed temporal upregulation, indicating possible association with response to valproate treatment. A total of 58 SNVs in 22 genes and 37 SNVs in 16 genes showed nominally significant associations with the Alda A continuous score and the acute-VPAR group, respectively. No SNVs reached the genome-wide significance threshold; however, three SNVs (rs115788299, rs11563197, and rs117669164) in the secreted phosphoprotein 2 gene reached a gene-based false discovery rate-corrected significance threshold with response to valproate treatment. Significant markers were associated with the pathophysiological processes of bipolar disorders, including the immune response, acute phase reaction, and coagulation cascade. These results suggest that valproate effectively suppresses mechanisms associated with disease progression. CONCLUSIONS: The markers identified in this study could be valuable indicators of the underlying mechanisms associated with response to valproate treatment.

Development of Prediction Model for Intensive Care Unit Admission Based on Heart Rate Variability: A Case-Control Matched Analysis.

This study aimed to develop a predictive model for intensive care unit (ICU) admission by using heart rate variability (HRV) data. This retrospective case-control study used two datasets (emergency department [ED] patients admitted to the ICU, and patients in the operating room without ICU admission) from a single academic tertiary hospital. HRV metrics were measured every 5 min using R-peak-to-R-peak (R-R) intervals. We developed a generalized linear mixed model to predict ICU admission and assessed the area under the receiver operating characteristic curve (AUC). Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated from the coefficients. We analyzed 610 (ICU: 122; non-ICU: 488) patients, and the factors influencing the odds of ICU admission included a history of diabetes mellitus (OR [95% CI]: 3.33 [1.71-6.48]); a higher heart rate (OR [95% CI]: 3.40 [2.97-3.90] per 10-unit increase); a higher root mean square of successive R-R interval differences (RMSSD; OR [95% CI]: 1.36 [1.22-1.51] per 10-unit increase); and a lower standard deviation of R-R intervals (SDRR; OR [95% CI], 0.68 [0.60-0.78] per 10-unit increase). The final model achieved an AUC of 0.947 (95% CI: 0.906-0.987). The developed model effectively predicted ICU admission among a mixed population from the ED and operating room.

L-DOPA regulates neuroinflammation and Aß pathology through NEP and ADAM17 in a mouse model of AD.

Dopamine plays important roles in cognitive function and inflammation and therefore is involved in the pathogenesis of neurodegenerative diseases, including Alzheimer's disease (AD). Drugs that increase or maintain dopamine levels in the brain could be a therapeutic strategy for AD. However, the effects of dopamine and its precursor levodopa (L-DOPA) on Aß/tau pathology in vivo and the underlying molecular mechanisms have not been studied in detail. Here, we investigated whether L-DOPA treatment alters neuroinflammation, Aß pathology, and tau phosphorylation in 5xFAD mice, a model of AD. We found that L-DOPA administration significantly reduced microgliosis and astrogliosis in 5xFAD mice. In addition, L-DOPA treatment significantly decreased Aß plaque number by upregulating NEP and ADAM17 levels in 5xFAD mice. However, L-DOPA-treated 5xFAD mice did not exhibit changes in tau hyperphosphorylation or tau kinase levels. These data suggest that L-DOPA alleviates neuroinflammatory responses and Aß pathology but not tau pathology in this mouse model of AD.

The characteristics and clinical outcomes of a pluripotent high-risk group with the potential to develop a diverse range of psychiatric disorders.

