An introduction to causal inference

From explainable to interpretable deep learning for natural language processing in healthcare: How far from reality?

Deep learning (DL) has substantially enhanced natural language processing (NLP) in healthcare research. However, the increasing complexity of DL-based NLP necessitates transparent model interpretability, or at least explainability, for reliable decision-making. This work presents a thorough scoping review of explainable and interpretable DL in healthcare NLP. The term "eXplainable and Interpretable Artificial Intelligence" (XIAI) is introduced to distinguish XAI from IAI. Different models are further categorized based on their functionality (model-, input-, output-based) and scope (local, global). Our analysis shows that attention mechanisms are the most prevalent emerging IAI technique. The use of IAI is growing, distinguishing it from XAI. The major challenges identified are that most XIAI does not explore "global" modelling processes, the lack of best practices, and the lack of systematic evaluation and benchmarks. One important opportunity is to use attention mechanisms to enhance multi-modal XIAI for personalized medicine. Additionally, combining DL with causal logic holds promise. Our discussion encourages the integration of XIAI in Large Language Models (LLMs) and domain-specific smaller models. In conclusion, XIAI adoption in healthcare requires dedicated in-house expertise. Collaboration with domain experts, end-users, and policymakers can lead to ready-to-use XIAI methods across NLP and medical tasks. While challenges exist, XIAI techniques offer a valuable foundation for interpretable NLP algorithms in healthcare.

AI analysis of super-resolution microscopy: Biological discovery in the absence of ground truth

Super-resolution microscopy, or nanoscopy, enables the use of fluorescent-based molecular localization tools to study molecular structure at the nanoscale level in the intact cell, bridging the mesoscale gap to classical structural biology methodologies. Analysis of super-resolution data by artificial intelligence (AI), such as machine learning, offers tremendous potential for the discovery of new biology, that, by definition, is not known and lacks ground truth. Herein, we describe the application of weakly supervised paradigms to super-resolution microscopy and its potential to enable the accelerated exploration of the nanoscale architecture of subcellular macromolecules and organelles.

Integrating Multi-Omics with environmental data for precision health: A novel analytic framework and case study on prenatal mercury induced childhood fatty liver disease

BACKGROUND

Precision Health aims to revolutionize disease prevention by leveraging information across multiple omic datasets (multi-omics). However, existing methods generally do not consider personalized environmental risk factors (e.g., environmental pollutants).

OBJECTIVE

To develop and apply a precision health framework which combines multiomic integration (including early, intermediate, and late integration, representing sequential stages at which omics layers are combined for modeling) with mediation approaches (including high-dimensional mediation to identify biomarkers, mediation with latent factors to identify pathways, and integrated/quasi-mediation to identify high-risk subpopulations) to identify novel biomarkers of prenatal mercury induced metabolic dysfunction-associated fatty liver disease (MAFLD), elucidate molecular pathways linking prenatal mercury with MAFLD in children, and identify high-risk children based on integrated exposure and multiomics data.

METHODS

This prospective cohort study used data from 420 mother-child pairs from the Human Early Life Exposome (HELIX) project. Mercury concentrations were determined in maternal or cord blood from pregnancy. Cytokeratin 18 (CK-18; a MAFLD biomarker) and five omics layers (DNA Methylation, gene transcription, microRNA, proteins, and metabolites) were measured in blood in childhood (age 6-10 years).

RESULTS

Each standard deviation increase in prenatal mercury was associated with a 0.11 [95% confidence interval: 0.02-0.21] standard deviation increase in CK-18. High dimensional mediation analysis identified 10 biomarkers linking prenatal mercury and CK-18, including six CpG sites and four transcripts. Mediation with latent factors identified molecular pathways linking mercury and MAFLD, including altered cytokine signaling and hepatic stellate cell activation. Integrated/quasi-mediation identified high risk subgroups of children based on unique combinations of exposure levels, omics profiles (driven by epigenetic markers), and MAFLD.

CONCLUSIONS

Prenatal mercury exposure is associated with elevated liver enzymes in childhood, likely through alterations in DNA methylation and gene expression. Our analytic framework can be applied across many different fields and serve as a resource to help guide future precision health investigations.

Propensity Weighted federated learning for treatment effect estimation in distributed imbalanced environments

Estimating treatment effects from observational data in medicine using causal inference is a very relevant task due to the abundance of observational data and the ethical and cost implications of conducting randomized experiments or experimental interventions. However, how could we estimate the effect of a treatment in a hospital that has very restricted access to treatment? In this paper, we want to address the problem of distributed causal inference, where hospitals not only have different distributions of patients, but also different treatment assignment criteria. Furthermore, it is necessary to take into account that due to privacy restrictions, personal patient data cannot be shared between hospitals. To address this problem, we propose an adaptation of the federated learning algorithm FederatedAveraging to one of the most advanced models for the prediction of treatment effects based on neural networks, TEDVAE. Our algorithm adaptation takes into account the shift in the treatment distribution between hospitals and is therefore called Propensity WeightedFederatedAveraging (PW FedAvg). As the distributions of the assignment of treatments become more unbalanced between the nodes, the estimation of causal effects becomes more challenging. The experiments show that PW FedAvg manages to reduce errors in the estimation of individual causal effects when imbalances are large, compared to VanillaFedAvg and other federated learning-based causal inference algorithms based on the application of federated learning to linear parametric models, Gaussian Processes and Random Fourier Features.

Statistical characteristics of analytical studies published in Peruvian medical journals from 2021 to 2022: A methodological study

OBJECTIVE

While statistical analysis plays a crucial role in medical science, some published studies might have utilized suboptimal analysis methods, potentially undermining the credibility of their findings. Critically appraising analytical approaches can help elevate the standard of evidence and ensure clinicians and other stakeholders have trustworthy results on which to base decisions. The aim of the present study was to examine the statistical characteristics of original articles published in Peruvian medical journals in 2021-2022.

DESIGN AND SETTING

We performed a methodological study of articles published between 2021 and 2022 from nine medical journals indexed in SciELO-Peru, Scopus, and Medline. We included original articles that conducted analytical analyses (i.e., association between variables). The statistical variables assessed were: statistical software used for analysis, sample size, and statistical methods employed (measures of effect), controlling for confounders, and the method employed for confounder control or epidemiological approaches.

RESULTS

We included 313 articles (ranging from 11 to 77 across journals), of which 67.7% were cross-sectional studies. While 90.7% of articles specified the statistical software used, 78.3% omitted details on sample size calculation. Descriptive and bivariate statistics were commonly employed, whereas measures of association were less common. Only 13.4% of articles (ranging from 0% to 39% across journals) presented measures of effect controlling for confounding and explained the criteria for selecting such confounders.

