A Bayesian approach to Mendelian randomization using summary statistics in the univariable and multivariable settings with correlated pleiotropy.

Mendelian randomization uses genetic variants as instrumental variables to make causal inferences on the effect of an exposure on an outcome. Due to the recent abundance of high-powered genome-wide association studies, many putative causal exposures of interest have large numbers of independent genetic variants with which they associate, each representing a potential instrument for use in a Mendelian randomization analysis. Such polygenic analyses increase the power of the study design to detect causal effects; however, they also increase the potential for bias due to instrument invalidity. Recent attention has been given to dealing with bias caused by correlated pleiotropy, which results from violation of the "instrument strength independent of direct effect" assumption. Although methods have been proposed that can account for this bias, a number of restrictive conditions remain in many commonly used techniques. In this paper, we propose a Bayesian framework for Mendelian randomization that provides valid causal inference under very general settings. We propose the methods MR-Horse and MVMR-Horse, which can be performed without access to individual-level data, using only summary statistics of the type commonly published by genome-wide association studies, and can account for both correlated and uncorrelated pleiotropy. In simulation studies, we show that the approach retains type I error rates below nominal levels even in high-pleiotropy scenarios. We demonstrate the proposed approaches in applied examples in both univariable and multivariable settings, some with very weak instruments.

Bacterial lifestyle shapes pangenomes.

Pangenomes vary across bacteria. Some species have fluid pangenomes, with a high proportion of genes varying between individual genomes. Other species have less fluid pangenomes, with different genomes tending to contain the same genes. Two main hypotheses have been suggested to explain this variation: differences in species' bacterial lifestyle and effective population size. However, previous studies have not been able to test between these hypotheses because the different features of lifestyle and effective population size are highly correlated with each other, and phylogenetically conserved, making it hard to disentangle their relative importance. We used phylogeny-based analyses, across 126 bacterial species, to tease apart the causal role of different factors. We found that pangenome fluidity was lower in i) host-associated compared with free-living species and ii) host-associated species that are obligately dependent on a host, live inside cells, and are more pathogenic and less motile. In contrast, we found no support for the competing hypothesis that larger effective population sizes lead to more fluid pangenomes. Effective population size appears to correlate with pangenome variation because it is also driven by bacterial lifestyle, rather than because of a causal relationship.

A diverse portfolio of marine protected areas can better advance global conservation and equity.

Marine protected areas (MPAs) are widely used for ocean conservation, yet the relative impacts of various types of MPAs are poorly understood. We estimated impacts on fish biomass from no-take and multiple-use (fished) MPAs, employing a rigorous matched counterfactual design with a global dataset of >14,000 surveys in and around 216 MPAs. Both no-take and multiple-use MPAs generated positive conservation outcomes relative to no protection (58.2% and 12.6% fish biomass increases, respectively), with smaller estimated differences between the two MPA types when controlling for additional confounding factors (8.3% increase). Relative performance depended on context and management: no-take MPAs performed better in areas of high human pressure but similar to multiple-use in remote locations. Multiple-use MPA performance was low in high-pressure areas but improved significantly with better management, producing similar outcomes to no-take MPAs when adequately staffed and appropriate use regulations were applied. For priority conservation areas where no-take restrictions are not possible or ethical, our findings show that a portfolio of well-designed and well-managed multiple-use MPAs represents a viable and potentially equitable pathway to advance local and global conservation.

Using genetic association data to guide drug discovery and development: Review of methods and applications.

