Multimodal Predictive Modeling: Scalable Imaging Informed Approaches to Predict Future Brain Health.

Background: Predicting future brain health is a complex endeavor that often requires integrating diverse data sources. The neural patterns and interactions iden-tified through neuroimaging serve as the fundamental basis and early indica-tors that precede the manifestation of observable behaviors or psychological states. New Method: In this work, we introduce a multimodal predictive modeling approach that leverages an imaging-informed methodology to gain insights into fu-ture behavioral outcomes. We employed three methodologies for evalua-tion: an assessment-only approach using support vector regression (SVR), a neuroimaging-only approach using random forest (RF), and an image-assisted method integrating the static functional network connectivity (sFNC) matrix from resting-state functional magnetic resonance imaging (rs-fMRI) alongside assessments. The image-assisted approach utilized a partially con-ditional variational autoencoder (PCVAE) to predict brain health constructs in future visits from the behavioral data alone. Results: Our performance evaluation indicates that the image-assisted method ex-cels in handling conditional information to predict brain health constructs in subsequent visits and their longitudinal changes. These results suggest that during the training stage, the PCVAE model effectively captures relevant in-formation from neuroimaging data, thereby potentially improving accuracy in making future predictions using only assessment data. Comparison with Existing Methods: The proposed image-assisted method outperforms traditional assessment-only and neuroimaging-only approaches by effectively integrating neuroimag-ing data with assessment factors. Conclusion: This study underscores the potential of neuroimaging-informed predictive modeling to advance our comprehension of the complex relationships between cognitive performance and neural connectivity. Highlights: Multifaceted perspective for studying longitudinal brain health changes.Showcases the versatility of methodologies through assessment-only, neuroimaging-only, and image-assisted predictive approaches.Provides predictive insights by revealing the neural patterns corresponding to alterations in behavior.

A SIMPLE AND FLEXIBLE TEST OF SAMPLE EXCHANGEABILITY WITH APPLICATIONS TO STATISTICAL GENOMICS.

In scientific studies involving analyses of multivariate data, basic but important questions often arise for the researcher: Is the sample exchangeable, meaning that the joint distribution of the sample is invariant to the ordering of the units? Are the features independent of one another, or perhaps the features can be grouped so that the groups are mutually independent? In statistical genomics, these considerations are fundamental to downstream tasks such as demographic inference and the construction of polygenic risk scores. We propose a non-parametric approach, which we call the V test, to address these two questions, namely, a test of sample exchangeability given dependency structure of features, and a test of feature independence given sample exchangeability. Our test is conceptually simple, yet fast and flexible. It controls the Type I error across realistic scenarios, and handles data of arbitrary dimensions by leveraging large-sample asymptotics. Through extensive simulations and a comparison against unsupervised tests of stratification based on random matrix theory, we find that our test compares favorably in various scenarios of interest. We apply the test to data from the 1000 Genomes Project, demonstrating how it can be employed to assess exchangeability of the genetic sample, or find optimal linkage disequilibrium (LD) splits for downstream analysis. For exchangeability assessment, we find that removing rare variants can substantially increase the p-value of the test statistic. For optimal LD splitting, the V test reports different optimal splits than previous approaches not relying on hypothesis testing. Software for our methods is available in R (CRAN: flintyR) and Python (PyPI: flintyPy).

Stable population structure in Europe since the Iron Age, despite high mobility.

Ancient DNA research in the past decade has revealed that European population structure changed dramatically in the prehistoric period (14,000-3000 years before present, YBP), reflecting the widespread introduction of Neolithic farmer and Bronze Age Steppe ancestries. However, little is known about how population structure changed from the historical period onward (3000 YBP - present). To address this, we collected whole genomes from 204 individuals from Europe and the Mediterranean, many of which are the first historical period genomes from their region (e.g. Armenia and France). We found that most regions show remarkable inter-individual heterogeneity. At least 7% of historical individuals carry ancestry uncommon in the region where they were sampled, some indicating cross-Mediterranean contacts. Despite this high level of mobility, overall population structure across western Eurasia is relatively stable through the historical period up to the present, mirroring geography. We show that, under standard population genetics models with local panmixia, the observed level of dispersal would lead to a collapse of population structure. Persistent population structure thus suggests a lower effective migration rate than indicated by the observed dispersal. We hypothesize that this phenomenon can be explained by extensive transient dispersal arising from drastically improved transportation networks and the Roman Empire's mobilization of people for trade, labor, and military. This work highlights the utility of ancient DNA in elucidating finer scale human population dynamics in recent history.

