Biology subjects

Barzilay, R.

Publications and source records attributed to Barzilay, R..

16 recordsLinked to original sources

bioRxiv2026

Neuroanatomical Subtypes of Callous-Unemotional Traits in a Community Sample of Youth

ImportanceCallous Unemotional (CU) traits are associated with significant clinical and neurophysiological heterogeneity that may affect treatment effectiveness. ObjectiveTo uncover neuroanatomical subtypes of CU traits using weakly-supervised machine learning and assess whether emerging subtypes differ on relevant clinical, temperamental, and environmental constructs. Design, Setting, and ParticipantsImaging data was from the longitudinal Adolescent, Brain, Cognitive Development (ABCD) Study. Participants were 9-10 years old at baseline (M=9.925, 69.9% male) and included 222 children with CU traits and 234 typically developing controls matched on age, sex and income. Main Outcomes and MeasuresThe weakly-supervised heterogeneity through discriminative analysis (HYDRA) model was trained on grey matter (GM) volumes from 84 regions of interest (ROIs) and tested for reproducibility using cross-validation and permutation testing. Derived subtypes were compared cross-sectionally and prospectively on relevant clinical, temperamental, and environmental measures and subsequent GM volume at 2-year follow up. ResultsHYDRA revealed an optimal 2-subtype solution within children with CU traits. Subtypes showed comparable levels of aggression that were significantly higher than typically developing controls. At the same time, subtype 1 had larger GM volume, fewer internalizing symptoms, and less adversity exposure, while subtype 2 was characterized by smaller GM volume, more internalizing symptoms, and more adversity exposure. Conclusions and RelevanceThis study provides evidence of neuroanatomically distinct subtypes of CU traits characterized by different clinical and etiological profiles, with implications for diagnosis and treatment.

neuroscience↗

bioRxiv2026

White matter reflects the childhood exposome

The childhood environment is critical for brain development. However, most neuroimaging studies examine individual environmental measures (e.g., socioeconomic status) or a limited set of exposures, obscuring how the combination of complex, real-world exposures jointly influence brain development. Here we investigated how white matter shape and tissue properties are linked to the childhood exposome, a multidimensional measure capturing over 300 environmental exposures. Using multi-shell diffusion MRI from 8,183 children (ages 9-10) in the ABCD study, we quantified microstructural and macrostructural properties across 62 person-specific white matter tracts. The exposome showed widespread and highly replicable associations with both white matter microstructure and macrostructure: more advantaged environments were associated with larger tract macrostructure and lower orientation dispersion. Principal component analysis revealed that the dominant axis of exposome-white matter covariation aligns with the cortical sensorimotor-association hierarchy, such that tracts spanning this hierarchy exhibit the strongest associations with the exposome. Multivariate models demonstrated that patterns of white matter features explained 25% of the variance in the exposome in unseen individuals. Notably, white matter-based prediction of cognition was markedly reduced after accounting for the exposome ([~]82% reduction in explained variance), indicating that brain-cognition associations overlap substantially with variance captured by the exposome. These findings generalized to independent data from the Healthy Brain Network (n=869), which differs substantially from ABCD in MRI acquisition, participant selection, and childhood environments. Together, these results suggest that white matter architecture strongly reflects the childhood environment.

neuroscience↗

bioRxiv2026

Associations between the environment, brain, mental health, and cognition across adolescence

Adolescence is a critical period for the development of the brain, cognition, and mental health, which are shaped by a wide range of environmental factors. In the present study, we analysed the Adolescent Brain and Cognitive Development (ABCD) dataset to examine how a range of proximal (e.g., socioeconomic status, familial circumstances) and distal (e.g., neighbourhood conditions, access to healthcare and education) environmental factors are associated with changes in centile-based measures of brain structure, and whether these brain differences subsequently mediate variations in cognition and mental health. We analysed these associations both at baseline (N = 6,911; 3,605 M, 3,606 F; mean age = 9.93) and longitudinally across three timepoints (N = 1,628; 879 M, 749 F; ages 8-15). At baseline, a more advantaged proximal and distal environment was associated with larger volumes across the whole brain relative to age- and sex-matched peers, which, in turn, mediated better mental health outcomes and cognitive performance. In the longitudinal analysis, the childhood environment predicted changes in brain structure across adolescence, and these structural changes predicted changes in mental health and cognition. The childhood environment also predicted cognitive but not mental health changes across adolescence, suggesting that these associations may already be established early in adolescence. These findings provide insight into how environmental and neural factors shape adolescent mental health and cognition, with potential implications for early intervention strategies aimed at promoting positive developmental outcomes.