CONCLUSION

This study revealed important statistical deficiencies within analytical studies published in Peruvian journals, including inadequate reporting of sample sizes, absence of measures of association and confounding control, and suboptimal explanations regarding the methodologies employed for adjusted analyses. These findings highlight the need for better statistical reporting and researcher-editor collaboration to improve the quality of research production and dissemination in Peruvian journals.

Use and reporting of inverse-probability-of-treatment weighting for multicategory treatments in medical research: a systematic review

OBJECTIVES

Causal inference methods for observational data represent an alternative to randomised controlled trials when they are not feasible or when real-world evidence is sought. Inverse-probability-of-treatment weighting (IPTW) is one of the most popular approaches to account for confounding in observational studies. In medical research, IPTW is mainly applied to estimate the causal effect of a binary treatment, even when the treatment has in fact multiple categories, despite the availability of IPTW estimators for multiple treatment categories. This raises questions about the appropriateness of the use of IPTW in this context. Therefore, we conducted a systematic review of medical publications reporting the use of IPTW in the presence of a multi-category treatment. Our objectives were to investigate the frequency of use and the implementation of these methods in practice, and to assess the quality of their reporting.

STUDY DESIGN AND SETTING

Using Pubmed, Embase and Web of Science, we screened 5660 articles and retained 106 articles in the final analysis that were from 17 different medical areas. This systematic review is registered on PROSPERO (CRD42022352669).

RESULTS

The number of treatment groups varied between 3 and 9, with a large majority of articles (90 [84.9%]) including 3 or 4 groups. The most commonly used method for estimating the weights was multinomial regression (51 [48.1%]) and generalized boosted models (48 [45.3%]). The covariates of the weight model were reported in 91 articles (85.9 %). Twenty-six articles (24.5 %) did not discuss the balance of covariates after weighting, and only 16 articles (15.1 %) referred to the assumptions needed to obtain correct inferences.

CONCLUSION

The results of this systematic review illustrate that medical publications scarcely use IPTW methods for more than two treatment categories. Among the publications that did, the quality of reporting was suboptimal, in particular in regard to the assumptions and model building. IPTW for multi-category treatments could be applied more broadly in medical research, and the application of the proposed guidelines in this context will help researchers to report their results and to ensure reproducibility of their research.

Navigating the Complexities of Artificial Intelligence-Enabled Real-World Data Collection for Oncology Pharmacovigilance

This new editorial discusses the promise and challenges of successful integration of natural language processing methods into electronic health records for timely, robust, and fair oncology pharmacovigilance.

Mechanism of nearshore sediment-facilitated oil transport: New insights from causal inference analysis

A quantitative understanding of spilled oil transport in a nearshore environment is challenging due to the complex physicochemical processes in aqueous conditions. The physicochemical processes involved in oil sinking mainly include oil dispersion, sediment settling, and oil-sediment interaction. For the first time, this work attempts to address the sinking mechanism in petroleum contaminant transport using structural causal models based on observed data. The effects of nearshore salinity distribution from the estuary to the ocean on those three processes are examined. The causal inference reveals sediment settling is the crucial process for oil sinking. Salinity indirectly affects oil sinking by promoting sediment settling rather than directly affecting oil-sediment interaction. The increase of salinity from 0‰ to 35‰ provides a natural enhancement for sediment settling. Notably, unbiased causal effect estimates demonstrate the strongest positive causal effect on the settling efficiency of sediments is posed by increasing oil dispersion effectiveness, with a normalized value of 1.023. The highest strength of the causal relationship between oil dispersion and sediment settling highlights the importance of the dispersing characteristics of spilled oil to sediment-facilitated oil transport. The employed logic, a data-driven method, will shed light on adopting advanced causal inference tools to unravel the complicated contaminants' transport.

Evaluating the heterogeneous effect of extended culture to blastocyst transfer on the implantation outcome via causal inference in fresh ICSI cycles

PURPOSE

In IVF treatments, extended culture to single blastocyst transfer is the recommended protocol over cleavage-stage transfer. However, evidence-based criteria for assessing the heterogeneous implications on implantation outcomes are lacking. The purpose of this work is to estimate the causal effect of blastocyst transfer on implantation outcome.

METHODS

We fit a causal forest model using a multicenter observational dataset that includes an exogenous source of variability in treatment assignment and has a strong claim for satisfying the assumptions needed for valid causal inference from observational data.

RESULTS

We quantified the probability difference in embryo implantation if transferred as a blastocyst versus cleavage stage. Blastocyst transfer increased the average implantation rate; however, we revealed a subpopulation of embryos whose implantation potential is predicted to increase via cleavage-stage transfer.

CONCLUSION

Relative to the current policy, the proposed embryo transfer policy retrospectively improves implantation rate from 0.2 to 0.27. Our work demonstrates the efficacy of implementing causal inference in reproductive medicine and motivates its utilization in medical disciplines that are dominated by retrospective datasets.

Association of antibiotic treatment with survival outcomes in treatment-naïve melanoma patients receiving immune checkpoint blockade

PURPOSE

The interaction of gut microbiome and immune system is being studied with increasing interest. Disturbing factors, such as antibiotics may impact the immune system via gut and interfere with tumor response to immune checkpoint blockade (ICB).

METHODS

In this multicenter retrospective cohort study exclusively treatment-naïve patients with cutaneous or mucosal melanoma treated with first-line anti-PD-1 based ICB for advanced, non-resectable disease between 06/2013 and 09/2018 were included. Progression-free (PFS), and overall survival (OS) according to antibiotic exposure (within 60 days prior to ICB and after the start of ICB vs. no antibiotic exposure) were analyzed. To account for immortal time bias, data from patients with antibiotics during ICB were analyzed separately in the time periods before and after start of antibiotics.

RESULTS

Among 578 patients with first-line anti-PD1 based ICB, 7% of patients received antibiotics within 60 days prior to ICB and 19% after starting ICB. Antibiotic exposure prior to ICB was associated with worse PFS (adjusted HR 1.75 [95% CI 1.22-2.52]) and OS (adjusted HR 1.64 [95% CI 1.04-2.58]) by multivariate analysis adjusting for potential confounders. The use of antibiotics after the start of ICB had no effect on either PFS (adjusted HR 1.19; 95% CI 0.89-1.60) or OS (adjusted HR 1.08; 95% CI 0.75-1.57).

CONCLUSIONS

Antibiotic exposure within 60 days prior to ICB seems to be associated with worse PFS and OS in melanoma patients receiving first-line anti-PD1 based therapy, whereas antibiotics after the start of ICB do not appear to affect PFS or OS.