Toward precision brain health: accurate prediction of a cognitive index trajectory using neuroimaging metrics.

The goal of precision brain health is to accurately predict individuals' longitudinal patterns of brain change. We trained a machine learning model to predict changes in a cognitive index of brain health from neurophysiologic metrics. A total of 48 participants (ages 21-65) completed a sensorimotor task during 2 functional magnetic resonance imaging sessions 6 mo apart. Hemodynamic response functions (HRFs) were parameterized using traditional (amplitude, dispersion, latency) and novel (curvature, canonicality) metrics, serving as inputs to a neural network model that predicted gain on indices of brain health (cognitive factor scores) for each participant. The optimal neural network model successfully predicted substantial gain on the cognitive index of brain health with 90% accuracy (determined by 5-fold cross-validation) from 3 HRF parameters: amplitude change, dispersion change, and similarity to a canonical HRF shape at baseline. For individuals with canonical baseline HRFs, substantial gain in the index is overwhelmingly predicted by decreases in HRF amplitude. For individuals with non-canonical baseline HRFs, substantial gain in the index is predicted by congruent changes in both HRF amplitude and dispersion. Our results illustrate that neuroimaging measures can track cognitive indices in healthy states, and that machine learning approaches using novel metrics take important steps toward precision brain health.

Bayesian estimation of gene constraint from an evolutionary model with gene features.

Measures of selective constraint on genes have been used for many applications including clinical interpretation of rare coding variants, disease gene discovery, and studies of genome evolution. However, widely-used metrics are severely underpowered at detecting constraint for the shortest ~25% of genes, potentially causing important pathogenic mutations to be over-looked. We developed a framework combining a population genetics model with machine learning on gene features to enable accurate inference of an interpretable constraint metric, s het . Our estimates outperform existing metrics for prioritizing genes important for cell essentiality, human disease, and other phenotypes, especially for short genes. Our new estimates of selective constraint should have wide utility for characterizing genes relevant to human disease. Finally, our inference framework, GeneBayes, provides a flexible platform that can improve estimation of many gene-level properties, such as rare variant burden or gene expression differences.

Predicting the effect of CRISPR-Cas9-based epigenome editing.

Epigenetic regulation orchestrates mammalian transcription, but functional links between them remain elusive. To tackle this problem, we here use epigenomic and transcriptomic data from 13 ENCODE cell types to train machine learning models to predict gene expression from histone post-translational modifications (PTMs), achieving transcriptome-wide correlations of ~ 0.70 - 0.79 for most samples. In addition to recapitulating known associations between histone PTMs and expression patterns, our models predict that acetylation of histone subunit H3 lysine residue 27 (H3K27ac) near the transcription start site (TSS) significantly increases expression levels. To validate this prediction experimentally and investigate how engineered vs. natural deposition of H3K27ac might differentially affect expression, we apply the synthetic dCas9-p300 histone acetyltransferase system to 8 genes in the HEK293T cell line. Further, to facilitate model building, we perform MNase-seq to map genome-wide nucleosome occupancy levels in HEK293T. We observe that our models perform well in accurately ranking relative fold changes among genes in response to the dCas9-p300 system; however, their ability to rank fold changes within individual genes is noticeably diminished compared to predicting expression across cell types from their native epigenetic signatures. Our findings highlight the need for more comprehensive genome-scale epigenome editing datasets, better understanding of the actual modifications made by epigenome editing tools, and improved causal models that transfer better from endogenous cellular measurements to perturbation experiments. Together these improvements would facilitate the ability to understand and predictably control the dynamic human epigenome with consequences for human health.

And the award goes to the Matthew Effect: Examining external status as a predictor of productivity and opportunity.