neuroscience↗

bioRxiv2026

The PMADS Project: A Longitudinal Multimodal Cohort Study to Understand Risk for Perinatal Mood and Anxiety Disorders

BackgroundPerinatal mood and anxiety disorders (PMADs) are among the most common and consequential complications of pregnancy. The perinatal period is also characterized by profound hormonal fluctuations and large-scale brain plasticity. However, the mechanisms linking these neurobiological changes to psychiatric risk are poorly understood. Prospective, clinically informed studies are needed to identify quantitative biomarkers and clarify pathways linking perinatal neurobiology to PMADs risk. MethodsThis report describes the design of a prospective, longitudinal cohort study integrating multimodal neuroimaging, biofluid sampling, and deep clinical phenotyping to enable precision characterization of neurobiological trajectories of PMADs risk. Twenty-five individuals at elevated risk for PMADs will be recruited prior to conception and followed across six in-person timepoints spanning the menstrual cycle, pregnancy, and early postpartum, with additional remote follow-ups through the first postpartum year. Data collection includes high-resolution structural MRI, functional brain mapping using multi-echo resting-state fMRI, diffusion MRI, arterial spin labeling, ultra-high field MR-based techniques for measuring glutamate (GluCEST and 1HMRS), biofluid sampling, and comprehensive clinical, behavioral, and cognitive assessments. Structured clinical interviews assess categorical diagnoses while dimensional symptom measures capture heterogeneity and transdiagnostic features of perinatal psychopathology. Longitudinal analyses will model nonlinear trajectories of brain and symptom change across the perinatal period as well as evaluate whether preconception network features and menstrual cycle-related brain changes are associated with subsequent perinatal symptom emergence. DiscussionThis cohort study establishes a longitudinal, multimodal framework for investigating neurobiological changes across the transition to pregnancy in individuals at elevated risk for PMADs. By anchoring pregnancy-related brain changes to preconception and menstrual cycle-related variability within the same individuals, this study is designed to evaluate associations between preconception hormone sensitivity, pregnancy-induced neuroplasticity, and PMADs risk. The resulting dataset will provide a deeply phenotyped longitudinal resource for investigating brain-behavior relationships across the perinatal period. Findings are expected to inform future larger-scale studies aimed at advancing mechanistic understanding of PMADs, improving individualized risk stratification, and supporting development of personalized preventive and neuromodulatory interventions.

neuroscience↗

bioRxiv2026

Menarche onset is an inflection point for mental health and brain development

Menarche is a normative milestone of female puberty, yet its role in adolescent mental health and brain development remains poorly understood. Using longitudinal data from 5,016 females (7 annual visits, ages 10-16 years) in the Adolescent Brain Cognitive Development Study, we found that menarche onset functions as an inflection point for the development of internalizing symptoms and gross brain morphometry. The onset of menarche, largely independent of timing and socio-environmental factors, preceded a significant spike in internalizing symptoms, while altering the rate of ongoing structural brain development. Following menarche onset, individuals with faster declines in gray matter volume and surface area also had heightened internalizing symptoms. These findings suggest that menarche is not only a reproductive milestone but a neuroendocrine driver of adolescent brain and mental health trajectories. This normative and easily identifiable marker could define a critical window for mental health screenings with greater precision than current age-based guidelines.

neuroscience↗

bioRxiv2025

Conditional-longitudinal brain growth charts detect MRI changes with birth weight and psychopathology

