A5087090835- Bioinformatics and Genomic Networks
- Diabetes Management and Research
- Gene expression and cancer classification
- Bayesian Modeling and Causal Inference
- Statistical Methods and Inference
- Genetic Associations and Epidemiology
- Liver Disease Diagnosis and Treatment
- Diabetes and associated disorders
- Pancreatic function and diabetes
- Machine Learning in Bioinformatics
- Explainable Artificial Intelligence (XAI)
- Cardiovascular Function and Risk Factors
- Single-cell and spatial transcriptomics
- Cancer Genomics and Diagnostics
- Topological and Geometric Data Analysis
- Diabetes Treatment and Management
- Complex Network Analysis Techniques
- Time Series Analysis and Forecasting
- Bariatric Surgery and Outcomes
- Alzheimer's disease research and treatments
- Gaussian Processes and Bayesian Inference
- Clinical Laboratory Practices and Quality Control
- Adipokines, Inflammation, and Metabolic Diseases
- Machine Learning and Data Classification
- Advanced Clustering Algorithms Research
Center for Systems Biology
2022-2025
Massachusetts General Hospital
2021-2025
Harvard University
2021-2025
University of California, Los Angeles
2024
University of Genoa
2018-2021
Abstract Obesity is a major public health challenge. Glucagon-like peptide-1 receptor agonists (GLP1-RA) and bariatric surgery (BS) are effective weight loss interventions; however, the genetic factors influencing treatment response remain largely unexplored. Moreover, most previous studies have focused on race ethnicity rather than ancestry. Here we analyzed 10,960 individuals from 9 multiancestry biobank across 6 countries to assess impact of known loss. Between 12 months, GLP1-RA users...
OBJECTIVE To examine the accuracy of different periods continuous glucose monitoring (CGM), hemoglobin A1c (HbA1c), and their combination for estimating mean glycemia over 90 days (AG90). RESEARCH DESIGN AND METHODS We retrospectively studied 985 CGM with <10% missing data from 315 adults (86% whom had type 1 diabetes) paired HbA1c measurements. The impact red blood cell age as a proxy nonglycemic effects on was estimated using published theoretical models in comparison empirical...
Background: Uveal melanoma (UM), a rare cancer of the eye, is characterized by initiating mutations in genes G-protein subunit alpha Q (GNAQ), 11 (GNA11), cysteinyl leukotriene receptor 2 (CYSLTR2), and phospholipase C beta 4 (PLCB4) metastasis-promoting splicing factor 3B1 (SF3B1), serine arginine rich (SRSF2), BRCA1-associated protein 1 (BAP1). Here, we tested hypothesis that additional mutations, though occurring only few cases (“secondary drivers”), might influence tumor development....
Pregnancy alters hematologic state as measured by complete blood counts (CBC), but the longitudinal changes in CBC indices that define healthy pregnancies are not well established. Our objectives were (1) to gestational age-specific reference intervals for CBCs and their a large United States-based cohort (2) use these examine associations between extreme values risk of obstetric complications. Retrospective study including electronic health record-based discovery validation cohorts....
Artificial intelligence (AI) applied to single-cell data has the potential transform our understanding of biological systems by revealing patterns and mechanisms that simpler traditional methods miss. Here, we develop a general-purpose, interpretable AI pipeline consisting two deep learning models: Multi- Input Set Transformer++ (MIST) model for prediction FastShap interpretability. We apply this large set routine clinical containing measurements circulating red blood cells (RBC), white...
Electronic Health Records (EHR)-linked biobanks have emerged as promising tools for precision medicine, enabling the integration of clinical and molecular data individual risk assessment. Association studies performed in biobank can connect common genetic variation to phenotypes, such through use polygenic scores (PGS), which are starting utility aiding clinician decision making. However, while aggregate large amounts effectively studies, most employ various opt-in consent protocols, and, a...
In many applications of finance, biology and sociology, complex systems involve entities interacting with each other. These processes have the peculiarity evolving over time comprising latent factors, which influence system without being explicitly measured. this work we present variable time-varying graphical lasso (LTGL), a method for multivariate time-series modelling that considers hidden or unmeasurable factors. The estimation contribution factors is embedded in model produces both...
Abstract Genome–wide association studies (GWAS) have revealed a plethora of putative susceptibility genes for Alzheimer’s disease (AD), with the sole exception APOE gene unequivocally validated in independent study. Considering that etiology complex diseases like AD could depend on functional multiple interaction network, here we proposed an alternative GWAS analysis strategy based (i) multivariate methods and (ii) telescope approach, order to guarantee identification correlated variables,...
