YFYuan Fang

Yuan Fang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Yuan Fang has a total of 49 co-authors. Here are the top co-authors by number of shared publications:

  • Jesse Mez

    Boston University Alzheimer’s Disease Research Center and CTE Center, Boston University Chobanian and Avedisian School of Medicine, Boston, MA, United States

    6 shared publications
  • Claudia L. Satizabal

    Department of Neurology, Boston University Chobanian and Avedisian School of Medicine, Boston, MA, United States

    5 shared publications
  • Joanne M. Murabito

    Framingham Heart Study, National Heart, Lung, and Blood Institute and Boston University Chobanian and Avedisian School of Medicine, Framingham, MA, United States

    5 shared publications
  • Kathryn L. Lunetta

    Department of Biostatistics, Boston University School of Public Health, Boston, MA, United States

    5 shared publications
  • Margaret F. Doyle

    Department of Pathology and Laboratory Medicine, Larner College of Medicine, University of Vermont, Burlington, VT, United States

    5 shared publications
  • Michael L. Alosco

    Boston University Alzheimer’s Disease Research Center and CTE Center, Boston University Chobanian and Avedisian School of Medicine, Boston, MA, United States

    5 shared publications

Latest Publications

  • Open Access

    Logistic-t multinomial mixture model for clustering for microbiome data

    by Wenshu Dai, Yuan Fang, Sanjeena Subedi

    The logistic-normal multinomial distribution has been used for modelling microbiome data obtained from high-throughput sequencing technologies, which are compositional in nature. A logistic-normal multinomial distribution is a hierarchical multinomial distribution that assumes the latent variable which are the additive log-ratio (ALR) transformed proportions in a multinomial distribution follows a Gaussian distribution. Model-based clustering algorithms have also been developed for clustering microbiome data based on the logistic-normal models. However, the Gaussian assumption may violated when the ALR transformed variable exhibit heavy-tailed distributions or has outliers. Our study introduces a novel mixture of logistic- t multinomial models that effectively address these challenges. Utilizing a multivariate t distribution for ALR transformed latent variables, the proposed approach provides the flexibility to capture heavy tails and thus better accommodates the outliers when clustering microbiome compositional data. We incorporate a variational Expectation-Maximization (EM) algorithm to facilitate efficient parameter estimation for intractable posterior distributions. Our model demonstrates competitive performance in terms of clustering accuracy and parameter recovery as compared to existing approaches through simulation studies and real data analysis, hence offering a robust tool for exploring the complex structures and heterogeneity of microbiome compositional data.

  • Open Access

    Modeling heterogeneity in cognitive trajectories in the Framingham Heart Study

    by Yuan Fang, Jiachen Chen, Jesse Mez, Claudia L. Satizabal, Michael L. Alosco, Wei Qiao Qiu, Margaret F. Doyle, Joanne M. Murabito, Kathryn L. Lunetta

    Introduction: The prevalence of cognitive impairment in the population is growing; however, there is substantial heterogeneity in the rate of decline across different cognitive domains. Harmonized factor scores measuring memory, executive function, and language domains have been created in the Framingham Heart Study (FHS).Methods: This work identified FHS participants with two or more repeated factor scores after age 60 and fitted latent class mixed models (LCMM) to cluster cognitive trajectories within each domain. Non-linear shapes of trajectories were modeled piecewise linearly, followed by stepwise selections to select cluster-specific change points.Results: We identified different latent classes of participants with early cognitive decline, compared to late decliners, for each domain. Ten-fold cross-validation yielded stable subgroupings. Our findings show latent-class-related differential patterns in cognitive aging in the FHS. We also investigated the association between identified latent classes with existing protein biomarkers of cognitive aging in a subsample of the study and found elevated levels of CD40L and CD14 were associated with a higher risk of early decline in memory and executive function domain, respectively.Discussion: In summary, our study advances the understanding of cognitive decline heterogeneity among FHS participants and sets the stage for further investigations into early intervention strategies and personalized approaches to mitigate cognitive aging risks.