Evidence on the validity of drug targets from randomized trials is reliable but typically expensive and slow to obtain. In contrast, evidence from conventional observational epidemiological studies is less reliable because of the potential for bias from confounding and reverse causation. Mendelian randomization is a quasi-experimental approach analogous to a randomized trial that exploits naturally occurring randomization in the transmission of genetic variants. In Mendelian randomization, genetic variants that can be regarded as proxies for an intervention on the proposed drug target are leveraged as instrumental variables to investigate potential effects on biomarkers and disease outcomes in large-scale observational datasets. This approach can be implemented rapidly for a range of drug targets to provide evidence on their effects and thus inform on their priority for further investigation. In this review, we present statistical methods and their applications to showcase the diverse opportunities for applying Mendelian randomization in guiding clinical development efforts, thus enabling interventions to target the right mechanism in the right population group at the right time. These methods can inform investigators on the mechanisms underlying drug effects, their related biomarkers, implications for the timing of interventions, and the population subgroups that stand to gain the most benefit. Most methods can be implemented with publicly available data on summarized genetic associations with traits and diseases, meaning that the only major limitations to their usage are the availability of appropriately powered studies for the exposure and outcome and the existence of a suitable genetic proxy for the proposed intervention.

Multi-response Mendelian randomization: Identification of shared and distinct exposures for multimorbidity and multiple related disease outcomes.

The existing framework of Mendelian randomization (MR) infers the causal effect of one or multiple exposures on one single outcome. It is not designed to jointly model multiple outcomes, as would be necessary to detect causes of more than one outcome and would be relevant to model multimorbidity or other related disease outcomes. Here, we introduce multi-response Mendelian randomization (MR2), an MR method specifically designed for multiple outcomes to identify exposures that cause more than one outcome or, conversely, exposures that exert their effect on distinct responses. MR2 uses a sparse Bayesian Gaussian copula regression framework to detect causal effects while estimating the residual correlation between summary-level outcomes, i.e., the correlation that cannot be explained by the exposures, and vice versa. We show both theoretically and in a comprehensive simulation study how unmeasured shared pleiotropy induces residual correlation between outcomes irrespective of sample overlap. We also reveal how non-genetic factors that affect more than one outcome contribute to their correlation. We demonstrate that by accounting for residual correlation, MR2 has higher power to detect shared exposures causing more than one outcome. It also provides more accurate causal effect estimates than existing methods that ignore the dependence between related responses. Finally, we illustrate how MR2 detects shared and distinct causal exposures for five cardiovascular diseases in two applications considering cardiometabolic and lipidomic exposures and uncovers residual correlation between summary-level outcomes reflecting known relationships between cardiovascular diseases.

High-dimensional generalized median adaptive lasso with application to omics data.

Recently, there has been a growing interest in variable selection for causal inference within the context of high-dimensional data. However, when the outcome exhibits a skewed distribution, ensuring the accuracy of variable selection and causal effect estimation might be challenging. Here, we introduce the generalized median adaptive lasso (GMAL) for covariate selection to achieve an accurate estimation of causal effect even when the outcome follows skewed distributions. A distinctive feature of our proposed method is that we utilize a linear median regression model for constructing penalty weights, thereby maintaining the accuracy of variable selection and causal effect estimation even when the outcome presents extremely skewed distributions. Simulation results showed that our proposed method performs comparably to existing methods in variable selection when the outcome follows a symmetric distribution. Besides, the proposed method exhibited obvious superiority over the existing methods when the outcome follows a skewed distribution. Meanwhile, our proposed method consistently outperformed the existing methods in causal estimation, as indicated by smaller root-mean-square error. We also utilized the GMAL method on a deoxyribonucleic acid methylation dataset from the Alzheimer's disease (AD) neuroimaging initiative database to investigate the association between cerebrospinal fluid tau protein levels and the severity of AD.

Impaired perception of temporal contiguity between action and effect is associated with disorders of agency in schizophrenia.

Delusions of control in schizophrenia are characterized by the striking feeling that one's actions are controlled by external forces. We here tested qualitative predictions inspired by Bayesian causal inference models, which suggest that such misattributions of agency should lead to decreased intentional binding. Intentional binding refers to the phenomenon that subjects perceive a compression of time between their intentional actions and consequent sensory events. We demonstrate that patients with delusions of control perceived less self-agency in our intentional binding task. This effect was accompanied by significant reductions of intentional binding as compared to healthy controls and patients without delusions. Furthermore, the strength of delusions of control tightly correlated with decreases in intentional binding. Our study validated a critical prediction of Bayesian accounts of intentional binding, namely that a pathological reduction of the prior likelihood of a causal relation between one's actions and consequent sensory events-here captured by delusions of control-should lead to lesser intentional binding. Moreover, our study highlights the import of an intact perception of temporal contiguity between actions and their effects for the sense of agency.