It has been suggested that increased status that comes from being an award winner can generate enduring advantages that compound over one's career via the Matthew Effect. However, research in this area has yielded conflicting results and has been unable to isolate the unique effect of status on career outcomes from the positive endogenous characteristics of award winners. In the current research, we attempt to address previous research limitations and examine if winning an award is associated with career outcomes (i.e., opportunities and productivity) irrespective of individual productivity levels prior to receiving an award. We examined our research questions using observational data of National Hockey League (NHL) league championship winners and non-winners (N = 427). By using a team award and several different analytic approaches we were able to examine the unique effects of affiliation-based external status, generated from an award win, on career outcomes. Our results generally show support for the Matthew Effect and suggest that affiliation-based external status, achieved by an award win, provides access to increased opportunities, which ultimately results in more productivity. We discuss the importance of incorporating opportunity and investigating its role in the cumulative advantage process and implications of the results.

Systematic differences in discovery of genetic effects on gene expression and complex traits.

Most signals in genome-wide association studies (GWAS) of complex traits implicate noncoding genetic variants with putative gene regulatory effects. However, currently identified regulatory variants, notably expression quantitative trait loci (eQTLs), explain only a small fraction of GWAS signals. Here, we show that GWAS and cis-eQTL hits are systematically different: eQTLs cluster strongly near transcription start sites, whereas GWAS hits do not. Genes near GWAS hits are enriched in key functional annotations, are under strong selective constraint and have complex regulatory landscapes across different tissue/cell types, whereas genes near eQTLs are depleted of most functional annotations, show relaxed constraint, and have simpler regulatory landscapes. We describe a model to understand these observations, including how natural selection on complex traits hinders discovery of functionally relevant eQTLs. Our results imply that GWAS and eQTL studies are systematically biased toward different types of variant, and support the use of complementary functional approaches alongside the next generation of eQTL studies.

Scaling the discrete-time Wright-Fisher model to biobank-scale datasets.

The discrete-time Wright-Fisher (DTWF) model and its diffusion limit are central to population genetics. These models can describe the forward-in-time evolution of allele frequencies in a population resulting from genetic drift, mutation, and selection. Computing likelihoods under the diffusion process is feasible, but the diffusion approximation breaks down for large samples or in the presence of strong selection. Existing methods for computing likelihoods under the DTWF model do not scale to current exome sequencing sample sizes in the hundreds of thousands. Here, we present a scalable algorithm that approximates the DTWF model with provably bounded error. Our approach relies on two key observations about the DTWF model. The first is that transition probabilities under the model are approximately sparse. The second is that transition distributions for similar starting allele frequencies are extremely close as distributions. Together, these observations enable approximate matrix-vector multiplication in linear (as opposed to the usual quadratic) time. We prove similar properties for Hypergeometric distributions, enabling fast computation of likelihoods for subsamples of the population. We show theoretically and in practice that this approximation is highly accurate and can scale to population sizes in the tens of millions, paving the way for rigorous biobank-scale inference. Finally, we use our results to estimate the impact of larger samples on estimating selection coefficients for loss-of-function variants. We find that increasing sample sizes beyond existing large exome sequencing cohorts will provide essentially no additional information except for genes with the most extreme fitness effects.

Exploring how brain health strategy training informs the future of work.

Introduction: The workplace typically affords one of the longest periods for continued brain health growth. Brain health is defined by the World Health Organization (WHO) as the promotion of optimal brain development, cognitive health, and well-being across the life course, which we expanded to also include connectedness to people and purpose. This work was motivated by prior work showing individuals, outside of an aggregate setting, benefitted from training as measured by significant performance gains on a holistic BrainHealth Index and its factors (i.e., clarity, connectedness, emotional balance). The current research was conducted during the changing remote work practices emerging post-pandemic to test whether a capacity-building training would be associated with significant gains on measures of brain health and components of burnout. The study also tested the influence of utilization of training modules and days in office for individuals to inform workplace practices. Methods: We investigated whether 193 individuals across a firm's sites would improve on measures of brain health and burnout from micro-delivery of online tactical brain health strategies, combined with two individualized coaching sessions, and practical exercises related to work and personal life, over a six-month period. Brain health was measured using an evidenced-based measure (BrainHealth™ Index) with its components (clarity, connectedness, emotional balance) consistent with the WHO definition. Burnout was measured using the Maslach Burnout Inventory Human Services Survey. Days in office were determined by access to digital workplace applications from the firm's network. Regression analyses were used to assess relationships between change in BrainHealth factors and change in components of the Maslach Burnout Inventory. Results: Results at posttest indicated that 75% of the individuals showed gains on a composite BrainHealth Index and across all three composite factors contributing to brain health. Benefits were directly tied to training utilization such that those who completed the core modules showed the greatest gains. The current results also found an association between gains on both the connectedness and emotional balance brain health factors and reduced on burnout components of occupational exhaustion and depersonalization towards one's workplace. We found that fewer days in the office were associated with greater gains in the clarity factor, but not for connectedness and emotional balance. Discussion: These results support the value of a proactive, capacity-building training to benefit all employees to complement the more widespread limited offerings that address a smaller segment who need mental illness assistance programs. The future of work may be informed by corporate investment in focused efforts to boost collective brain capital through a human-centered, capacity-building approach. Efforts are underway to uncover the value of better brain health, i.e., Brainomics© - which includes economic, societal, and individual benefits.