Assessing the heterogeneity of the impact of COVID-19 incidence on all-cause excess mortality among healthcare districts in Lombardy, Italy, to evaluate the local response to the pandemic: an ecological study

OBJECTIVES

The fragmentation of the response to the COVID-19 pandemic at national, regional and local levels is a possible source of variability in the impact of the pandemic on society. This study aims to assess how much of this variability affected the burden of COVID-19, measured in terms of all-cause 2020 excess mortality.

DESIGN

Ecological retrospective study.

SETTING

Lombardy region of Italy, 2015-2020.

OUTCOME MEASURES

We evaluated the relationship between the intensity of the epidemics and excess mortality, assessing the heterogeneity of this relationship across the 91 districts after adjusting for relevant confounders.

RESULTS

The epidemic intensity was quantified as the COVID-19 hospitalisations per 1000 inhabitants. Five confounders were identified through a directed acyclic graph: age distribution, population density, pro-capita gross domestic product, restriction policy and population mobility.Analyses were based on a negative binomial regression model with district-specific random effects. We found a strong, positive association between COVID-19 hospitalisations and 2020 excess mortality (p<0.001), estimating that an increase of one hospitalised COVID-19 patient per 1000 inhabitants resulted in a 15.5% increase in excess mortality. After adjusting for confounders, no district differed in terms of COVID-19-unrelated excess mortality from the average district. Minimal heterogeneity emerged in the district-specific relationships between COVID-19 hospitalisations and excess mortality (6 confidence intervals out of 91 did not cover the null value).

CONCLUSIONS

The homogeneous effect of the COVID-19 spread on the excess mortality in the Lombardy districts suggests that, despite the unprecedented conditions, the pandemic reactions did not result in health disparities in the region.

Effects of Platelet Count on Blood Pressure: Evidence from Observational and Genetic Investigations

Platelet count has been associated with blood pressure, but whether this association reflects causality remains unclear. To strengthen the evidence, we conducted a traditional observational analysis in the Lifelines Cohort Study (n = 167,785), and performed bi-directional Mendelian randomization (MR) with summary GWAS data from the UK Biobank (n = 350,475) and the International Consortium of Blood Pressure (ICBP) (n = 299,024). Observational analyses showed positive associations between platelet count and blood pressure (OR = 1.12 per SD, 95% CI: 1.10 to 1.14 for hypertension; B = 0.07, 95% CI: 0.07 to 0.08 for SBP; B = 0.07 per SD, 95% CI: 0.06 to 0.07 for DBP). In MR, a genetically predicted higher platelet count was associated with higher SBP (B = 0.02 per SD, 95% CI = 0.00 to 0.04) and DBP (B = 0.03 per SD, 95% CI = 0.01 to 0.05). IVW models and sensitivity analyses of the association between platelet count and DBP were consistent, but not all sensitivity analyses were statistically significant for the platelet count-SBP relation. Our findings indicate that platelet count has modest but significant effects on SBP and DBP, suggesting causality and providing further insight into the pathophysiology of hypertension.

Mendelian randomization for nephrologists

Confounding is a major limitation of observational studies. Mendelian randomization (MR) is a powerful study design that uses genetic variants as instrumental variables to enable examination of the causal effect of an exposure on an outcome in observational data. With the emergence of large-scale genome-wide association studies in nephrology over the past decade, MR has become a popular method to establish causal inferences. However, MR is a complex and challenging methodology that requires careful consideration to ensure robust results. This review article aims to summarize the basic concepts of MR, its application and relevance in nephrology, and the methodological challenges and limitations as well as discuss the current guidelines for design and reporting. With reference to a clinically relevant example of examining the causal relationship between the estimated glomerular filtration rate and cancer, this review outlines the key steps to conducting an MR study, including the key considerations and potential pitfalls at each step. These include defining the clinical question, selecting the data sources, identifying and refining appropriate genetic variants by considering linkage disequilibrium and associations with potential confounders, harmonization of variants across data sets, validation of the genetic instrument by assessing its strength, estimation of the causal effects, confirming the validity of the findings, and interpreting and reporting results.

A new causal centrality measure reveals the prominent role of subcortical structures in the causal architecture of the extended default mode network

Network representation has been an incredibly useful concept for understanding the behavior of complex systems in social sciences, biology, neuroscience, and beyond. Network science is mathematically founded on graph theory, where nodal importance is gauged using measures of centrality. Notably, recent work suggests that the topological centrality of a node should not be over-interpreted as its dynamical or causal importance in the network. Hence, identifying the influential nodes in dynamic causal models (DCM) remains an open question. This paper introduces causal centrality for DCM, a dynamics-sensitive and causally-founded centrality measure based on the notion of intervention in graphical models. Operationally, this measure simplifies to an identifiable expression using Bayesian model reduction. As a proof of concept, the average DCM of the extended default mode network (eDMN) was computed in 74 healthy subjects. Next, causal centralities of different regions were computed for this causal graph, and compared against several graph-theoretical centralities. The results showed that the subcortical structures of the eDMN were more causally central than the cortical regions, even though the graph-theoretical centralities unanimously favored the latter. Importantly, model comparison revealed that only the pattern of causal centrality was causally relevant. These results are consistent with the crucial role of the subcortical structures in the neuromodulatory systems of the brain, and highlight their contribution to the organization of large-scale networks. Potential applications of causal centrality-to study causal models of other neurotypical and pathological functional networks-are discussed, and some future lines of research are outlined.

Voxel-based analysis: Roadmap for clinical translation

Voxel-based analysis (VBA) allows the full, 3-dimensional, dose distribution to be considered in radiotherapy outcome analysis. This provides new insights into anatomical variability of pathophysiology and radiosensitivity by removing the need for a priori definition of organs assumed to drive the dose response associated with patient outcomes. This approach may offer powerful biological insights demonstrating the heterogeneity of the radiobiology across tissues and potential associations of the radiotherapy dose with further factors. As this methodological approach becomes established, consideration needs to be given to translating VBA results to clinical implementation for patient benefit. Here, we present a comprehensive roadmap for VBA clinical translation. Technical validation needs to demonstrate robustness to methodology, where clinical validation must show generalisability to external datasets and link to a plausible pathophysiological hypothesis. Finally, clinical utility requires demonstration of potential benefit for patients in order for successful translation to be feasible. For each step on the roadmap, key considerations are discussed and recommendations provided for best practice.