Identifying causal subsequent memory effects.

Over 40 y of accumulated research has detailed associations between neuroimaging signals measured during a memory encoding task and later memory performance, across a variety of brain regions, measurement tools, statistical approaches, and behavioral tasks. But the interpretation of these subsequent memory effects (SMEs) remains unclear: if the identified signals reflect cognitive and neural mechanisms of memory encoding, then the underlying neural activity must be causally related to future memory. However, almost all previous SME analyses do not control for potential confounders of this causal interpretation, such as serial position and item effects. We collect a large fMRI dataset and use an experimental design and analysis approach that allows us to statistically adjust for nearly all known exogenous confounding variables. We find that, using standard approaches without adjustment, we replicate several univariate and multivariate subsequent memory effects and are able to predict memory performance across people. However, we are unable to identify any signal that reliably predicts subsequent memory after adjusting for confounding variables, bringing into doubt the causal status of these effects. We apply the same approach to subjects' judgments of learning collected following an encoding period and show that these behavioral measures of mnemonic status do predict memory after adjustments, suggesting that it is possible to measure signals near the time of encoding that reflect causal mechanisms but that existing neuroimaging measures, at least in our data, may not have the precision and specificity to do so.

Large-scale land acquisitions exacerbate local farmland inequalities in Tanzania.

Land inequality stalls economic development, entrenches poverty, and is associated with environmental degradation. Yet, rigorous assessments of land-use interventions attend to inequality only rarely. A land inequality lens is especially important to understand how recent large-scale land acquisitions (LSLAs) affect smallholder and indigenous communities across as much as 100 million hectares around the world. This paper studies inequalities in land assets, specifically landholdings and farm size, to derive insights into the distributional outcomes of LSLAs. Using a household survey covering four pairs of land acquisition and control sites in Tanzania, we use a quasi-experimental design to characterize changes in land inequality and subsequent impacts on well-being. We find convincing evidence that LSLAs in Tanzania lead to both reduced landholdings and greater farmland inequality among smallholders. Households in proximity to LSLAs are associated with 21.1% (P = 0.02) smaller landholdings while evidence, although insignificant, is suggestive that farm sizes are also declining. Aggregate estimates, however, hide that households in the bottom quartiles of farm size suffer the brunt of landlessness and land loss induced by LSLAs that combine to generate greater farmland inequality. Additional analyses find that land inequality is not offset by improvements in other livelihood dimensions, rather farm size decreases among households near LSLAs are associated with no income improvements, lower wealth, increased poverty, and higher food insecurity. The results demonstrate that without explicit consideration of distributional outcomes, land-use policies can systematically reinforce existing inequalities.

Network inference from short, noisy, low time-resolution, partial measurements: Application to C. elegans neuronal calcium dynamics.

Network link inference from measured time series data of the behavior of dynamically interacting network nodes is an important problem with wide-ranging applications, e.g., estimating synaptic connectivity among neurons from measurements of their calcium fluorescence. Network inference methods typically begin by using the measured time series to assign to any given ordered pair of nodes a numerical score reflecting the likelihood of a directed link between those two nodes. In typical cases, the measured time series data may be subject to limitations, including limited duration, low sampling rate, observational noise, and partial nodal state measurement. However, it is unknown how the performance of link inference techniques on such datasets depends on these experimental limitations of data acquisition. Here, we utilize both synthetic data generated from coupled chaotic systems as well as experimental data obtained from Caenorhabditis elegans neural activity to systematically assess the influence of data limitations on the character of scores reflecting the likelihood of a directed link between a given node pair. We do this for three network inference techniques: Granger causality, transfer entropy, and, a machine learning-based method. Furthermore, we assess the ability of appropriate surrogate data to determine statistical confidence levels associated with the results of link-inference techniques.

simmrd: An open-source tool to perform simulations in Mendelian randomization.