A genetic history of continuity and mobility in the Iron Age central Mediterranean.

The Iron Age was a dynamic period in central Mediterranean history, with the expansion of Greek and Phoenician colonies and the growth of Carthage into the dominant maritime power of the Mediterranean. These events were facilitated by the ease of long-distance travel following major advances in seafaring. We know from the archaeological record that trade goods and materials were moving across great distances in unprecedented quantities, but it is unclear how these patterns correlate with human mobility. Here, to investigate population mobility and interactions directly, we sequenced the genomes of 30 ancient individuals from coastal cities around the central Mediterranean, in Tunisia, Sardinia and central Italy. We observe a meaningful contribution of autochthonous populations, as well as highly heterogeneous ancestry including many individuals with non-local ancestries from other parts of the Mediterranean region. These results highlight both the role of local populations and the extreme interconnectedness of populations in the Iron Age Mediterranean. By studying these trans-Mediterranean neighbours together, we explore the complex interplay between local continuity and mobility that shaped the Iron Age societies of the central Mediterranean.

Bayesian estimation of gene constraint from an evolutionary model with gene features.

Measures of selective constraint on genes have been used for many applications including clinical interpretation of rare coding variants, disease gene discovery, and studies of genome evolution. However, widely-used metrics are severely underpowered at detecting constraint for the shortest ~25% of genes, potentially causing important pathogenic mutations to be overlooked. We developed a framework combining a population genetics model with machine learning on gene features to enable accurate inference of an interpretable constraint metric, shet. Our estimates outperform existing metrics for prioritizing genes important for cell essentiality, human disease, and other phenotypes, especially for short genes. Our new estimates of selective constraint should have wide utility for characterizing genes relevant to human disease. Finally, our inference framework, GeneBayes, provides a flexible platform that can improve estimation of many gene-level properties, such as rare variant burden or gene expression differences.

Expanding the stdpopsim species catalog, and lessons learned for realistic genome simulations.

Simulation is a key tool in population genetics for both methods development and empirical research, but producing simulations that recapitulate the main features of genomic datasets remains a major obstacle. Today, more realistic simulations are possible thanks to large increases in the quantity and quality of available genetic data, and the sophistication of inference and simulation software. However, implementing these simulations still requires substantial time and specialized knowledge. These challenges are especially pronounced for simulating genomes for species that are not well-studied, since it is not always clear what information is required to produce simulations with a level of realism sufficient to confidently answer a given question. The community-developed framework stdpopsim seeks to lower this barrier by facilitating the simulation of complex population genetic models using up-to-date information. The initial version of stdpopsim focused on establishing this framework using six well-characterized model species (Adrion et al., 2020). Here, we report on major improvements made in the new release of stdpopsim (version 0.2), which includes a significant expansion of the species catalog and substantial additions to simulation capabilities. Features added to improve the realism of the simulated genomes include non-crossover recombination and provision of species-specific genomic annotations. Through community-driven efforts, we expanded the number of species in the catalog more than threefold and broadened coverage across the tree of life. During the process of expanding the catalog, we have identified common sticking points and developed the best practices for setting up genome-scale simulations. We describe the input data required for generating a realistic simulation, suggest good practices for obtaining the relevant information from the literature, and discuss common pitfalls and major considerations. These improvements to stdpopsim aim to further promote the use of realistic whole-genome population genetic simulations, especially in non-model organisms, making them available, transparent, and accessible to everyone.

