PLOS Biology

The COVID-19 pandemic has brought substantial attention to the systems used to communicate biomedical research. In particular, the need to rapidly and credibly communicate research findings has led many stakeholders to encourage researchers to adopt open science practices such as posting preprints and sharing data. To examine the degree to which this has led to the adoption of such practices, we examined the "openness" of a sample of 539 published papers describing the results of randomized controlled trials testing interventions to prevent or treat COVID-19. The majority (56%) of the papers in this sample were free to read at the time of our investigation and 23.56% were preceded by preprints. However, there is no guarantee that the papers without an open license will be available without a subscription in the future, and only 49.61% of the preprints we identified were linked to the subsequent peer-reviewed version. Of the 331 papers in our sample with statements identifying if (and how) related datasets were available, only a paucity indicated that data was available in a repository that facilitates rapid verification and reuse. Our results demonstrate that, while progress has been made, there is still a significant mismatch between aspiration and the practice of open science in an important area of the COVID-19 literature. Open MaterialsWe are committed to making the details of our research process as open as possible. The data and code that underlie our analyses are archived and published through the Dryad Data Repository (https://doi.org/10.5061/dryad.mkkwh7137). Documentation and instructions for manuscript screening and data extraction are available on Protocols.io (https://dx.doi.org/10.17504/protocols.io.x54v9jx7zg3e/v1). Author contributions are outlined in Supplementary Table 1. O_TBL View this table: org.highwire.dtl.DTLVardef@774764org.highwire.dtl.DTLVardef@f03612org.highwire.dtl.DTLVardef@6e16ccorg.highwire.dtl.DTLVardef@19ac3eborg.highwire.dtl.DTLVardef@1b47f40_HPS_FORMAT_FIGEXP M_TBL O_FLOATNOSupplementary Table 1.C_FLOATNO O_TABLECAPTIONAuthor Information and Contributions C_TABLECAPTION C_TBL

Emergency Use Authorization for remdesivir months after discovery of COVID-19 is unprecedented. Typically, decades of research and public-sector funding are required to establish the mature body of foundational research requisite for efficient, targeted drug discovery and development. This work quantifies the body of research related to remdesivirs biological target, RNA-dependent RNA polymerase (RdRp), or parent chemical structure, nucleoside analogs (NcAn), through 2019, as well as NIH funding for this research 2000-2019. There were 6,567 RdRp-related publications in PubMed, including 1,263 with NIH support, and 11,073 NcAn-related publications, including 2,319 with NIH support. NIH support for RdRp research comprised 2,203 Project Years with Costs of $1,875 million. NIH support for NcAn research comprised 4,607 Project Years with Costs of $4,612 million. Research Project grants accounted for 63% and 48% of Project Years for RdRp and NcAn respectively, but only 19% and 12% of Project Costs. Analytical modeling of research maturation estimates that RdRp and NcAn research passed an established maturity threshold in 2008 and 1994 respectively. Of 97 investigational compounds targeting RdRp since 1989, the three authorized for use entered clinical trials after both thresholds. This work demonstrates the scale of foundational research on the biological target and parent chemical structure of remdesivir that supported its discovery and development for COVID-19. This work identifies $6.5 billion in NIH funding for research leading to remdesivir, underscoring the role of public sector investments in basic research and research infrastructure that underlie new drugs and the response to emergent disease. SIGNIFICANCE STATEMENTEmergency Use Authorization of remdesivir for treating COVID-19 four months after discovery of this virus was enabled by decades of research on the drugs biological target as well as other medicines with related chemical structures. The NIH contributed 6,800 years of grant funding to this research, totaling $6.5 billion (2000-2019), including funding for both investigator-initiated research and research infrastructure. Of this, $46.5 million was for research directly related to remdesivir. This analysis demonstrates the importance of a robust body of foundational research in responding rapidly to emergent diseases, and the substantial NIH contribution to this research. It also underscores the scale and significance of the public-sector investments that enable new drug discovery and development.