ImportanceBrain maturation varies between individuals, particularly during dynamic developmental periods like adolescence. Directly assessing differences in longitudinal trajectories can reveal deviations from normative patterns. ObjectiveWe present novel conditional-longitudinal normative models that characterize variability in brain maturation. We utilize these models to examine whether differences in longitudinal trajectories are associated with birth weight (BW), gestational age (GA), and longitudinal psychopathology derived from behavioral assessments. DesignCross-sectional and conditional-longitudinal normative models were developed for brain volumes derived from the first two neuroimaging timepoints from the Adolescent Brain Cognitive Development (ABCD) Study. Conditional-longitudinal models index an individuals expected brain volume at follow-up conditioned on their baseline measurement. Models were fit with split-half cross-validation on demographically matched samples. SettingThe ABCD Study is a multi-site, population-based study ParticipantsParticipants were excluded based on imaging quality flags and missing data, leaving 10,830 at baseline and 7,262 at follow-up. ExposuresBW and GA were derived from parent-report questionnaires. General psychopathology scores were calculated using a bifactor model. Main Outcomes and MeasuresWe calculated cross-sectional and conditional-longitudinal centiles, respectively quantifying individual deviations in size and change between timepoints. Sensitivity analyses included covariates for parental income and education as well as current weight and height. ResultsThe sample was 10,830 at baseline (48.2% F,age 9-10y) and 7,262 at follow-up (46.6% F,age 11-13y). Conditional-longitudinal centiles were sensitive to individual differences in brain change between timepoints. Lower BW was associated with lower conditional-longitudinal centiles, suggesting larger decreases in brain volumes over time (27 regions pfdr<0.05, {beta}max=0.08). Lower conditional-longitudinal centiles were associated with greater increases in psychopathology scores, suggesting with increased psychopathology brain volumes show greater decrease (37 regions pfdr<0.05, {beta}max=0.06). Notably, changes in psychopathology were not related to brain size at either timepoint, indexed by cross-sectional centiles. Conclusions and RelevanceModels that capture individual-level deviations from expected growth trajectories, rather than static positions on a growth curve, are particularly informative for assessing developmental change. Novel conditional-longitudinal models address this gap in lifespan brain imaging. Using this framework, we demonstrate robust associations between individual trajectory deviations, perinatal adversity, and longitudinally assessed mental health symptoms. Condition-longitudinal models hold promise for applications across psychiatric neuroscience, from development to aging. Key Points QuestionHow do differences in brain maturation trajectories, quantified by novel conditional-longitudinal models, relate to perinatal factors and mental health in adolescence? FindingsIn this longitudinal analysis of neuroimaging data from the Adolescent Brain Cognitive Development (ABCD) Study, conditional-longitudinal normative models revealed that trajectories of brain maturation in adolescence are associated with birth weight, and with longitudinal changes in mental health. MeaningConditional longitudinal models detect inter-individual variability in brain maturation, which is related to both perinatal factors and concurrent changes in psychopathology.

neuroscience↗

bioRxiv2025

BoltzGen: Toward Universal Binder Design

We introduce BoltzGen, an all-atom generative model for designing proteins and peptides across all modalities to bind a wide range of biomolecular targets. BoltzGen builds strong structural reasoning capabilities about target-binder interactions into its generative design process. This is achieved by unifying design and structure prediction, resulting in a single model that also reaches state-of-the-art folding performance. BoltzGens generation process can be controlled with a flexible design specification language over covalent bonds, structure constraints, binding sites, and more. We experimentally validate these capabilities in eight diverse design campaigns with functional and affinity readouts across 26 targets. In our experiments, binder modalities span from nanobodies to disulfide-bonded peptides, and targets from disordered proteins to small molecules. In particular, we identify nanobody binders for novel targets with low similarity to proteins with already known bound structures. We release model weights, data, and both inference and training code at: https://github.com/HannesStark/boltzgen.

bioengineering↗

bioRxiv2025

Multivariate environmental exposures are reflected in whole-brain functional connectivity and cognition in youth