Mendelian randomization (MR) has become a popular tool for inferring causality of risk factors on disease. There are currently over 45 different methods available to perform MR, reflecting this extremely active research area. It would be desirable to have a standard simulation environment to objectively evaluate the existing and future methods. We present simmrd, an open-source software for performing simulations to evaluate the performance of MR methods in a range of scenarios encountered in practice. Researchers can directly modify the simmrd source code so that the research community may arrive at a widely accepted framework for researchers to evaluate the performance of different MR methods.

Combining evidence from Mendelian randomization and colocalization: Review and comparison of approaches.

Mendelian randomization and colocalization are two statistical approaches that can be applied to summarized data from genome-wide association studies (GWASs) to understand relationships between traits and diseases. However, despite similarities in scope, they are different in their objectives, implementation, and interpretation, in part because they were developed to serve different scientific communities. Mendelian randomization assesses whether genetic predictors of an exposure are associated with the outcome and interprets an association as evidence that the exposure has a causal effect on the outcome, whereas colocalization assesses whether two traits are affected by the same or distinct causal variants. When considering genetic variants in a single genetic region, both approaches can be performed. While a positive colocalization finding typically implies a non-zero Mendelian randomization estimate, the reverse is not generally true: there are several scenarios which would lead to a non-zero Mendelian randomization estimate but lack evidence for colocalization. These include the existence of distinct but correlated causal variants for the exposure and outcome, which would violate the Mendelian randomization assumptions, and a lack of strong associations with the outcome. As colocalization was developed in the GWAS tradition, typically evidence for colocalization is concluded only when there is strong evidence for associations with both traits. In contrast, a non-zero estimate from Mendelian randomization can be obtained despite only nominally significant genetic associations with the outcome at the locus. In this review, we discuss how the two approaches can provide complementary information on potential therapeutic targets.

Investigating the shared genetic architecture of uterine leiomyoma and breast cancer: A genome-wide cross-trait analysis.

Little is known regarding the shared genetic architecture or causality underlying the phenotypic association observed for uterine leiomyoma (UL) and breast cancer (BC). Leveraging summary statistics from the hitherto largest genome-wide association study (GWAS) conducted in each trait, we investigated the genetic overlap and causal associations of UL with BC overall, as well as with its subtypes defined by the status of estrogen receptor (ER). We observed a positive genetic correlation between UL and BC overall (rg = 0.09, p = 6.00 × 10-3), which was consistent in ER+ subtype (rg = 0.06, p = 0.01) but not in ER- subtype (rg = 0.06, p = 0.08). Partitioning the whole genome into 1,703 independent regions, local genetic correlation was identified at 22q13.1 for UL with BC overall and with ER+ subtype. Significant genetic correlation was further discovered in 9 out of 14 functional categories, with the highest estimates observed in coding, H3K9ac, and repressed regions. Cross-trait meta-analysis identified 9 novel loci shared between UL and BC. Mendelian randomization demonstrated a significantly increased risk of BC overall (OR = 1.09, 95% CI = 1.01-1.18) and ER+ subtype (OR = 1.09, 95% CI = 1.01-1.17) for genetic liability to UL. No reverse causality was found. Our comprehensive genome-wide cross-trait analysis demonstrates a shared genetic basis, pleiotropic loci, as well as a putative causal relationship between UL and BC, highlighting an intrinsic link underlying these two complex female diseases.

Integrating Evidence to Guide Use of Biologics and Small Molecules for Inflammatory Bowel Diseases.

Advances in science have led to the development of multiple biologics and small molecules for the treatment of inflammatory bowel diseases (IBDs). This growth in advanced medical therapies has been accompanied by an increase in methodological innovation to study and compare therapies. Guidelines provide an evidence-based approach to integrating therapies into routine practice, but they are often unable to provide timely recommendations as new therapies come to market, and they have limited incorporation of real-world evidence when making recommendations. This limits the scope and usability of guidelines, and a gap remains in defining how best to position and integrate advanced medical therapies for IBD. In this review, we provide a framework for clinicians and researchers to understand key differences in sources of evidence, how different methodologies are applied to study the comparative effectiveness of advanced medical therapies in IBD, and considerations for how these sources of evidence can be used to better integrate current guideline recommendations. Over time, we anticipate this framework will allow for a transition to living guidelines and/or practice recommendations.