Scaling the Discrete-time Wright Fisher model to biobank-scale datasets.

The Discrete-Time Wright Fisher (DTWF) model and its large population diffusion limit are central to population genetics. These models describe the forward-in-time evolution of the frequency of an allele in a population and can include the fundamental forces of genetic drift, mutation, and selection. Computing like-lihoods under the diffusion process is feasible, but the diffusion approximation breaks down for large sample sizes or in the presence of strong selection. Unfortunately, existing methods for computing likelihoods under the DTWF model do not scale to current exome sequencing sample sizes in the hundreds of thousands. Here we present an algorithm that approximates the DTWF model with provably bounded error and runs in time linear in the size of the population. Our approach relies on two key observations about Binomial distributions. The first is that Binomial distributions are approximately sparse. The second is that Binomial distributions with similar success probabilities are extremely close as distributions, allowing us to approximate the DTWF Markov transition matrix as a very low rank matrix. Together, these observations enable matrix-vector multiplication in linear (as opposed to the usual quadratic) time. We prove similar properties for Hypergeometric distributions, enabling fast computation of likelihoods for subsamples of the population. We show theoretically and in practice that this approximation is highly accurate and can scale to population sizes in the billions, paving the way for rigorous biobank-scale population genetic inference. Finally, we use our results to estimate how increasing sample sizes will improve the estimation of selection coefficients acting on loss-of-function variants. We find that increasing sample sizes beyond existing large exome sequencing cohorts will provide essentially no additional information except for genes with the most extreme fitness effects.

The ENCODE Imputation Challenge: a critical assessment of methods for cross-cell type imputation of epigenomic profiles.

A promising alternative to comprehensively performing genomics experiments is to, instead, perform a subset of experiments and use computational methods to impute the remainder. However, identifying the best imputation methods and what measures meaningfully evaluate performance are open questions. We address these questions by comprehensively analyzing 23 methods from the ENCODE Imputation Challenge. We find that imputation evaluations are challenging and confounded by distributional shifts from differences in data collection and processing over time, the amount of available data, and redundancy among performance measures. Our analyses suggest simple steps for overcoming these issues and promising directions for more robust research.

Whole-genome sequencing reveals a complex African population demographic history and signatures of local adaptation.

We conduct high coverage (>30×) whole-genome sequencing of 180 individuals from 12 indigenous African populations. We identify millions of unreported variants, many predicted to be functionally important. We observe that the ancestors of southern African San and central African rainforest hunter-gatherers (RHG) diverged from other populations >200 kya and maintained a large effective population size. We observe evidence for ancient population structure in Africa and for multiple introgression events from "ghost" populations with highly diverged genetic lineages. Although currently geographically isolated, we observe evidence for gene flow between eastern and southern Khoesan-speaking hunter-gatherer populations lasting until ∼12 kya. We identify signatures of local adaptation for traits related to skin color, immune response, height, and metabolic processes. We identify a positively selected variant in the lightly pigmented San that influences pigmentation in vitro by regulating the enhancer activity and gene expression of PDPK1.

Understanding the Relationships Between Health Care Providers' Moral Distress and Patients' Quality of Death.

Introduction: Moral distress is a commonly recognized phenomenon among health care providers; however, the experience of moral distress by staff caring for patients who die during an acute care hospital stay has not been previously examined. It also remains unclear how the quality of a death may impact moral distress among these providers. Objectives: We sought to understand levels of moral distress experienced by intern physicians and nurses who provided care during a patient's final 48 hours of life, and how the perceived quality of death impacted moral distress. Materials and Methods: We utilized a mixed-method prospective cohort design, surveying nurses and interns following inpatient hospital deaths at an academic safety-net hospital in the United States. Participants completed surveys and answered open-ended questions to evaluate moral distress and the quality of the patient's death. Results: A total of 126 surveys were sent to nurses and interns caring for 35 patients who died, with 46 surveys completed. Overall moderate-to-high levels of moral distress were identified among participants, and we found that higher levels of moral distress correlated with lower perceived quality of death. We identified five themes in our qualitative analysis highlighting the challenges nurses and interns face in end-of-life care, including the following: poor communication, unexpected deaths, patient suffering, resource limitations, and failure to prioritize a patient's wishes or best interests. Conclusions: Nurses and interns experience moderate-to-high levels of moral distress when caring for dying patients. Lower quality of end-of-life care is associated with higher levels of moral distress.