Correlation of serum biochemical parameters and saliva pH in healthy individuals

Saliva has the potential to work alongside needles in standard medical diagnosis. Yet the number of studies aimed at deciphering the biochemical communication between saliva and the rest of the body's systems is still very limited. The aim of this study is to investigate the interfluid interaction between saliva and serum by determining the correlation between saliva pH and serum biochemical parameters under mild conditions. Ultimately, using saliva may provide a stress-free diagnostic tool, but more ambitiously, the pH of saliva could present a genuine cost-effective screening tool that may immensely benefit areas with limited access to health care and diagnostic labs. Saliva and blood samples were collected from 43 randomly selected children (7-12 years), living in Jeddah, free from obesity and chronic or systemic body and mouth diseases. A complete serum biochemical analysis was performed, and the salivary pH of all samples was measured immediately at the time of collection. The correlations between saliva pH and serum biochemical parameters were investigated using Univariate and multiple linear regression models. Our results showed that pH has a weak significant positive correlation with total protein and a negative weak significant correlation with urea. Weak correlations suggest the existence of more serum factors to be investigated for their effect on the pH using a stepwise multiple linear regression. The multiple linear models' calculated saliva pH values were close to the measured values, demonstrating its possible capacity to predict saliva pH using serum parameters. The regression model's successful prediction of saliva pH using serum biochemicals reflects the significant correlations between the body fluids' parameters and invites more research to elucidate these relationships.

A Latent Change Score Approach to Understanding Chronic Bodily Pain Outcomes Following Knee Arthroplasty: A Secondary Analysis of Longitudinal Data

BACKGROUND

The extent to which chronic bodily pain changes following total knee arthroplasty (TKA) is unknown. We determined the extent of chronic bodily pain changes at 1 year following TKA.

METHODS

Data from our randomized trial of pain coping skills, which revealed no effect of the studied interventions, were used. The presence and severity of chronic pain in 16 body regions, excluding the surgically treated knee, were determined prior to and 1 year following surgery. The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain scale was used to quantify the extent of surgical knee pain. Latent change score (LCS) models were used to determine the extent to which true chronic bodily pain scores change after TKA.

RESULTS

The mean age of the sample of 367 participants was 63.4 ± 8.0 years, and 247 (67%) were female. LCS analyses showed significant 20% to 54% reductions in pain in the surgically treated lower limb (not including the surgically treated knee), pain in the non-surgically treated lower limb, and whole body pain. In bivariate LCS analyses, greater improvement in the WOMAC pain score, indicating surgical benefit of TKA, led to greater improvement in all 4 bodily pain areas beyond the surgically treated knee, even after controlling for the latent change in pain catastrophizing.

CONCLUSIONS

Clinically important chronic bodily pain reductions occurred following TKA and may be causally linked to the surgical procedure. Reduction in chronic bodily pain in sites other than the surgically treated knee is an additional benefit of TKA.

LEVEL OF EVIDENCE

Prognostic Level II . See Instructions for Authors for a complete description of levels of evidence.

Reverse causation biases weighted cumulative exposure model estimates, but can be investigated in sensitivity analyses

OBJECTIVES

To examine the effects of reverse causation on estimates from the weighted cumulative exposure (WCE) model that is used in pharmacoepidemiology to explore drug-health outcome associations, and to identify sensitivity analyses for revealing such effects.

STUDY DESIGN AND SETTING

314,099 patients with diabetes under Clalit Health Services, Israel, were followed over 2002-2012. The association between metformin and pancreatic cancer (PC) was explored using a WCE model within the framework of discrete-time Cox regression. We used computer simulations to explore the effects of reverse causation on estimates of a WCE model and to examine sensitivity analyses for revealing and adjusting for reverse causation. We then applied those sensitivity analyses to our data.

RESULTS

Simulation demonstrated bias in the weighted cumulative exposure model and showed that sensitivity analysis could reveal and adjust for these biases. In our data, a positive association was observed (hazard ratio (HR) = 3.24, 95% confidence interval (CI): 2.24-4.73) with metformin exposure in the previous 2 years. After applying sensitivity analysis, assuming reverse causation operated up to 4 years before cancer diagnosis, the association between metformin and PC was no longer apparent.

CONCLUSION

Reverse causation can cause substantial bias in the WCE model. When suspected, sensitivity analyses based on causal analysis are advocated.

Mammogram Uptake from Social Determinants of Health Can Be Lost in Translation to Individual Patients

Purpose

The objective of this study is to describe patterns in barriers to breast cancer screening uptake with the end goal of improving screening adherence and decreasing the burden of mortality due to breast cancer. This study looks at social determinants of health and their association to screening and mortality. It also investigates the extent that models trained on county data are generalizable to individuals.

Methods

County level screening uptake and age adjusted mortality due to breast cancer are combined with the Centers for Disease Controls Social Vulnerability Index (SVI) to train a model predicting screening uptake rates. Patterns learned are then applied to de-identified electronic medical records from individual patients to make predictions on mammogram screening follow through.

Results

Accurate predictions can be made about a county's breast cancer screening uptake with the SVI. However, the association between increased screening, and decreased age adjusted mortality, doesn't hold in areas with a high proportion of minority residents. It is also shown that patterns learned from county SVI data have little discriminative power at the patient level.

Conclusion

This study demonstrates that social determinants in the SVI can explain much of the variance in county breast cancer screening rates. However, these same patterns fail to discriminate which patients will have timely follow through of a mammogram screening test. This study also concludes that the core association between increased screening and decreased age adjusted mortality does not hold in high proportion minority areas.

Reproduction patterns among non-Hodgkin lymphoma survivors by subtype in Sweden, Denmark and Norway: A population-based matched cohort study

Previous studies concerning reproductive patterns among non-Hodgkin lymphoma (NHL) survivors are scarce and those available have reported conflicting results. Treatment regimens vary considerably between aggressive and indolent NHL and studies of reproductive patterns by subtypes are warranted. In this matched cohort study, we identified all NHL patients aged 18-40 years and diagnosed between 2000 and 2018 from the Swedish and Danish lymphoma registers, and the clinical database at Oslo University Hospital (n = 2090). Population comparators were matched on sex, birth year and country (n = 19 427). Hazard ratios (HRs) were estimated using Cox regression. Males and females diagnosed with aggressive lymphoma subtypes had lower childbirth rates (HRfemale : 0.43, 95% CI: 0.31-0.59, HRmale : 0.61, 95% CI: 0.47-0.78) than comparators during the first 3 years after diagnosis. For indolent lymphomas, childbirth rates were not significantly different from comparators (HRfemale : 0.71, 95% CI: 0.48-1.04, HRmale : 0.94, 95% CI: 0.70-1.27) during the same period. Childbirth rates reached those of comparators for all subtypes after 3 years but the cumulative incidence of childbirths was decreased throughout the 10-year follow-up for aggressive NHL. Children of NHL patients were more likely to be born following assisted reproductive technology than those of comparators, except for male indolent lymphoma patients. In conclusion, fertility counselling is particularly important for patients with aggressive NHL.