This work characterizes the NIH contribution to vaccine technologies being employed in "warp speed" development of vaccines for COVID-19, as well as the lack of sustained NIH funding for published research against recognized epidemic threats. Using quantitative methods, we examined the advance of published research on ten of the vaccine technologies incorporated in the 165 candidate vaccines entering development through July 2020 as well as the NIH funding that supported this research. Live, attenuated virus, inactivated virus, and adjuvant technologies have been used in successful products since the 1950s and continue to exhibit steady advance. Synthetic (recombinant) vaccines, viral vectors, DNA, and TLR9 agonists as adjuvants emerged since the 1980s, and exhibit a logistic, "S-curve" pattern of growth characteristic of emerging technologies that have passed an analytically-defined established point. In contrast, mRNA, virus-like particle, and nanoparticle technologies show exponential growth characteristic of technologies short of their established points. The body of research and NIH funding for established and emerging vaccine technologies exhibited sustained growth through the late 2010s, supported by > 16,000 project years of NIH funding totaling over $17.2 billion (2000-2019), the majority through cooperative agreements and intramural programs. NIH funding for published research on vaccines for recognized zoonotic threats including coronavirus, Zika, Ebola, and dengue, however, has been inconsistent and reactive to disease outbreaks. These data are considered in the context of the high failure rate for candidate vaccines and evidence that technological maturity is a significant factor in the efficiency of product development. Sustained funding for both enabling technologies and vaccine development is essential to ensure a rapid response to COVID and future pandemic threats. SIGNIFICANCE STATEMENTThis work examines the advance of research and NIH funding for technologies being employed in "warp speed" development of COVID-19 vaccines in the context of evidence that mature technologies have a greater likelihood of generating successful products. We show that candidate vaccines for COVID-19 employ a variety of established and still-emerging technologies, and identify $17.2 billion in NIH funding for this research from 2000-2019. In contrast, NIH funding for published research for vaccines on recognized pandemic threats has been inconsistent. This work highlights the significance and scale of the NIH contribution to vaccine technologies and the lack of sustained initiatives for vaccine development.

Both citation and funding metrics converge in shaping current perceptions of academic success. We aimed to evaluate what proportion of the most-cited USA-based biomedical scientists are funded by biomedical federal agencies and whether funded scientists are more cited than not funded ones. We linked a Scopus-based database on top-cited researchers (n=75,316 USA-based) and the NIH RePORTER database of 33 biomedical federal agencies (n=204,603 grant records) with matching based on name and institution. The 40,887 USA-based top-cited scientists who were allocated to any of 69 scientific subfields highly related to biomedicine were considered in the main analysis. The proportion of USA-based top-cited biomedical scientists (based on career-long citation impact) who had received any federal funding from biomedical research agencies was 63% for any funding (1996-2022), 21% for recent funding (2015-2022), and 14% for current funding (2021-2022). Respective proportions were 65%, 31%, and 21%, when top-cited scientists based on recent single year impact were considered. There was large variability across scientific subfields. No subfield had more than 31% of its top-cited USA-based scientists (career-long impact) currently funded. Funded top-cited researchers were overall more cited than non-funded top-cited scientists, e.g. mean (median) 14,420 (8983) versus 8,445 (4613) (p<0.001) and a substantial difference remained (, after adjusting for subfield and years since first publication. Differences were more prominent in some specific biomedical subfields. Overall, biomedical federal funding has offered support to approximately two-thirds of the top-cited biomedical scientists at some point during the last quarter century, but only a small minority of top-cited scientists have current federal biomedical funding. The large unevenness across subfields needs to be addressed with ways that improve equity, efficiency, excellence, and translational potential.

Little attention has been paid to governances impacts on the evolution of SARS-CoV-2, the virus that causes COVID-19. To evaluate such impacts on the evolution of vaccine resistance, we analyzed a stochastic compartmental model to quantify the risk a mutant strain capable of evading immunity emerges post-vaccine rollout. We calibrated the model with publicly available data for four territories in the Western Hemisphere qualitatively differing in pandemic interventions. The model shows an immune-evading strain to be readily selected over all infectivities in Texas. In Panama, only a high level of transmission permits immune evasion to evolve. No invasion appears likely in Costa Rica and Uruguay. Programs combining pharmaceutical and nonpharmaceutical interventions are best positioned to remove the epidemiological space SARS-CoV-2 needs to evolve vaccine resistance. One Sentence SummaryModes of governance and production help set the evolutionary trajectories of vaccine resistance in SARS-CoV-2 before vaccine campaigns begin.