Each individuals complex, multidimensional environment, known as their "exposome", plays an essential role in shaping cognitive neurodevelopment. Understanding the mechanisms whereby childrens exposome influences their development is crucial to facilitate the design of interventions to foster positive developmental trajectories for all youth. Recent work has identified a general exposome factor associated with socio-economic inequality that is strongly related to cognition and individual differences in the spatial organization of functional brain networks in youth. Building on these findings, the current study explores whether alterations in functional connectivity may represent a potential mechanism linking variation in the exposome to cognitive performance. We apply a data-driven, cross-validated, whole-brain machine learning approach, connectome-based statistical inference, to identify patterns of functional connectivity associated with exposome scores among early adolescents enrolled in the Adolescent Brain Cognitive Development (ABCD) Study using data collected during three cognitive tasks and during rest. Additionally, we investigate whether the identified patterns of functional connectivity relate to individual differences in cognitive performance across three domains: General Cognition, Executive Functioning, and Learning/Memory. Models incorporating 10-fold cross-validation over 100 iterations identified consistent functional connections associated with the exposome across task and rest conditions (model performance: ns = 6,137 - 8,391, rs = 0.34 - 0.44, ps <.001). Results were robust across data collection sites and functional connections common across all significant models were associated with cognitive performance across domains (ps < 0.0009). Collectively, these findings reveal that multidimensional environmental exposures are reflected in patterns of functional connectivity and relate to cognitive functioning among youth. HighlightsO_LIPatterns of functional connectivity reflect multivariate environmental exposures. C_LIO_LIConsistent exposome-related connectivity patterns were observed across brain states. C_LIO_LIThe identified general exposome network was associated with cognitive performance across domains. C_LI Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=102 SRC="FIGDIR/small/688261v1_ufig1.gif" ALT="Figure 1"> View larger version (33K): org.highwire.dtl.DTLVardef@b5cc3aorg.highwire.dtl.DTLVardef@172f954org.highwire.dtl.DTLVardef@14f0d01org.highwire.dtl.DTLVardef@deccb5_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2025

Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction

Accurately modeling biomolecular interactions is a central challenge in modern biology. While recent advances, such as AlphaFold3 and Boltz-1, have substantially improved our ability to predict biomolecular complex structures, these models still fall short in predicting binding affinity, a critical property underlying molecular function and therapeutic efficacy. Here, we present Boltz-2, a new structural biology foundation model that exhibits strong performance for both structure and affinity prediction. Boltz-2 introduces controllability features including experimental method conditioning, distance constraints, and multi-chain template integration for structure prediction, and is, to our knowledge, the first AI model to approach the performance of free-energy perturbation (FEP) methods in estimating small molecule-protein binding affinity. Crucially, it achieves strong correlation with experimental readouts on many benchmarks, while being at least 1000x more computationally efficient than FEP. By coupling Boltz-2 with a generative model for small molecules, we demonstrate an effective workflow to find diverse, synthesizable, high-affinity binders, as estimated by absolute FEP simulations on the TYK2 target. To foster broad adoption and further innovation at the intersection of machine learning and biology, we are releasing Boltz-2 weights, inference, and training code 1 under a permissive open license, providing a robust and extensible foundation for both academic and industrial research.

molecular biology↗

bioRxiv2025

Atom level enzyme active site scaffolding using RFdiffusion2

De novo enzyme design starts from ideal active site descriptions consisting of constellations of catalytic residue functional groups around reaction transition state(s), and seeks to generate protein structures that can accurately hold the site in place. Highly active enzymes have been designed starting from such descriptions using the generative AI method RFdiffusion [1-3], but there are two current methodological limitations. First, the geometry of the active site can only be specified at the residue level, so for each catalytic residue functional group placed around the reaction transition state, the possible locations of the residue backbone must be enumerated by building side chain rotamers back from the functional group. Second, the location of the catalytic residues along the sequence must be specified in advance, which considerably limits the space of solutions which can be sampled. Here we describe a new deep generative method, Rosetta Fold diffusion 2 (RFdiffusion2), that solves both problems, enabling enzymes to be designed from sequence agnostic descriptions of functional group locations without inverse rotamer generation. We first evaluate RFdiffusion2 on an in silico enzyme design benchmark of 41 diverse active sites and find that it is able to successfully build proteins scaffolding all 41 sites, compared to 16/41 with prior state-of-the-art deep learning methods. Next, we design enzymes around three diverse catalytic sites and characterize the designs experimentally; in each case we identify active catalysts in testing less than 96 sequences. RFdiffusion2 demonstrates the potential of atomic resolution generative models for the design of de novo enzymes directly from their reaction mechanisms.