Interaction between travel restriction policies and the spread of COVID-19

To investigate the interaction between travel restriction policies and the spread of COVID-19, we collected data on human mobility trends, population density, Gross Domestic Product (GDP) per capita, daily new confirmed cases (or deaths), and the total confirmed cases (or deaths), as well as governmental travel restriction policies from 33 countries. The data collection period was from April 2020 to February 2022, resulting in 24,090 data points. We then developed a structural causal model to describe the causal relationship between these variables. Using the Dowhy method to solve the developed model, we found several significant results that passed the refutation test. Specifically, travel restriction policies played an important role in slowing the spread of COVID-19 until May 2021. International travel controls and school closures had an impact on reducing the spread of the pandemic beyond the impact of travel restrictions. Additionally, May 2021 marked a turning point in the spread of COVID-19 as it became more infectious, but the mortality rate gradually decreased. The impact of travel restriction policies on human mobility and the pandemic diminished over time. Overall, the cancellation of public events and restrictions on public gatherings were more effective than other travel restriction policies. Our findings provide insights into the effects of travel restriction policies and travel behavioral changes on the spread of COVID-19, while controlling for informational and other confounding variables. This experience can be applied in the future to respond to emergent infectious diseases.

U-shaped association between low-density lipoprotein cholesterol levels and risk of all-cause mortality mediated by post-stroke infection in acute ischemic stroke

During the acute stage of ischemic stroke, it remains unclear how to interpret the low low-density lipoprotein cholesterol (LDL-C) level. We aimed to evaluate the association between LDL-C levels, post-stroke infection, and all-cause mortality. 804,855 ischemic stroke patients were included. Associations between LDL-C levels, infection, and mortality risk were estimated by multivariate logistic regression models and displayed by restricted cubic spline curves. Mediation analysis was performed under counterfactual framework to elucidate the mediation effect of post-stroke infection. The association between LDL-C and mortality risk was U-shaped. The nadir in LDL-C level with the lowest mortality risk was 2.67 mmol/L. Compared with the group with LDL-C = 2.50-2.99 mmol/L, the multivariable-adjusted odds ratio for mortality was 2.22 (95% confidence intervals (CI): 1.77-2.79) for LDL-C <1.0 mmol/L and 1.22 (95% CI: 0.98-1.50) for LDL-C ≥5.0 mmol/L. The association between LDL-C and all-cause mortality was 38.20% (95% CI: 5.96-70.45, P = 0.020) mediated by infection. After stepwise excluding patients with increasing numbers of cardiovascular risk factors, the U-shaped association between LDL-C and all-cause mortality and the mediation effects of infection remained consistent with the primary analysis, but the LDL-C interval with the lowest mortality risk increased progressively. The mediation effects of infection were largely consistent with the primary analysis in subgroups of age ≥65 years, female, body mass index <25 kg/m², and National Institutes of Health Stroke Scale ≥16. During the acute stage of ischemic stroke, there is a U-shaped association between LDL-C level and all-cause mortality, where post-stroke infection is an important mediating mechanism.

Simple design and analysis strategies for solving problems in observational orthopaedic clinical research

Randomized controlled trials are the gold standard to establishing causal relationships in clinical research. However, these studies are expensive and time consuming to conduct. This article aims to provide orthopaedic surgeons and clinical researchers with methodology to optimize inference and minimize bias in observational studies that are often much more feasible to undertake. To mitigate the risk of bias arising from their nonexperimental design, researchers must first understand the ways in which measured covariates can influence treatment, outcomes, and missingness of follow-up data. With knowledge of these relationships, researchers can then build causal diagrams to best understand how to control sources of bias. Some common techniques for controlling for bias include matching, regression, stratification, and propensity score analysis. Selection bias may result from loss to follow-up and missing data. Strategies such as multiple imputation and time-to-event analysis can be useful for handling missingness. For longitudinal data, repeated measures allow observational studies to best summarize the impact of the intervention over time. Clinical researchers familiar with fundamental concepts of causal inference and techniques reviewed in this article will have the power to improve the quality of inferences made from clinical research in orthopaedic trauma surgery.

New proposal to address mediation analysis interrogations by using genetic variants as instrumental variables

The application of causal mediation analysis (CMA) considering the mediation effect of a third variable is increasing in epidemiological studies; however, this requires fitting strong assumptions on confounding bias. To address this limitation, we propose an extension of CMA combining it with Mendelian randomization (MRinCMA). We applied the new approach to analyse the causal effect of obesity and diabetes on pancreatic cancer, considering each factor as potential mediator. To check the performance of MRinCMA under several conditions/scenarios, we used it in different simulated data sets and compared it with structural equation models. For continuous variables, MRinCMA and structural equation models performed similarly, suggesting that both approaches are valid to obtain unbiased estimates. When noncontinuous variables were considered, MRinCMA presented, overall, lower bias than structural equation models. By applying MRinCMA, we did not find any evidence of causality of obesity or diabetes on pancreatic cancer. With this new methodology, researchers would be able to address CMA hypotheses by appropriately accounting for the confounding bias assumption regardless of the conditions used in their studies in different settings.

A connection between survival multistate models and causal inference for external treatment interruptions

Recently, treatment interruptions such as a clinical hold in randomized clinical trials have been investigated by using a multistate model approach. The phase III clinical trial START (Stimulating Targeted Antigenic Response To non-small-cell cancer) with primary endpoint overall survival was temporarily placed on hold for enrollment and treatment by the US Food and Drug Administration (FDA). Multistate models provide a flexible framework to account for treatment interruptions induced by a time-dependent external covariate. Extending previous work, we propose a censoring and a filtering approach both aimed at estimating the initial treatment effect on overall survival in the hypothetical situation of no clinical hold. A special focus is on creating a link to causal inference. We show that calculating the matrix of transition probabilities in the multistate model after application of censoring (or filtering) yields the desired causal interpretation. Assumptions in support of the identification of a causal effect by censoring (or filtering) are discussed. Thus, we provide the basis to apply causal censoring (or filtering) in more general settings such as the COVID-19 pandemic. A simulation study demonstrates that both causal censoring and filtering perform favorably compared to a naïve method ignoring the external impact.