BACKGROUND: Recent studies have indicated that clinical high risk for psychosis (CHR-P) is highly specific for psychotic disorders other than pluripotential to various serious mental illnesses. However, not all CHR-P develop psychotic disorder only, and psychosis can occur in non-psychotic disorders as well. Our prospective cohort study aims to investigate the characteristics and clinical outcomes of a pluripotent high-risk group with the potential to develop a diverse range of psychiatric disorders. METHODS: The SPRIM study is a prospective naturalistic cohort program that focuses on the early detection of those at risk of developing serious mental illness, including psychosis (CHR-P), bipolar (CHR-B), and depressive disorder (CHR-D), as well as undifferentiated risk participants (UCHR). Our study has a longitudinal design with a baseline assessment and eight follow-up evaluations at 6, 12, 18, 24, 30, 36, 42, and 48 months to determine whether participants have transitioned to psychosis or mood disorders. RESULTS: The SPRIM sample consisted of 90 CHR participants. The total cumulative incidence rate of transition was 53.3% (95% CI 32.5-77.2). CHR-P, CHR-B, CHR-D, and UCHR had cumulative incidence rates of 13.7% (95% CI 3.4-46.4), 52.4% (95% CI 28.1-81.1), 66.7% (95% CI 24.6-98.6) and 54.3% (95% CI 20.5-93.1), respectively. The cumulative incidence of psychosis, bipolar, and depressive disorder among all participants was 3.3% (95% CI 0.8-11.5), 45.7% (95% CI 24.4-73.6), and 11.2% (95% CI 3.1-36.2), respectively. CONCLUSIONS: Our study suggests that the concept of pluripotent high-risk for a diverse range of psychiatric disorders is an integrative approach to examining transdiagnostic interactions between illnesses with a high transition rate and minimizing stigma.

Individual-specific postural discomfort prediction using decision tree models.

The objective of the current study was to explore the utilization of the decision tree (DT) algorithm to model posture-discomfort relationships at the individual level. The DT algorithm has the advantage that it makes no assumptions about the distribution of data, is robust in representing non-linear data with noise, and produces white-box models that are interpretable. Individual-level modelling is essential for examining individual-specific postural discomfort perception processes and understanding the inter-individual variability. It also has practical applications, including the development of individual-specific digital human models and more precise and informative population accommodation analysis. Individual-specific DT models were generated using postural discomfort rating data for various seated upper body postures to predict discomfort based on postural and task variables. The individual-specific DT models accurately predicted postural discomfort and revealed large inter-individual variability in the modelling results. DT modelling is expected to greatly facilitate investigating the human discomfort perception process.

Developing theragnostics for Alzheimer's disease: Insights from cancer treatment.

The prevalence of Alzheimer's disease (AD) and its associated economic and societal burdens are on the rise, but there are no curative treatments for AD. Interestingly, this neurodegenerative disease shares several biological and pathophysiological features with cancer, including cell-cycle dysregulation, angiogenesis, mitochondrial dysfunction, protein misfolding, and DNA damage. However, the genetic factors contributing to the overlap in biological processes between cancer and AD have not been actively studied. In this review, we discuss the shared biological features of cancer and AD, the molecular targets of anticancer drugs, and therapeutic approaches. First, we outline the common biological features of cancer and AD. Second, we describe several anticancer drugs, their molecular targets, and their effects on AD pathology. Finally, we discuss how protein-protein interactions (PPIs), receptor inhibition, immunotherapy, and gene therapy can be exploited for the cure and management of both cancer and AD. Collectively, this review provides insights for the development of AD theragnostics based on cancer drugs and molecular targets.

Efficacy of Learning Disorder Treatment for Reading or Mathematics Disorders: An Open Study.

Objectives: This study aimed to identify the effectiveness of treatment programs for children with reading (RD) or mathematics disorders (MD). Structured treatment programs were developed to improve phonological awareness and number sense among children and adolescents with RD or MD, respectively, and the effectiveness of the learning disorder treatment programs were evaluated. Methods: We used standardized, objective diagnostic, and evaluation tools not only to recruit participants with RD, MD, or comorbid attention deficit and hyperactivity disorder, but also to assess the effectiveness of the treatments regarding both improved core neurocognitive deficits of RD or MD and academic achievement. Forty children with RD or MD received one-on-one treatments from therapists. Results: In the RD group, treatment effects were observed in all subtests. In the word and paragraph reading tests, the accuracy rates and fluency improved. The results of the phonological working memory test, word-sound correspondence test, and rapid automatic naming tests also improved. In the MD group, the accuracy rate and fluency on the arithmetic test improved. An increase in the accuracy rate in the size and distance comparison tests and a decrease in the error rate in the estimation test were also observed. However, there were no improvements in reaction time in these subtests. Conclusion: Learning disorder treatment programs that focus on improving phonological awareness or number sense in children with RD or MD improved achievement, phonological awareness, and number sense.