  • Open Access

    The Association of Circulating Immune Cells With Cognitive Function, Brain Imaging, and Incident All-cause and Alzheimer’s Dementia: The Framingham Offspring Study

    by Yumeng Cao, Lindsay R Salvati, Jiachen Chen, Ahmed Ragab, Jesse Mez, Claudia L Satizabal, Michael L Alosco, Yuan Fang, Wei Qiao Qiu, Kathryn L Lunetta, et al.

    Emerging evidence supports the central role of the immune system in brain health, yet little is known about the role of circulating immune cells and cognitive function or brain health in dementia-free populations. We investigated the association of 43 immune cells with cognitive function, structural brain imaging, and incident dementia in Framingham Heart Study Offspring participants. Immune cells were phenotyped by flow cytometry. Linear mixed effects models were used for cross-sectional associations between immune cells and 4 cognitive domain scores and 13 brain magnetic resonance imaging measurements. Cox proportional hazards regression models tested the relationship between immune cells and time to dementia. Models were adjusted for age, sex, education, cytomegalovirus status, and APOE genotype, with further adjustment for cardiovascular risk factors. Data was further stratified by cytomegalovirus status. Among 795 participants with cellular phenotyping, cognitive testing and brain imaging data (mean age 61, 52% women), there were no associations between immune cells and cognitive test scores. Several significant associations between immune cells and regional brain magnetic resonance imaging measurements were observed. Higher CD8+ cells [CD8+CD45RO−CCR7−CD27− (Teff), CD8+CD45RA+CD28−CD57+(TEMRA), CD8+CD27−CD28−] associated with greater cerebrum gray and frontal gray matter volumes and inclusion of cardiovascular risk factors strengthened the association. Among CMV+ participants, CD8+TEMRA and CD8+Teff cells were significantly associated with higher total gray and frontal gray matter volumes. No significant associations were observed between immune cells and incident all-cause or Alzheimer’s disease dementia. The pathobiology underpinning the associations between immune cells and brain volumes require further study and validation in diverse samples.

  • Open Access

    Association of DMD Gene Variant Classes With Motor Outcomes in a Drug Registration Clinical Trial Setting

    by Yuan Fang, Utkarsh J. Dang, Katherine I. Illei, Paula Clemens, Eric P. Hoffman

    Duchenne muscular dystrophy (DMD) is caused by pathogenic variants of the DMD gene, leading to the loss of dystrophin. Clinical variability in DMD can complicate interpretation of interventional data in clinical trials. One source of clinical variability is allelic heterogeneity (different pathogenic variants with different effects on dystrophin protein expression). We sought to determine whether gene variant classes in clinical trial participants potentially affect clinical trial data interpretation. We analyzed 186 vamorolone trial participants with DMD (VBP15-002/003; VBP15-004) who were 4 to <7 years old and steroid-naïve at baseline. We stratified participants into gene variant classes by either variant location in the gene affecting different gene promoters (5′ [Dp427-only] vs 3′ [Dp427+other isoforms]) or residual dystrophin levels (null vs possible non-null [5′ gene variants, exon 44 skippable, splice site]). We evaluated associations with baseline motor outcomes and treatment response (prednisone and vamorolone). Participants with variants in ex63 and downstream (null for Dp427+Dp140+Dp71 protein isoforms) showed poorer baseline motor outcomes for time to stand from supine velocity than those with variants in ex1-44 (Dp427 only). No significant baseline differences were found between likely null and possible non-null variants. Participants with only Dp427 involvement showed significantly better treatment response for the 6-minute walk distance. Most of the comparisons of baseline motor function and treatment response were similar between variant classes. The large variation in baseline motor function in young, steroid-naïve patients with DMD is only minimally explained by different gene variant classes. While there is strong literature support for 3′ variants leading to a more severe motor and cognitive DMD phenotype, we found this variant class under-represented in our clinical trials. This suggests that they may fail inclusion criteria (failure to follow commands; poor motor function). Subgroup analyses in DMD clinical trials at a young age range based on gene variant class may not reveal significant differences and would be relatively noninformative.