Atrial Fibrillation and the Risk of Ventricular Arrhythmias and Cardiac Arrest: A Nationwide Population-Based Study

BACKGROUND

Atrial fibrillation (AF) has been linked to an increased risk of ventricular arrhythmias (VAs) and sudden death. We investigated this association in hospitalised patients in France.

METHODS

All hospitalised patients from 2013 were identified from the French National database and included if they had at least 5 years of follow-up data.

RESULTS

Overall, 3,381,472 patients were identified. After excluding 35,834 with a history of VAs and cardiac arrest, 3,345,638 patients were categorised into two groups: no AF (n = 3,033,412; mean age 57.2 ± 21.4; 54.3% female) and AF (n = 312,226; 78.1 ± 10.6; 44.0% female). Over a median follow-up period of 5.4 years (interquartile range (IQR) 5.0-5.8 years), the incidence (2.23%/year vs. 0.56%/year) and risk (hazard ratio (HR) 3.657 (95% confidence interval (CI) 3.604-3.711)) of VAs and cardiac arrest were significantly higher in AF patients compared to non-AF patients. This was still significant after adjusting for confounders, with a HR of 1.167 (95% CI 1.111-1.226) and in the 1:1 propensity score-matched analysis (n = 289,332 per group), with a HR of 1.339 (95% CI 1.313-1.366). In the mediation analysis, the odds of cardiac arrest were significantly mediated by AF-associated VAs, with an OR of 1.041 (95% CI 1.040-1.042).

CONCLUSION

In hospitalised French patients, AF was associated with an increased risk of VAs and sudden death.

Birth order and morbidity and mortality to hospital discharge among inborn very low-birthweight, very preterm twin infants admitted to neonatal intensive care: a retrospective cohort study

OBJECTIVE

To know the association of birth order with the risk of morbidity and mortality in very low-birthweight (VLBW) twin infants less than 32 weeks' gestational age (GA).

DESIGN

Retrospective cohort study.

SETTING

Infants admitted to the collaborating centres of the Spanish SEN1500 neonatal network.

PATIENTS

Liveborn VLBW twin infants, with GA from 23⁺⁰ weeks to 31⁺⁶ weeks, without congenital anomalies, admitted from 2011 to 2020. Outborn patients were excluded.

MAIN OUTCOME MEASURES

Respiratory distress syndrome (RDS), patent ductus arteriosus, bronchopulmonary dysplasia (BPD), necrotising enterocolitis, major brain damage (MBD), late-onset neonatal sepsis, severe retinopathy of prematurity, survival and survival without morbidity. Crude and adjusted incidence rate ratios were calculated.

RESULTS

Among 2111 twin pairs included, the second twin had higher risk (adjusted risk ratio (aRR) of RDS (aRR 1.08, 95% CI 1.03 to 1.12) and need for surfactant (aRR1.10, 95% CI 1.05 to 1.16). No other significant differences were found, neither in survival (aRR 1.01, 95% CI 0.99 to 1.03) nor in survival without BPD (aRR 1.02, 95% CI 0.99 to 1.05), survival without MBD (aRR 1.02, 95% CI 0.99 to 1.06) nor in survival without major morbidity (aRR 0.97, 95% CI 0.92 to 1.03). However, second twins born by caesarean section (C-section) after a vaginally delivered first twin had less overall survival and survival without MBD.

CONCLUSION

In modern perinatology, second twins are still more unstable immediately after birth and require more resuscitation. After admission to the neonatal intensive care unit, they are at increased risk of RDS, but not other conditions, except for second twins delivered by C-section after a first twin delivered vaginally, who have decreased overall survival and survival without major brain injury.

Determining asthma endotypes and outcomes: Complementing existing clinical practice with modern machine learning

There is unprecedented opportunity to use machine learning to integrate high-dimensional molecular data with clinical characteristics to accurately diagnose and manage disease. Asthma is a complex and heterogeneous disease and cannot be solely explained by an aberrant type 2 (T2) immune response. Available and emerging multi-omics datasets of asthma show dysregulation of different biological pathways including those linked to T2 mechanisms. While T2-directed biologics have been life changing for many patients, they have not proven effective for many others despite similar biomarker profiles. Thus, there is a great need to close this gap to understand asthma heterogeneity, which can be achieved by harnessing and integrating the rich multi-omics asthma datasets and the corresponding clinical data. This article presents a compendium of machine learning approaches that can be utilized to bridge the gap between predictive biomarkers and actual causal signatures that are validated in clinical trials to ultimately establish true asthma endotypes.

Understanding and overcoming the technical challenges in using in silico predictions in regulatory decisions of complex toxicological endpoints - A pesticide perspective for regulatory toxicologists with a focus on machine learning models

There are many challenges that must be overcome before in silico toxicity predictions are ripe for regulatory decision-making. Today, mandates in the United States of America and the European Union to avoid animal usage in toxicity testing is driving the need to consider alternative technologies, including Quantitative Structure Activity Relationship (QSAR) models, and read across approaches. However, when adopting new methods, it is critical that both new approach developers as well as regulatory users understand the strengths and challenges with these new approaches. In this paper, we identify potential sources of bias in machine learning methods specific to toxicity predictions, that may impact the overall performance of in silico models. We also discuss ways to mitigate these biases. Based on our experiences, the most prevalent sources of bias include class imbalance (differing numbers of "toxic" vs "nontoxic" compounds), limited numbers of chemicals within a particular chemistry, and biases within the studies that make up the database used for model building, as well as model evaluation biases. While this is already complex for repeated dose toxicity, in reproduction and developmental toxicity a further level of complexity is introduced by the need to evaluate effects on individual animal and litter basis (e.g., a hierarchal structure). We also discuss key considerations developers and regulators need to make when they use machine learning models to predict chemical safety. Our objective is for our paper to serve as a desk reference for model developers and regulators as they evaluate machine learning models and as they make decisions using these models.

Targeting keystone species helps restore the dysbiosis of butyrate-producing bacteria in nonalcoholic fatty liver disease

The dysbiosis of the gut microbiome is one of the pathogenic factors of nonalcoholic fatty liver disease (NAFLD) and also affects the treatment and intervention of NAFLD. Among gut microbiomes, keystone species that regulate the integrity and stability of an ecological community have become the potential intervention targets for NAFLD. Here, we collected stool samples from 22 patients with nonalcoholic steatohepatitis (NASH), 25 obese patients, and 16 healthy individuals from New York for 16S rRNA gene sequencing. An algorithm was implemented to identify keystone species based on causal inference theories and dynamic intervention simulation. External validation was performed in an independent cohort from California. Eight keystone species in the gut of NAFLD, represented by Porphyromonas loveana, Alistipes indistinctus, and Dialister pneumosintes, were identified, which could efficiently restore the microbial composition of the NAFLD toward a normal gut microbiome with 92.3% recovery. These keystone species regulate intestinal amino acid metabolism and acid-base environment to promote the growth of the butyrate-producing Lachnospiraceae and Ruminococcaceae species that are significantly reduced in NAFLD patients. Our findings demonstrate the importance of keystone species in restoring the microbial composition toward a normal gut microbiome, suggesting a novel potential microbial treatment for NAFLD.

