A5001438573- Schizophrenia research and treatment
- Gene expression and cancer classification
- Mental Health Research Topics
- Advanced Causal Inference Techniques
- Bioinformatics and Genomic Networks
- Personality Disorders and Psychopathology
- Statistical Methods and Inference
- Genetic Associations and Epidemiology
- Mental Health and Psychiatry
- Stress Responses and Cortisol
- Machine Learning in Bioinformatics
- Computational Drug Discovery Methods
- Statistical Methods and Bayesian Inference
- Traditional Chinese Medicine Analysis
- Cancer-related molecular mechanisms research
- Sleep and related disorders
- AI in cancer detection
- Advanced biosensing and bioanalysis techniques
- Heart Rate Variability and Autonomic Control
- Tryptophan and brain disorders
- Traumatic Brain Injury Research
- Resilience and Mental Health
- Algorithms and Data Compression
- Biomedical Text Mining and Ontologies
- Cell Image Analysis Techniques
Stony Brook University
2016-2024
Icahn School of Medicine at Mount Sinai
2017
Structural models of psychopathology provide dimensional alternatives to traditional categorical classification systems. Competing models, such as the bifactor and correlated factors are typically compared via statistical indices assess how well each model fits same data. However, simulation studies have found evidence for probifactor fit index bias in several psychological research domains. The present study sought extend this psychopathology, wherein has received much attention, but its...
Kraepelin considered declining course a hallmark of schizophrenia, but others have suggested that outcomes usually stabilize or improve after treatment initiation. The authors investigated this question in an epidemiologically defined cohort with psychotic disorders followed for 20 years first hospitalization.
Abstract Understanding whether and how the schizophrenia polygenic risk score (SZ PRS) predicts course of illness could improve diagnosis prognostication in psychotic disorders. We tested SZ PRS symptoms, cognition, severity, diagnostic changes over 20 years following first admission. The Suffolk County Mental Health Project is an inception cohort study first-admission patients with psychosis. Patients were assessed six times years, 249 provided DNA. Geographically- demographically-matched...
Heterogeneity of psychosis presents significant challenges for classification. Between 2 and 12 symptom dimensions have been proposed, consensus is lacking. The present study sought to identify uniquely informative models by comparing the validity these alternatives. An epidemiologic cohort 628 first-admission inpatients with was interviewed 6 times over decades completed an electrophysiological assessment error processing at year 20. We first analyzed a comprehensive set 49 symptoms rated...
The associations among normal personality and many mental disorders are well established, but it remains unclear whether how symptoms of schizophrenia schizotypal traits align with the taxonomy. This study examined joint factor structure personality, schizotypy, in people psychotic (n = 288) never-psychotic adults 257) Suffolk County Mental Health Project. First, we evaluated (positive negative mistrust) traits. In both psychotic-disorder groups, best-fitting model had 5 factors:...
Advances in medical technology have allowed for customized prognosis, diagnosis, and treatment regimens that utilize multiple heterogeneous data sources. Multiple kernel learning (MKL) is well suited the integration of high throughput MKL remains to be under-utilized by genomic researchers partly due lack unified guidelines its use, benchmark datasets. We provide three implementations R. These methods are applied simulated illustrate can select appropriate models. also apply combine clinical...
Abstract Motivation A gradient boosting decision tree (GBDT) is a powerful ensemble machine-learning method that has the potential to accelerate biomarker discovery from high-dimensional molecular data. Recent algorithmic advances, such as extreme (XGB) and light (LGB), have rendered GBDT training more efficient, scalable accurate. However, these modern techniques not yet been widely adopted in discovering biomarkers for censored survival outcomes, which are key clinical outcomes or...
CRISPR is a versatile gene editing tool which has revolutionized genetic research in the past few years. Optimizing sgRNA design to improve efficiency of target/DNA cleavage critical ensure success screens.By borrowing knowledge from oligonucleotide and nucleosome occupancy models, we systematically evaluated candidate features computed number nucleic acid, thermodynamic secondary structure models on real datasets. Our results showed that taking into account position-dependent dinucleotide...