  • Open Access

    Genomic and Proteomic Analyses of Bacterial Communities of Ixodes scapularis Ticks from Broome County, New York

    by Michel Shamoon-Pour, Emily H. Canessa, John Macher, Amaan Fruitwala, Emma Draper, Benjamin Policriti, Matthew Chin, Matthew Nunez, Paul Puccio, Yuan Fang, et al.

    The microbial communities of Ixodes scapularis, the primary vector of Lyme disease in North America, exhibit regional variations that may affect pathogen transmission and vector competence. We analyzed bacterial communities in I. scapularis ticks collected from Broome County, New York, using 16S rRNA gene sequencing (18 ticks) as well as mass spectrometry-based proteomics (36 ticks). According to the 16S rRNA analysis, the endosymbiont Rickettsia buchneri was the most abundant species, with significantly higher (p = 0.0011) abundance in females (54.76%) compared to males (31.15%). We detected Borreliella burgdorferi in 44.44% of ticks and Anaplasma phagocytophilum in two nymphs but in high relative abundances (12.73% and 46.46%). Male ticks exhibited higher bacterial diversity, although the community composition showed no significant clustering by sex or life stage. Co-occurrence analysis revealed negative associations between R. buchneri and Pseudomonas (p = 0.0245), but no associations with B. burgdorferi. Proteomic analysis identified 12 R. buchneri-specific proteins, additionally detecting the protozoan pathogen Babesia microti in 18.18% of females. These findings provide the first comprehensive characterization of I. scapularis microbiomes in the Southern Tier region of New York and suggest broader distribution of R. buchneri across tick life stages than previously recognized, with potential implications for pathogen transmission dynamics.


  • Open Access

    Tools for clinical pharmacists to identify medications that can cause drug‐induced cognitive impairment: A systematic review

    by Jacob Stanton, Michael R. Gionfriddo, Gail B. Rattinger, Yuan Fang, Mia E. Lussier

    Cognitive impairment affects around 19% of community‐dwelling older adults in the United States. Pharmacists can play a significant role in identifying and deprescribing cognitive‐impairing medications through cognitive‐focused medication reviews or as part of comprehensive medication management. Multiple tools exist to help pharmacists and other practitioners identify these medications, yet a comprehensive description of these tools has not been undertaken. We searched EBSCO MEDLINE, PubMed, EMBASE, CINAHL, and PsycInfo from inception to August 4, 2023, using search terms relative to medications and cognition. Studies were included if they detailed a tool to assess the cognitive burden of a medication regimen. Abstracts and full texts were screened independently and in duplicate. Any disagreements at the level of full‐text screening were resolved by consensus. Eligible studies underwent data extraction by two authors independently and in duplicate using a data extraction form. Our search identified 13 126 articles of which only four tools were initially included. Additionally, we identified seven articles from the reference lists of excluded review articles for a final inclusion count of 11 tools. Articles were published between 2001 and 2017. Nine of 11 tools (82%) were focused solely on anticholinergic medications, while only one tool indicated a clinically relevant cutoff value and alternative medication recommendations. While there are many tools to assess the cognitive burden from medications, many are limited by the development date, focus on anticholinergic medications, or lack clinically useful information. Further research is needed to develop more comprehensive, up‐to‐date, and readily implementable tools that clinical pharmacists can utilize to assess medication‐induced cognitive risks. Such as tool could be applied by pharmacists in many settings to optimize medication use and improve patient outcomes.