Bayesian semiparametric model for sequential treatment decisions with informative timing.

We develop a Bayesian semiparametric model for the impact of dynamic treatment rules on survival among patients diagnosed with pediatric acute myeloid leukemia (AML). The data consist of a subset of patients enrolled in a phase III clinical trial in which patients move through a sequence of four treatment courses. At each course, they undergo treatment that may or may not include anthracyclines (ACT). While ACT is known to be effective at treating AML, it is also cardiotoxic and can lead to early death for some patients. Our task is to estimate the potential survival probability under hypothetical dynamic ACT treatment strategies, but there are several impediments. First, since ACT is not randomized, its effect on survival is confounded over time. Second, subjects initiate the next course depending on when they recover from the previous course, making timing potentially informative of subsequent treatment and survival. Third, patients may die or drop out before ever completing the full treatment sequence. We develop a generative Bayesian semiparametric model based on Gamma Process priors to address these complexities. At each treatment course, the model captures subjects' transition to subsequent treatment or death in continuous time. G-computation is used to compute a posterior over potential survival probability that is adjusted for time-varying confounding. Using our approach, we estimate the efficacy of hypothetical treatment rules that dynamically modify ACT based on evolving cardiac function.

Practical causal mediation analysis: extending nonparametric estimators to accommodate multiple mediators and multiple intermediate confounders.

Mediation analysis is appealing for its ability to improve understanding of the mechanistic drivers of causal effects, but real-world data complexities challenge its successful implementation, including (i) the existence of post-exposure variables that also affect mediators and outcomes (thus, confounding the mediator-outcome relationship), that may also be (ii) multivariate, and (iii) the existence of multivariate mediators. All three challenges are present in the mediation analysis we consider here, where our goal is to estimate the indirect effects of receiving a Section 8 housing voucher as a young child on the risk of developing a psychiatric mood disorder in adolescence that operate through mediators related to neighborhood poverty, the school environment, and instability of the neighborhood and school environments, considered together and separately. Interventional direct and indirect effects (IDE/IIE) accommodate post-exposure variables that confound the mediator-outcome relationship, but currently, no readily implementable nonparametric estimator for IDE/IIE exists that allows for both multivariate mediators and multivariate post-exposure intermediate confounders. The absence of such an IDE/IIE estimator that can easily accommodate both multivariate mediators and post-exposure confounders represents a significant limitation for real-world analyses, because when considering each mediator subgroup separately, the remaining mediator subgroups (or a subset of them) become post-exposure intermediate confounders. We address this gap by extending a recently developed nonparametric estimator for the IDE/IIE to allow for easy incorporation of multivariate mediators and multivariate post-exposure confounders simultaneously. We apply the proposed estimation approach to our analysis, including walking through a strategy to account for other, possibly co-occurring intermediate variables when considering each mediator subgroup separately.

Tree-based subgroup discovery using electronic health record data: heterogeneity of treatment effects for DTG-containing therapies.

The rich longitudinal individual level data available from electronic health records (EHRs) can be used to examine treatment effect heterogeneity. However, estimating treatment effects using EHR data poses several challenges, including time-varying confounding, repeated and temporally non-aligned measurements of covariates, treatment assignments and outcomes, and loss-to-follow-up due to dropout. Here, we develop the subgroup discovery for longitudinal data algorithm, a tree-based algorithm for discovering subgroups with heterogeneous treatment effects using longitudinal data by combining the generalized interaction tree algorithm, a general data-driven method for subgroup discovery, with longitudinal targeted maximum likelihood estimation. We apply the algorithm to EHR data to discover subgroups of people living with human immunodeficiency virus who are at higher risk of weight gain when receiving dolutegravir (DTG)-containing antiretroviral therapies (ARTs) versus when receiving non-DTG-containing ARTs.