Causal inference under over-simplified longitudinal causal models

Many causal models of interest in epidemiology involve longitudinal exposures, confounders and mediators. However, repeated measurements are not always available or used in practice, leading analysts to overlook the time-varying nature of exposures and work under over-simplified causal models. Our objective is to assess whether - and how - causal effects identified under such misspecified causal models relates to true causal effects of interest. We derive sufficient conditions ensuring that the quantities estimated in practice under over-simplified causal models can be expressed as weighted averages of longitudinal causal effects of interest. Unsurprisingly, these sufficient conditions are very restrictive, and our results state that the quantities estimated in practice should be interpreted with caution in general, as they usually do not relate to any longitudinal causal effect of interest. Our simulations further illustrate that the bias between the quantities estimated in practice and the weighted averages of longitudinal causal effects of interest can be substantial. Overall, our results confirm the need for repeated measurements to conduct proper analyses and/or the development of sensitivity analyses when they are not available.

Common permutation methods in animal social network analysis do not control for non-independence

The non-independence of social network data is a cause for concern among behavioural ecologists conducting social network analysis. This has led to the adoption of several permutation-based methods for testing common hypotheses. One of the most common types of analysis is nodal regression, where the relationships between node-level network metrics and nodal covariates are analysed using a permutation technique known as node-label permutations. We show that, contrary to accepted wisdom, node-label permutations do not automatically account for the non-independences assumed to exist in network data, because regression-based permutation tests still assume exchangeability of residuals. The same assumption also applies to the quadratic assignment procedure (QAP), a permutation-based method often used for conducting dyadic regression. We highlight that node-label permutations produce the same p-values as equivalent parametric regression models, but that in the presence of non-independence, parametric regression models can also produce accurate effect size estimates. We also note that QAP only controls for a specific type of non-independence between edges that are connected to the same nodes, and that appropriate parametric regression models are also able to account for this type of non-independence. Based on this, we suggest that standard parametric models could be used in the place of permutation-based methods. Moving away from permutation-based methods could have several benefits, including reducing over-reliance on p-values, generating more reliable effect size estimates, and facilitating the adoption of causal inference methods and alternative types of statistical analysis.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00265-022-03254-x.

Knowledge of the perturbation design is essential for accurate gene regulatory network inference

The gene regulatory network (GRN) of a cell executes genetic programs in response to environmental and internal cues. Two distinct classes of methods are used to infer regulatory interactions from gene expression: those that only use observed changes in gene expression, and those that use both the observed changes and the perturbation design, i.e. the targets used to cause the changes in gene expression. Considering that the GRN by definition converts input cues to changes in gene expression, it may be conjectured that the latter methods would yield more accurate inferences but this has not previously been investigated. To address this question, we evaluated a number of popular GRN inference methods that either use the perturbation design or not. For the evaluation we used targeted perturbation knockdown gene expression datasets with varying noise levels generated by two different packages, GeneNetWeaver and GeneSpider. The accuracy was evaluated on each dataset using a variety of measures. The results show that on all datasets, methods using the perturbation design matrix consistently and significantly outperform methods not using it. This was also found to be the case on a smaller experimental dataset from E. coli. Targeted gene perturbations combined with inference methods that use the perturbation design are indispensable for accurate GRN inference.

Global Change and Emerging Infectious Diseases

Our world is undergoing rapid planetary changes driven by human activities, often mediated by economic incentives and resource management, affecting all life on Earth. Concurrently, many infectious diseases have recently emerged or spread into new populations. Mounting evidence suggests that global change-including climate change, land-use change, urbanization, and global movement of individuals, species, and goods-may be accelerating disease emergence by reshaping ecological systems in concert with socioeconomic factors. Here, we review insights, approaches, and mechanisms by which global change drives disease emergence from a disease ecology perspective. We aim to spur more interdisciplinary collaboration with economists and identification of more effective and sustainable interventions to prevent disease emergence. While almost all infectious diseases change in response to global change, the mechanisms and directions of these effects are system specific, requiring new, integrated approaches to disease control that recognize linkages between environmental and economic sustainability and human and planetary health.

Application of Inverse-Probability-of-Treatment Weighting to Estimate the Effect of Daytime Sleepiness in Patients with Obstructive Sleep Apnea

Rationale: Continuous positive airway pressure (CPAP), the first line therapy for obstructive sleep apnea (OSA), is considered effective in reducing daytime sleepiness. Its efficacy relies on adequate adherence, often defined as >4 hours per night. However, this binary threshold may limit our understanding of the causal effect of CPAP adherence and daytime sleepiness, and a multilevel approach for CPAP adherence can be more appropriate. Objectives: In this study, we show how two causal inference methods can be applied on observational data for the estimation of the effect of different ranges of CPAP adherence on daytime sleepiness as measured by the Epworth Sleepiness Scale (ESS). Methods: Data were collected from a large prospective observational French cohort for patients with OSA. Four groups of CPAP adherence were considered (0-4, 4-6, 6-7, and 7-10 h per night). Multivariable regression, inverse-probability-of-treatment weighting (IPTW), and inverse propensity weighting with regression adjustment (IPW-RA) were used to assess the impact of CPAP adherence level on daytime sleepiness. Results: In this study, 9,244 patients with OSA treated by CPAP were included. The mean initial ESS score was 11 (±5.2), with a mean reduction of 4 points (±5.1). Overall, there was evidence of the causal effect of CPAP adherence on daytime sleepiness which was mainly observed between the lower CPAP adherence group (0-4 h) compared with the higher CPAP adherence group (7-10 h). There are no differences by considering higher level of CPAP adherence (>4 h). Conclusions: We showed that IPTW and IPW-RA can be easily implemented to answer questions regarding causal effects using observational data when randomized trials cannot be conducted. Both methods give a direct causal interpretation at the population level and allow the assessment of the appropriate consideration of measured confounders.