Effect of information provision on parental intention toward COVID-19 vaccination for children: a nationwide survey experiment.

The reluctance of parents to vaccinate their children against COVID-19 was prevalent particularly when uncertainty over vaccination outcomes prevailed. We conducted a nationwide randomized online survey experiment to assess the effect of information provision on parental intention for COVID-19 vaccination before the government started vaccination for children in South Korea. Parents of elementary school children were provided with either no information (Control), information on vaccine profile (vaccine informed group; VI), or COVID-19 (disease informed group; DI). Among 359,110 participants, parental intention for vaccination of children was significantly higher in both VI and DI groups compared with the Control group. In terms of effect size, information on COVID-19 vaccine increased likelihood to vaccinate by 1620 per 100,000 parents and reduced vaccine hesitancy by 1340 per 100,000 parents. Consistently with the positive effect on vaccination intention, both VI and DI interventions increased participants' perceptions on vaccination benefits being higher than its risks and vaccination risks being lower than health risks of COVID-19 infection, and self-reported trust in COVID-19 information. Our results lend strong support to the claim that the provision of targeted, tailored information on COVID-19 vaccine and infection increases parental intention to vaccinate children and reduces vaccine hesitancy.

PANCDR: precise medicine prediction using an adversarial network for cancer drug response.

Pharmacogenomics aims to provide personalized therapy to patients based on their genetic variability. However, accurate prediction of cancer drug response (CDR) is challenging due to genetic heterogeneity. Since clinical data are limited, most studies predicting drug response use preclinical data to train models. However, such models might not be generalizable to external clinical data due to differences between the preclinical and clinical datasets. In this study, a Precision Medicine Prediction using an Adversarial Network for Cancer Drug Response (PANCDR) model is proposed. PANCDR consists of two sub-models, an adversarial model and a CDR prediction model. The adversarial model reduces the gap between the preclinical and clinical datasets, while the CDR prediction model extracts features and predicts responses. PANCDR was trained using both preclinical data and unlabeled clinical data. Subsequently, it was tested on external clinical data, including The Cancer Genome Atlas and brain tumor patients. PANCDR outperformed other machine learning models in predicting external test data. Our results demonstrate the robustness of PANCDR and its potential in precision medicine by recommending patient-specific drug candidates. The PANCDR codes and data are available at https://github.com/DMCB-GIST/PANCDR.

Changes in the intrinsic severity of severe acute respiratory syndrome coronavirus 2 according to the emerging variant: a nationwide study from February 2020 to June 2022, including comparison with vaccinated populations.

BACKGROUND: As the population acquires immunity through vaccination and natural infection of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), understanding the intrinsic severity of coronavirus disease (COVID-19) is becoming challenging. We aimed to evaluate the intrinsic severity regarding circulating variants of SARS-CoV-2 and to compare this between vaccinated and unvaccinated individuals. METHODS: With unvaccinated and initially infected confirmed cases of COVID-19, we estimated the case severity rate (CSR); case fatality rate (CFR); and mortality rate (MR), including severe/critical cases and deaths, stratified by age and compared by vaccination status according to the period regarding the variants of COVID-19 and vaccination. The overall rate was directly standardized with age. RESULTS: The age-standardized CSRs (aCSRs) of the unvaccinated group were 2.12%, 5.51%, and 0.94% in the pre-delta, delta, and omicron period, respectively, and the age-standardized CFRs (aCFRs) were 0.60%, 2.49%, and 0.63% in each period, respectively. The complete vaccination group had lower severity than the unvaccinated group over the entire period showing under 1% for the aCSR and 0.5% for the aCFR. The age-standardized MR of the unvaccinated group was 448 per million people per month people in the omicron period, which was 11 times higher than that of the vaccinated group. In terms of age groups, the CSR and CFR sharply increased with age from the 60 s and showed lower risk reduction in the 80 s when the period changed to the omicron period. CONCLUSIONS: The intrinsic severity of COVID-19 was the highest in the delta period, with over 5% for the aCSR, whereas the completely vaccinated group maintained below 1%. This implies that when the population is vaccinated, the impact of COVID-19 will be limited, even if a new mutation appears. Moreover, considering the decreasing intrinsic severity, the response to COVID-19 should prioritize older individuals at a higher risk of severe disease.