Robust discovery of gene regulatory networks from single-cell gene expression data by Causal Inference Using Composition of Transactions.

Gene regulatory networks (GRNs) drive organism structure and functions, so the discovery and characterization of GRNs is a major goal in biological research. However, accurate identification of causal regulatory connections and inference of GRNs using gene expression datasets, more recently from single-cell RNA-seq (scRNA-seq), has been challenging. Here we employ the innovative method of Causal Inference Using Composition of Transactions (CICT) to uncover GRNs from scRNA-seq data. The basis of CICT is that if all gene expressions were random, a non-random regulatory gene should induce its targets at levels different from the background random process, resulting in distinct patterns in the whole relevance network of gene-gene associations. CICT proposes novel network features derived from a relevance network, which enable any machine learning algorithm to predict causal regulatory edges and infer GRNs. We evaluated CICT using simulated and experimental scRNA-seq data in a well-established benchmarking pipeline and showed that CICT outperformed existing network inference methods representing diverse approaches with many-fold higher accuracy. Furthermore, we demonstrated that GRN inference with CICT was robust to different levels of sparsity in scRNA-seq data, the characteristics of data and ground truth, the choice of association measure and the complexity of the supervised machine learning algorithm. Our results suggest aiming at directly predicting causality to recover regulatory relationships in complex biological networks substantially improves accuracy in GRN inference.

Proteomics: Progress and Promise of High-Throughput Proteomics in Chronic Kidney Disease.

Current proteomic tools permit the high-throughput analysis of the blood proteome in large cohorts, including those enriched for chronic kidney disease (CKD) or its risk factors. To date, these studies have identified numerous proteins associated with cross-sectional measures of kidney function, as well as with the longitudinal risk of CKD progression. Representative signals that have emerged from the literature include an association between levels of testican-2 and favorable kidney prognosis and an association between levels of TNFRSF1A and TNFRSF1B and worse kidney prognosis. For these and other associations, however, understanding whether the proteins play a causal role in kidney disease pathogenesis remains a fundamental challenge, especially given the strong impact that kidney function can have on blood protein levels. Prior to investing in dedicated animal models or randomized trials, methods that leverage the availability of genotyping in epidemiologic cohorts-including Mendelian randomization, colocalization analyses, and proteome-wide association studies-can add evidence for causal inference in CKD proteomics research. In addition, integration of large-scale blood proteome analyses with urine and tissue proteomics, as well as improved assessment of posttranslational protein modifications (e.g., carbamylation), represent important future directions. Taken together, these approaches seek to translate progress in large-scale proteomic profiling into the promise of improved diagnostic tools and therapeutic target identification in kidney disease.

Mendelian randomization for causal inference accounting for pleiotropy and sample structure using genome-wide summary statistics.

Mendelian randomization (MR) is a valuable tool for inferring causal relationships among a wide range of traits using summary statistics from genome-wide association studies (GWASs). Existing summary-level MR methods often rely on strong assumptions, resulting in many false-positive findings. To relax MR assumptions, ongoing research has been primarily focused on accounting for confounding due to pleiotropy. Here, we show that sample structure is another major confounding factor, including population stratification, cryptic relatedness, and sample overlap. We propose a unified MR approach, MR-APSS, which 1) accounts for pleiotropy and sample structure simultaneously by leveraging genome-wide information; and 2) allows the inclusion of more genetic variants with moderate effects as instrument variables (IVs) to improve statistical power without inflating type I errors. We first evaluated MR-APSS using comprehensive simulations and negative controls and then applied MR-APSS to study the causal relationships among a collection of diverse complex traits. The results suggest that MR-APSS can better identify plausible causal relationships with high reliability. In particular, MR-APSS can perform well for highly polygenic traits, where the IV strengths tend to be relatively weak and existing summary-level MR methods for causal inference are vulnerable to confounding effects.