Negative cognitive schema modification as mediator of symptom improvement after electroconvulsive therapy in major depressive disorder

BACKGROUND

Electroconvulsive therapy (ECT) is a potent option for treatment-resistant major depressive disorder (MDD). Cognitive models of depression posit that negative cognitions and underlying all-or-nothing negative schemas contribute to and perpetuate depressed mood. This study investigates whether ECT can modify negative schemas, potentially via memory reactivation, and whether such changes are related to MDD symptom improvement.

METHOD

Seventy-two patients were randomized to either an emotional memory reactivation electroconvulsive therapy (EMR-ECT) or control memory reactivation electroconvulsive therapy (CMR-ECT) intervention prior to ECT-sessions in a randomized controlled trail. Emotional memories associated with patients' depression were reactivated before ECT-sessions. At baseline and after the ECT-course, negative schemas and depression severity were assessed using the Dysfunctional Attitude Scale (DAS) and Hamilton Depression Rating Scale HDRS. Mediation analyses were used to examine whether the effects of ECT on HDRS-scores were mediated by changes in DAS-scores or vice versa.

RESULTS

Post-ECT DAS-scores were significantly lower compared to baseline. Post-ECT, the mean HDRS-score of the whole sample (15.10 ± 8.65 [SD]; n = 59) was lower compared to baseline (24.83 ± 5.91 [SD]). Multiple regression analysis showed no significant influence of memory reactivation on schema improvement. Path analysis showed that depression improvement was mediated by improvement of negative cognitive schemas.

CONCLUSION

ECT is associated with improvement of negative schemas, which appears to mediate the improvement of depressive symptoms. An emotional memory intervention aimed to modify negative schemas showed no additional effect.

If you move, I move: The social influence effect on residential mobility

There are many theories that account for why households move between residential areas. In this paper, we advance on this by formulating a new mechanism whereby a household’s probability of leaving a neighborhood is informed by the number of other households who have previously left that neighborhood. We call this mechanism: the social influence (SI) effect. By applying matching to Swedish register data for Stockholm County (1998–2017), and after adjusting for theoretically relevant confounders from the existing literature, we find that SI has a significant effect on neighborhood out-mobility. Furthermore, we find that the SI effect is moderated by the visibility with which others’ behaviors is observed, measured as the number of previous out-movers, the distance to ego, and its salience in the social environment. Our study also discusses some ways in which SI might be entangled with other mechanisms, and outlines future directions from which studies of residential segregation dynamics might be approached.

Identification and robust estimation of swapped direct and indirect effects: Mediation analysis with unmeasured mediator-outcome confounding and intermediate confounding

Counterfactual-model-based mediation analysis can yield substantial insight into the causal mechanism through the assessment of natural direct effects (NDEs) and natural indirect effects (NIEs). However, the assumptions regarding unmeasured mediator-outcome confounding and intermediate mediator-outcome confounding that are required for the determination of NDEs and NIEs present practical challenges. To address this problem, we introduce an instrumental blocker, a novel quasi-instrumental variable, to relax both of these assumptions, and we define a swapped direct effect (SDE) and a swapped indirect effect (SIE) to assess the mediation. We show that the SDE and SIE are identical to the NDE and NIE, respectively, based on a causal interpretation. Moreover, the empirical expressions of the SDE and SIE are derived with and without an intermediate mediator-outcome confounder. Then, a multiply robust estimation method is derived to mitigate the model misspecification problem. We prove that the proposed estimator is consistent, asymptotically normal, and achieves the semiparametric efficiency bound. As an illustration, we apply the proposed method to genomic datasets of lung cancer to investigate the potential role of the epidermal growth factor receptor in the treatment of lung cancer.

Insights into the Profile of the Human Expiratory Microbiota and Its Associations with Indoor Microbiotas

Microorganisms residing in the human respiratory tract can be exhaled, and they constitute a part of environmental microbiotas. However, the expiratory microbiota community and its associations with environmental microbiotas remain poorly understood. Here, expiratory bacteria and fungi and the corresponding microbiotas from the living environments were characterized by DNA amplicon sequencing of residents' exhaled breath condensate (EBC) and environmental samples collected from 14 residences in Nanjing, China. The microbiotas of EBC samples, with a substantial heterogeneity, were found to be as diverse as those of skin, floor dust, and airborne microbiotas. Model fitting results demonstrated the role of stochastic processes in the assembly of the expiratory microbiota. Using a fast expectation-maximization algorithm, microbial community analysis revealed that expiratory microbiotas were differentially associated with other types of microbiotas in a type-dependent and residence-specific manner. Importantly, the expiratory bacteria showed a composition similarity with airborne bacteria in the bathroom and kitchen environments with an average of 12.60%, while the expiratory fungi showed a 53.99% composition similarity with the floor dust fungi. These differential patterns indicate different relationships between expiratory microbiotas and the airborne microbiotas and floor dust microbiotas. The results here illustrated for the first time the associations between expiratory microbiotas and indoor microbiotas, showing a potential microbial exchange between the respiratory tract and indoor environment. Thus, improved hygiene and ventilation practices can be implemented to optimize the indoor microbial exposome, especially in indoor bathrooms and kitchens.

Essential Regression: A generalizable framework for inferring causal latent factors from multi-omic datasets

High-dimensional cellular and molecular profiling of biological samples highlights the need for analytical approaches that can integrate multi-omic datasets to generate prioritized causal inferences. Current methods are limited by high dimensionality of the combined datasets, the differences in their data distributions, and their integration to infer causal relationships. Here, we present Essential Regression (ER), a novel latent-factor-regression-based interpretable machine-learning approach that addresses these problems by identifying latent factors and their likely cause-effect relationships with system-wide outcomes/properties of interest. ER can integrate many multi-omic datasets without structural or distributional assumptions regarding the data. It outperforms a range of state-of-the-art methods in terms of prediction. ER can be coupled with probabilistic graphical modeling, thereby strengthening the causal inferences. The utility of ER is demonstrated using multi-omic system immunology datasets to generate and validate novel cellular and molecular inferences in a wide range of contexts including immunosenescence and immune dysregulation.

Materials Data toward Machine Learning: Advances and Challenges

Machine learning (ML) is believed to have enabled a paradigm shift in materials research, and in practice, ML has demonstrated its power in speeding up the cost-efficient discovery of new materials and autonomizing materials laboratories. In this Perspective, current research progress in materials data which are the backbones of ML are reviewed, focusing on high-throughput data generation, standardized data storage, and data representation. More importantly, the challenging issues in materials data that should be overcome to unlock the full potential of ML in materials research and development, including classic 5V (volume, velocity, variety, veracity, and value) issues, 3M (multicomponent, multiscale, and multistage) challenges, co-mining of experimental and computational data, and materials data toward transferable/explainable ML or causal ML, are discussed.