The current study aimed to evaluate the efficacy and safety of Compound Danshen Dripping Pills (CDDP) in improving cardiac function among patients with acute anterior ST-segment elevation myocardial infarction (AAMI). Between February 2021 2023, 247 eligible AAMI after primary percutaneous coronary intervention (pPCI) were enrolled randomly assigned (1∶1) receive CDDP (<i>n </i>= 126) or placebo 121), a follow-up 48 weeks. Compared group, group demonstrated significant increase left...
Personality is a major predictor of many mental and physical disorders, but its contributions to illness course are understudied. The current study aimed explore whether personality associated with psychiatric medical over 10 years following trauma. World Trade Center (WTC) responders (N = 532) completed the inventory for DSM-5, which measures both broad domains narrow facets. Responders' health was assessed in decade WTC disaster during annual monitoring visits at Health Program clinic....
The Cox proportional hazard model is one of the most widely used methods in modeling time-to-event data health sciences. Due to simplicity partial likelihood function, many machine learning algorithms use it for survival data. However, due nature censored data, optimization problem becomes intractable when more complicated regularization employed, which necessary dealing with high dimensional omic In this paper, we show that a convex conjugate function loss based on Fenchel duality exists,...
Abstract Background Risk factors for depressive disorders (DD) change substantially over time, but the prognostic value of these changes remains unclear. Two basic types dynamic effects are possible. The ‘Risk Escalation hypothesis’ posits that worsening risk levels predicts DD onset above average level factors. Alternatively, ‘Chronic rather than first-onset DD. Methods We utilized data from ADEPT project, a cohort 496 girls (baseline age 13.5–15.5 years) community followed 3 years....
Abstract Advances in medical technology have allowed for customized prognosis, diagnosis, and personalized treatment regimens that utilize multiple heterogeneous data sources. Multiple kernel learning (MKL) is well suited integration of high throughput sources, however, there are currently no implementations MKL R. In this paper, we give some background material support vector machine (SVM) introduce an R package, RMKL, which provides C++ code to implement several algorithms classification...
Abstract High dimensional genomics data in biomedical sciences is an invaluable resource for constructing statistical prediction models. With the increasing knowledge of gene networks and pathways, such information can be utilized models to improve accuracy enhance model interpretability. However, certain scenarios network structure may only partially known or subject inaccuracy. Thus, performance incorporating compromised. In this paper, we propose a weighted sparse learning method by...
Abstract The search for prognostic biomarkers capable of predicting patient outcomes, by analyzing gene expression in tissue samples and other molecular profiles, remains largely on single-gene-based or global-gene-search approaches. Gene-centric approaches, while foundational, fail to capture the higher-order dependencies that reflect activities co-regulated processes, pathway alterations, regulatory networks, all which are crucial determining outcomes complex diseases like cancer. Here, we...
Abstract Understanding whether and how the schizophrenia polygenic risk score (SZ PRS) predicts course of illness could improve diagnostics prognostication in psychotic disorders. We tested SZ PRS symptoms, cognition, severity, diagnostic changes over 20 years following first admission. The Suffolk County Mental Health Project is an inception cohort study first-admission patients with psychosis. Patients were assessed six times years, 249 provided DNA. Geographically- demographically-matched...
Nonparametric tests are popular and efficient methods of detecting jumps in high- frequency financial data. Each method has its own advantages disadvantages, their performances may be affected by underlying noise dynamic structures. To address this, we proposed a robust p-value pooling that aims to combine the each method. We focus on model validation within Monte Carlo framework assess reproducibility false discovery rate (FDR). Reproducible analyses via correspondence curve an...
Abstract The Cox proportional hazard model is the most widely used method in modeling time-to-event data health sciences. A common form of loss function machine learning for survival also mainly based on partial likelihood function, due to its simplicity. However, optimization problem becomes intractable when more complicated regularization employed with function. In this paper, we show that a convex conjugate Fenchel Duality exists, and provides an alternative framework primal form....
Abstract Motivation Gradient boosting decision tree (GBDT) is a powerful ensemble machine learning method that has the potential to accelerate biomarker discovery from high-dimensional molecular data. Recent algorithmic advances, such as Extreme Boosting (XGB) and Light (LGB), have rendered GBDT training more efficient, scalable accurate. These modern techniques, however, not yet been widely adopted in biomarkers based on patient survival data, which are key clinical outcomes or endpoints...