biochemistry↗

bioRxiv2024

Boltz-1: Democratizing Biomolecular Interaction Modeling

Understanding biomolecular interactions is fundamental to advancing fields like drug discovery and protein design. In this paper, we introduce BO_SCPLOWOLTZC_SCPLOW-1, an open-source deep learning model incorporating innovations in model architecture, speed optimization, and data processing achieving AO_SCPLOWLPHAC_SCPLOWFO_SCPLOWOLDC_SCPLOW3-level accuracy in predicting the 3D structures of biomolecular complexes. BO_SCPLOWOLTZC_SCPLOW-1 demonstrates a performance on-par with state-of-the-art commercial models on a range of diverse benchmarks, setting a new benchmark for commercially accessible tools in structural biology. Further, we push the boundary of capabilities of these models with BO_SCPLOWOLTZC_SCPLOWO_SCPCAP-C_SCPCAPO_SCPLOWSTEERINGC_SCPLOW, a new inference time steering technique that is able to fix hallucinations and non-physical predictions from the models. By releasing the training and inference code, model weights, datasets, and benchmarks under the MIT open license, we aim to foster global collaboration, accelerate discoveries, and provide a robust platform for advancing biomolecular modeling.

biophysics↗

bioRxiv2023

Improving influenza A vaccine strain selection through deep evolutionary models

Current vaccines provide limited protection against rapidly evolving viruses. For example, the flu vaccines effectiveness has averaged below 40% for the past five years. Today, clinical outcomes of vaccine effectiveness can only be assessed retrospectively. Since vaccine strains are selected at least six months ahead of flu season, prospective estimation of their effectiveness is crucial but remains under-explored. In this paper, we propose an in-silico method named VaxSeer that selects vaccine strains based on their coverage scores, which quantifies expected vaccine effectiveness in future seasons. This score considers both the future dominance of circulating viruses and antigenic profiles of vaccine candidates. Based on historical WHO data, our approach consistently selects superior strains than the annual recommendations. Finally, the prospective coverage score exhibits a strong correlation with retrospective vaccine effectiveness and reduced disease burden, highlighting the promise of this framework in driving the vaccine selection process.

immunology↗

bioRxiv2023

A general exposome factor explains individual differences in functional brain network topography and cognition in youth

Our minds and brains are highly unique. Despite the long-recognized importance of the environment in shaping individual differences in cognitive neurodevelopment, only with the combination of deep phenotyping approaches and the availability of large-scale datasets have we been able to more comprehensively characterize the many inter-connected features of an individuals environment and experience ("exposome"). Moreover, despite clear evidence that brain organization is highly individualized, most neuroimaging studies still rely on group atlases to define functional networks, smearing away inter-individual variation in the spatial layout of functional networks across the cortex ("functional topography"). Here, we leverage the largest longitudinal study of brain and behavior development in the United States to investigate how an individuals exposome may contribute to functional brain network organization leading to differences in cognitive functioning. To do so, we apply three previously-validated data driven computational models to characterize an individuals multidimensional exposome, define individual-specific maps of functional brain networks, and measure cognitive functioning across broad domains. In pre-registered analyses replicated across matched discovery (n=5,139, 48.5% female) and replication (n=5,137, 47.1% female) samples, we find that a childs exposome is associated with multiple domains of cognitive functioning both at baseline assessment and two years later - over and above associations with baseline cognition. Cross-validated ridge regression models reveal that the exposome is reflected in childrens unique patterns of functional topography. Finally, we uncover both shared and unique contributions of the exposome and functional topography to cognitive abilities, finding that models trained on a single variable capturing a childs exposome can more accurately and parsimoniously predict future cognitive performance than models trained on a wealth of personalized neuroimaging data. This study advances our understanding of how childhood environments contribute to unique patterns of functional brain organization and variability in cognitive abilities.