Staphylococcus argenteus bacteremia in the Republic of Korea.

In 2015, Staphylococcus argenteus and Staphylococcus schweitzeri were proposed as new species, distinct from Staphylococcus aureus and collectively referred to as the S. aureus complex. However, no clinical reports of these new species exist in Korea. Upon the application of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) for all bloodstream isolates since September 2022, S. argenteus was identified in one patient. Therefore, we aimed to search for new species among the archives of the S. aureus bacteremia cohort and describe their clinical and microbiological characteristics. Among the 691 archived S. aureus isolates between 2012 and 2018, one was identified as S. argenteus via MALDI-TOF MS. Both S. argenteus isolates (one in 2022) were obtained from patients with extensive pneumonia accompanied by bacteremia and both cases had fatal outcomes. They harbored multiple virulence genes (clfA, clfB, fnbpA, sdrC, sdrD, sdrE, bbp, cna, see, seg, sei, blaZ, fnbpB, and map) but did not harbor mecA and pvl. No matched sequence type (ST) was found in either isolate, and both S. argenteus isolates were closely related to ST1594, ST1593, ST1793, and ST1303, which belonged to S. argenteus. S. argenteus accounted for <1% of the S. aureus complex but had clinical characteristics similar to S. aureus. Therefore, clinicians should be aware of these factors to avoid misidentifying these strains as coagulase-negative staphylococci, and appropriate reporting is required to minimize confusion.IMPORTANCEStaphylococcus argenteus, a member of Staphylococcus aureus complex, has been reported as an important pathogen that causes clinically invasive infections in humans similar to S. aureus. Clinical isolates of S. argenteus have been reported across the world, showing a large geographical difference in prevalence and genomic profile. However, there have been no clinical reports regarding this new species in Korea. This is the first report to investigate the clinical and genetic characteristics of S. argenteus identified in patients with bacteremia, and the proportion of S. argenteus bacteremia among S. aureus bacteremia cohort in Korea.

Integrated drug response prediction models pinpoint repurposed drugs with effectiveness against rhabdomyosarcoma.

Targeted therapies for inhibiting the growth of cancer cells or inducing apoptosis are urgently needed for effective rhabdomyosarcoma (RMS) treatment. However, identifying cancer-targeting compounds with few side effects, among the many potential compounds, is expensive and time-consuming. A computational approach to reduce the number of potential candidate drugs can facilitate the discovery of attractive lead compounds. To address this and obtain reliable predictions of novel cell-line-specific drugs, we apply prediction models that have the potential to improve drug discovery approaches for RMS treatment. The results of two prediction models were ensemble and validated via in vitro experiments. The computational models were trained using data extracted from the Genomics of Drug Sensitivity in Cancer database and tested on two RMS cell lines to select potential RMS drug candidates. Among 235 candidate drugs, 22 were selected following the result of the computational approach, and three candidate drugs were identified (NSC207895, vorinostat, and belinostat) that showed selective effectiveness in RMS cell lines in vitro via the induction of apoptosis. Our in vitro experiments have demonstrated that our proposed methods can effectively identify and repurpose drugs for treating RMS.

Risk Factors for Severe and Critical Coronavirus Disease 2019 in Children.