Diversity of ribosomes at the level of rRNA variation associated with human health and disease.

With hundreds of copies of ribosomal DNA (rDNA) it is unknown whether they possess sequence variations that ultimately form different types of ribosomes. Here, we developed an algorithm for variant-calling between paralog genes (termed RGA) and compared rDNA variations with rRNA variations from long-read sequencing of translating ribosomes (RIBO-RT). Our analyses identified dozens of highly abundant rRNA variants, largely indels, that are incorporated into translationally active ribosomes and assemble into distinct ribosome subtypes encoded on different chromosomes. We developed an in-situ rRNA sequencing method (SWITCH-seq) revealing that variants are co-expressed within individual cells and found that they possess different structures. Lastly, we observed tissue-specific rRNA-subtype expression and linked specific rRNA variants to cancer. This study therefore reveals the variation landscape of translating ribosomes within human cells.

Greater physical fitness (Vo2Max) in healthy older adults associated with increased integrity of the Locus Coeruleus-Noradrenergic system.

Physical activity (PA) is a key component for brain health and Reserve, and it is among the main dementia protective factors. However, the neurobiological mechanisms underpinning Reserve are not fully understood. In this regard, a noradrenergic (NA) theory of cognitive reserve (Robertson, 2013) has proposed that the upregulation of NA system might be a key factor for building reserve and resilience to neurodegeneration because of the neuroprotective role of NA across the brain. PA elicits an enhanced catecholamine response, in particular for NA. By increasing physical commitment, a greater amount of NA is synthetised in response to higher oxygen demand. More physically trained individuals show greater capabilities to carry oxygen resulting in greater Vo2max - a measure of oxygen uptake and physical fitness (PF). In the current study, we hypothesised that greater Vo2 max would be related to greater Locus Coeruleus (LC) MRI signal intensity. As hypothesised, greater Vo2max related to greater LC signal intensity across 41 healthy adults (age range 60-72). As a control procedure, in which these analyses were repeated for the other neuromodulators' seeds (for Serotonin, Dopamine and Acetylcholine), weaker associations emerged. This newly established link between Vo2max and LC-NA system offers further understanding of the neurobiology underpinning Reserve in relationship to PA. While this study supports Robertson's theory proposing the upregulation of the noradrenergic system as a possible key factor building Reserve, it also provide grounds for increasing LC-NA system resilience to neurodegeneration via Vo2max enhancement.

Effects of online brain training on self-reported mental health symptoms for generally healthy adults during the Covid-19 pandemic.

BACKGROUND: The cognitive training Strategic Memory Advanced Reasoning Training (SMART) has been shown to improve symptoms of depression, anxiety, and stress when completed using in-person delivery, but mental health outcomes have not yet been studied for online delivery of SMART. METHODS: Data was analyzed from 145 generally healthy adults participating in the BrainHealth Project pilot study who had access to 12 weeks of online self-paced SMART and self-reported mental health symptoms on the Depression Anxiety Stress Scale (DASS-21) pre- and post-training. We utilized linear models to examine the change in self-reported symptoms of depression, anxiety, and stress following the 12-week training period and to explore the influence of age, gender, and education on changes in symptomatology. Data from 44 participants who completed a follow-up DASS-21 6 months after completing SMART was used to explore the lasting impact of the training. RESULTS: Improvements in depression, anxiety, and stress symptoms were observed following online SMART, evidenced by a significant decrease in self-reported symptoms on the DASS-21. Improvement in self-reported mental health symptomatology was maintained or continued to improve 6-month post-training. No significant effect of gender was observed, but findings motivate additional exploration of the effects of education and age. CONCLUSION: Online SMART should be considered a low-cost, high-impact approach for supporting public mental health for generally healthy adults.