neuroscience↗

bioRxiv2022

Broadly applicable and accurate protein design by integrating structure prediction networks and diffusion generative models

There has been considerable recent progress in designing new proteins using deep learning methods1-9. Despite this progress, a general deep learning framework for protein design that enables solution of a wide range of design challenges, including de novo binder design and design of higher order symmetric architectures, has yet to be described. Diffusion models10,11 have had considerable success in image and language generative modeling but limited success when applied to protein modeling, likely due to the complexity of protein backbone geometry and sequence-structure relationships. Here we show that by fine tuning the RoseTTAFold structure prediction network on protein structure denoising tasks, we obtain a generative model of protein backbones that achieves outstanding performance on unconditional and topology-constrained protein monomer design, protein binder design, symmetric oligomer design, enzyme active site scaffolding, and symmetric motif scaffolding for therapeutic and metal-binding protein design. We demonstrate the power and generality of the method, called RoseTTAFold Diffusion (RFdiffusion), by experimentally characterizing the structures and functions of hundreds of new designs. In a manner analogous to networks which produce images from user-specified inputs, RFdiffusion enables the design of diverse, complex, functional proteins from simple molecular specifications.

biochemistry↗

bioRxiv2022

Personalized Functional Brain Network Topography Predicts Individual Differences in Youth Cognition

Individual differences in cognition during childhood are associated with important social, physical, and mental health outcomes in adolescence and adulthood. Given that cortical surface arealization during development reflects the brains functional prioritization, quantifying variation in the topography of functional brain networks across the developing cortex may provide insight regarding individual differences in cognition. We test this idea by defining personalized functional networks (PFNs) that account for interindividual heterogeneity in functional brain network topography in 9-10 year olds from the Adolescent Brain Cognitive DevelopmentSM Study. Across matched discovery (n=3,525) and replication (n=3,447) samples, the total cortical representation of fronto-parietal PFNs positively correlated with general cognition. Cross-validated ridge regressions trained on PFN topography predicted cognition across domains, with prediction accuracy increasing along the cortexs sensorimotor-association organizational axis. These results establish that functional network topography heterogeneity is associated with individual differences in cognition before the critical transition into adolescence.

neuroscience↗

bioRxiv2021

Stability of Polygenic Scores Across Discovery Genome-Wide Association Studies

Polygenic scores (PGS) are commonly evaluated in terms of their predictive accuracy at the population level by the proportion of phenotypic variance they explain. To be useful for precision medicine applications, they also need to be evaluated at the individual patient level when phenotypes are not necessarily already known. Hence, we investigated the stability of PGS in European-American (EUR)- and African-American (AFR)-ancestry individuals from the Philadelphia Neurodevelopmental Cohort (PNC) and the Adolescent Brain Cognitive Development (ABCD) cohort using different discovery GWAS for post-traumatic stress disorder (PTSD), type-2 diabetes (T2D), and height. We found that pairs of EUR-ancestry GWAS for the same trait had genetic correlations > 0.92. However, PGS calculated from pairs of sameancestry and different-ancestry GWAS had correlations that ranged from <0.01 to 0.74. PGS stability was higher for GWAS that explained more of the trait variance, with height PGS being more stable than PTSD or T2D PGS. Focusing on the upper end of the PGS distribution, different discovery GWAS do not consistently identify the same individuals in the upper quantiles, with the best case being 60% of individuals above the 80th percentile of PGS overlapping from one height GWAS to another. The degree of overlap decreases sharply as higher quantiles, less heritable traits, and different-ancestry GWAS are considered. PGS computed from different discovery GWAS have only modest correlation at the level of the individual patient, underscoring the need to proceed cautiously with integrating PGS into precision medicine applications.

genetics↗