BACKGROUND: Coronavirus disease 2019 (COVID-19) is generally mild in children; however, severe or critical cases may occur. In this nationwide study, we analyzed clinical manifestations in children diagnosed with severe acute respiratory syndrome coronavirus 2 to identify high-risk groups for severe or critical disease and compared the clinical features between the Delta- and Omicron-dominant periods. METHODS: Data were retrieved from the National Health Insurance Service (NHIS) database and merged with the Korea Disease Control and Prevention Agency-COVID-19-NHIS cohort, which includes information on COVID-19 cases and vaccination records. We included individuals <20 years old diagnosed with COVID-19 during both periods (Delta: July 25, 2021-January 15, 2022; Omicron: January 16, 2022-March 31, 2022). RESULTS: Proportion of severe or critical cases was higher during the Delta period than during the Omicron period. The Omicron period saw increased hospitalization for pneumonia and croup and increased likelihood of hospitalization for neurological manifestations. The risk of severe COVID-19 depended on age group (Delta: highest for 12-19 years; Omicron: 0-4 years). This risk was high in children with multiple complex chronic conditions during both periods and with obesity or asthma during the Delta but not during the Omicron period. Two-dose COVID-19 vaccination provided strong protection against severe disease in the Delta period (adjusted odds ratio: 0.20), with reduced effectiveness in the Omicron period (adjusted odds ratio: 0.91). However, it significantly reduced the risk of critical illness (adjusted odds ratio: 0.14). CONCLUSIONS: These findings can facilitate identification of children at high risk of severe or critical COVID-19, who may require intensive medical support, and development of vaccination policies.

Proteomic profiling in the progression of psychosis: Analysis of clinical high-risk, first episode psychosis, and healthy controls.

BACKGROUND AND HYPOTHESIS: Recent evidence has highlighted the benefits of early detection and treatment for better clinical outcomes in patients with psychosis. Biological markers of the disease have become a focal point of research. This study aimed to identify protein markers detectable in the early stages of psychosis and indicators of progression by comparing them with those of healthy controls (HC) and first episode psychosis (FEP). STUDY DESIGN: The participants comprised 28 patients in the clinical high-risk (CHR) group, 49 patients with FEP, and 61 HCs aged 15-35 years. Blood samples were collected and analyzed using multiple reaction monitoring-mass spectrometry to measure the expression of 158 peptide targets. Data were adjusted for age, sex, and use of psychotropic drugs. STUDY RESULTS: A total of 18 peptides (17 proteins) differed significantly among the groups. The protein PRDX2 was higher in the FEP group than in the CHR and HC groups and showed increased expression according to disease progression. The levels of six proteins were significantly higher in the FEP group than in the CHR group. Nine proteins differed significantly in the CHR group compared to the other groups. Sixteen proteins were significantly correlated with symptom severity. These proteins are primarily related to the coagulation cascade, inflammatory response, brain structure, and synaptic plasticity. CONCLUSIONS: Our findings suggested that peripheral protein markers reflect disease progression in patients with psychosis. Further longitudinal research is needed to confirm these findings and to identify the specific roles of these markers in the pathogenesis of schizophrenia.

Molecular data representation based on gene embeddings for cancer drug response prediction.

Cancer drug response prediction is a crucial task in precision medicine, but existing models have limitations in effectively representing molecular profiles of cancer cells. Specifically, when these models represent molecular omics data such as gene expression, they employ a one-hot encoding-based approach, where a fixed gene set is selected for all samples and omics data values are assigned to specific positions in a vector. However, this approach restricts the utilization of embedding-vector-based methods, such as attention-based models, and limits the flexibility of gene selection. To address these issues, our study proposes gene embedding-based fully connected neural networks (GEN) that utilizes gene embedding vectors as input data for cancer drug response prediction. The GEN allows for the use of embedding-vector-based architectures and different gene sets for each sample, providing enhanced flexibility. To validate the efficacy of GEN, we conducted experiments on three cancer drug response datasets. Our results demonstrate that GEN outperforms other recently developed methods in cancer drug prediction tasks and offers improved gene representation capabilities. All source codes are available at https://github.com/DMCB-GIST/GEN/ .