Objective To study the longitudinal effects of both glucocorticoids and tocilizumab, an interleukin‐6 receptor inhibitor, on hemoglobin A 1c (HbA ) levels during glucocorticoid tapering. Methods We analyzed patients with complete data from Giant Cell Arteritis Clinical Research Study (GiACTA) to investigate impact glycemic nonglycemic factors changes in HbA over 52‐week trial. cell arteritis (GCA) were randomized receive either tocilizumab or placebo addition glucocorticoids. used a...
Graphical models allow to describe the interplay among variables of a system through compact representation, suitable when relations evolve over time. For example, in biological setting, genes interact differently depending on external environmental or metabolic factors. To incorporate this dynamics viable strategy is estimate sequence temporally related graphs assuming similarity samples different time points. While adjacent points may direct analysis towards robust underlying graph,...
The increased availability of multivariate time-series asks for the development suitable methods able to holistically analyse them. To this aim, we propose a novel flexible method data-mining, forecasting and causal patterns detection that leverages coupling Hidden Markov Models Gaussian Graphical Models. Given non-stationary time-series, proposed simultaneously clusters time points while understanding probabilistic relationships among variables. clustering divides into stationary sub-groups...
More and more biologists bioinformaticians turn to machine learning analyze large amounts of data. In this context, it is crucial understand which the most suitable data analysis pipeline for achieving reliable results. This process may be challenging, due a variety factors, ones being type general goal (e.g., explorative or predictive). Life science sets require further consideration as they often contain measures with low signal-to-noise ratio, high-dimensional observations, relatively few...
Abstract The complete blood count is an important screening tool for healthy adults and the most commonly ordered test at periodic physical exams. However, results are usually interpreted relative to one-size-fits-all reference intervals, undermining goal of precision medicine tailor medical care needs individual patients based on their unique characteristics. Here we show that standard indices in have robust homeostatic setpoints patient-specific stable, with typical adult’s set 9...
<p dir="ltr">Objective. To examine the accuracy of different periods continuous glucose monitoring (CGM), hemoglobin A1c (HbA1c), and their combination for estimating mean glycemia over 90 days (AG90).</p><p dir="ltr">Research Design Methods. We retrospectively studied 985 90-day CGM with <10% missing data from 315 adults (86% type 1 diabetes) paired HbA1c measurements. The impact red blood cell age as a proxy non-glycemic effects on was estimated using published...
<p dir="ltr">Objective. To examine the accuracy of different periods continuous glucose monitoring (CGM), hemoglobin A1c (HbA1c), and their combination for estimating mean glycemia over 90 days (AG90).</p><p dir="ltr">Research Design Methods. We retrospectively studied 985 90-day CGM with <10% missing data from 315 adults (86% type 1 diabetes) paired HbA1c measurements. The impact red blood cell age as a proxy non-glycemic effects on was estimated using published...
Obesity is a significant public health concern. GLP-1 receptor agonists (GLP1-RA), predominantly in use as type 2 diabetes treatment, are promising pharmacological approach for weight loss, while bariatric surgery (BS) remains durable, but invasive, intervention. Despite observed heterogeneity loss effects, the genetic effects on from GLP1-RA and BS have not been extensively explored large sample sizes, most studies focused differences race ethnicity, rather than ancestry. We studied whether...
Abstract Background The complete blood count (CBC) is an important screening tool for healthy adults, commonly ordered at periodic physical exams. However, test results are typically interpreted using one-size-fits-all reference intervals (RIs), which don’t account the low index-of-individuality of most CBC markers. This undermines goal precision medicine - to tailor patient care based on individual characteristics. Methods Here we utilize a database &gt;50,000 outpatients with yearly...
Summary The purpose of this study is to identify a global and robust signature characterizing Alzheimer’s Disease (AD). Two public GWAS datasets were analyzed considering 3-fold kernel approach, based on SNPs, Genes Pathways analysis, two binary classifications tasks addressed: cases@controls APOE4 task. In the SNP ADNI-1 ADNI-2 datasets, chromosome 19 20 reached high classification accuracy. addition, functional characterization signatures found enriched same pathway (i.e., Neuroactive...
Multivariate time series analysis is becoming an integral part of data pipelines. Understanding the individual point connections between covariates as well how these change in non-trivial. To this aim, we propose a novel method that leverages on Hidden Markov Models and Gaussian Graphical -- Time Adaptive Model (TAGM). Our model generalization state-of-the-art methods for inference temporal graphical models, its formulation both aspects models providing better results than current methods....