  • Open Access

    AGING-RELATED IMMUNE CELLS PHENOTYPES AND ALL-CAUSE MORTALITY IN THE FRAMINGHAM HEART STUDY

    by Ahmed A Y Ragab, Margaret Doyle, Jiachen Chen, Yuan Fang, Kathryn Lunetta, Joanne Murabito

    Almost 2 billion people will likely be above the age of 60 by 2050. Aging is a process that is intrinsically complicated. One of the measures for immune system senescence is aging related immune cells phenotypes (ARIP)s. Besides CD8/CD4 ratio, new immunosenescence phenotypes have been proposed. The associations between ARIPs and all-cause mortality during long-term follow up is understudied. We profiled immune cells using flow cytometry and prospectively investigated ten different ARIPs, namely, CD4+CD27-, CD4+CD28-CD27-, CD8+CD27-, CD8+CD28-CD27-, CD4+/CD8+ ratio, CD4+(Tnaive/(T Central Memory (Tcm)+T Effector Memory (Tem)+T Effector (Teff)), CD8+ (Tnaive/(Tcm+Tem+Teff)), Granzyme B+CD8/Granzyme B+CD4 ratio, CD8+ Tc17/Treg ratio and CD4+ Th17/Treg ratio in relation to survival outcome among 990 dementia-free Framingham Heart Study (FHS) Offspring cohort participants who attended the seventh exam (1998-2001, mean age 62 years, range 40-88, 52% female). Cox proportional hazards regression models adjusting for age and sex with robust variance to account for family correlation were used to test for association between the ARIPs and hazard of death. During up to 20 years of follow up, the survival rate was 66%. Higher CD8+ Tc17/Treg ratio was significantly associated with better survival (HR:0.82 [0.7-0.94], p< 0.001). Higher CD4+ Th17/Treg ratio and CD4/CD8 ratio were also nominally associated with lower risk of death (HR:0.87, [0.75-0.97], p=0.01 and HR 0.9, [0.79-0.99], p=0.04, respectively). Other ARIPs were unassociated with all-cause mortality. We conclude that further investigation of the CD8+Tc17/Treg and CD4+ Th17/Treg ratios as ARIP biomarkers for risk of all-cause mortality is justified.

  • Open Access

    INFLAMMATORY BIOMARKERS ASSOCIATED WITH BRAIN MRI MEASURES: FRAMINGHAM HEART STUDY OFFSPRING COHORT

    by Jiachen Chen, Margaret Doyle, Yuan Fang, Ahmed A Y Ragab, Joanne Murabito, Kathryn Lunetta

    Brain MRI volumes measuring brain atrophy have been associated with risks for Alzheimer’s Disease (AD) and related dementia. Inflammatory biomarkers associated with brain MRI volumes may provide insight into the neuroinflammation associated with these diseases. The study aim is to identify circulating inflammatory biomarkers associated with total and regional brain MRI volumes in the Framingham Heart Study Offspring cohort. Participants (N=662, 52% female, mean age 62 years) free of dementia and stroke at the time of blood draw and who had MRI measures within five years were profiled using the OLINK Proteomics inflammation panel. Pairwise cross-sectional associations between 68 biomarkers and eight brain MRI volumes were investigated using linear mixed-effect models accommodating familial correlations and adjusting for covariates (age, age2, sex, time between blood draw and MRI measurement, age-sex interaction, and number of APOE ε2 and ε4 alleles), using FDR≤0.1 to declare significance. APOE genotype-stratified analyses were performed to explore effect modification. Higher levels of 8 proteins were significantly associated with smaller total brain volumes (TCBV), including CDCP1, HGF, IL6, IL8, MMP10, OPG, VEGFA, and 4E-BP1. Higher levels of SCF and TWEAK were significantly associated with larger TCBV, and higher levels of SCF were also associated with larger parietal gray matter volume. In APOE ε4 carriers, higher levels of IFNγ were associated with greater white matter hyperintensity volumes. Consistent with our findings, SCF has been shown to have neuroprotective effects in animal models. Further studies are needed to confirm these potential risk and protective factors and to elucidate mechanisms.