ObjectivesWe measured attitudes towards "immunity passports" in the context of COVID-19 of a large sample of scientists. Consensus of scientists opinions on a different aspect of immunity passports was assessed. MethodsWe designed and implemented a survey to capture what scientists from around the world and different scientific background think about immunity certification. The survey was sent to the corresponding authors of scholarly articles published in the last five years in the top 20-ranked journals in each of the 27 subject areas between May and June 2020. Responses from 12,738 scientists were captured, and their distribution was tabulated by participants in health science and other fields. Consensus of responses was calculated using a variant of Shannon Entropy, made suitable for the ordinal response variables. ResultsHalf of the scientists surveyed, regardless of academic background agree that a potential immunity passport program will be good for public health (50.2%) and the economy (54.4%), with 19.1% and 15.4% of participants disagree, respectively. A significant proportion of scientists raised concerns about immunity certification over fairness to others (36.5%) and social inequality (45.5%). There is little consensus in the different aspects of immunity passport among scientists. Overall, scientists with health background hold a more conservative view towards immunity certification. ConclusionsOur findings suggest a lack of general agreement regarding the potential health and economic benefits, societal costs, and ethical issues of an immunity certification program within the scientific community. Given the relevant and important implications of immunity passport due to the increasing vaccine availability and efficacy, more attention should be given to the discussion of the design and implementation of immunity certification program. Strengths and limitations of this studyO_LIFirst cross-disciplinary survey with a large and international sample size that enables mapping of scientists opinions and attitudes towards COVID-19 immunity certificates. C_LIO_LIFrom the survey responses, we measured, reported, and compared the levels of consensus of scientists between health-related and non-health-related discipline. C_LIO_LIResponse rate and sample representativeness are moderate. C_LI

Neisseria gonorrhoeae (NG) bacteria have evolved resistance to many of the antibiotics that have been used successfully to treat gonorrhoea infection. To gain a better understanding of potential treatment options for gonorrhoea, we extend a previously developed within-host mathematical model to integrate treatment dynamics by accounting for key pharmacokinetic (PK) and pharmacodynamic (PD) features. This extended model was used to investigate different treatment regimens for two potential treatment options, namely, monotreatment with gepotidacin, and dual treatment with gentamicin and azithromycin. The simulated treatment success rates aligned well with the, albeit limited, clinical trial data that are available. The simulation results indicated that antibiotic treatment failure is associated with failure to successfully clear intracellular NG (NG residing within epithelial cells and neutrophils) and that extracellular PK indices alone cannot differentiate between treatment success or failure. We found that the index defined by the ratio of area under the curve to minimum inhibitory concentration (AUC/MIC) index > 150h, evaluated using intracellular gepotidacin concentration, successfully distinguished between treatment success and failure. For the dual treatment regimen, AUC/MIC index > 140h evaluated using the simulated single drug concentration, representing the combined effect of gentamicin and azithromycin with the Loewe additivity concept, successfully differentiated between treatment success and failure. However, we found this PK threshold associated with dual treatment to be less informative than in the gepotidacin monotreatment case as a majority of samples below this threshold still resulted in infection clearance. Although previous experimental results on the killing of intracellular NG are scarce, our findings draw attention to the importance of further experiments on antibiotic killing of intracellular NG. This will be useful for testing putative new anti-gonorrhoea antibiotics. Author SummaryGonorrhoea is a sexually transmitted infection caused by bacteria of the species Neisseria gonorrhoeae (NG). Although gonorrhoea can be easily treated using antibiotics, due to the propensity of NG to acquire resistance to antimicrobials, available treatment options have greatly diminished and most of the antibiotics used to treat infection in the past are now removed from treatment recommendations. As clinical trials have limitations in terms of expense, duration and ethical constraints they are not ideal for optimising doses, regimens and drug combinations. In this case, simulations through within-host mathematical models are useful in determining the effective dosing regimens and to explore intracellular treatment effects for which there is little experimental evidence. Our simulations identified the importance of treating intracellular NG (NG residing within neutrophils and epithelial cells) and the importance of considering intracellular pharmacokinetic indices when differentiating treatment success and failure. With the use of this model, we can simulate a range of different treatment regimens and drug combinations to assess their effectiveness at various values of the minimum inhibitory concentration which can potentially be used to guide future clinical trial design.

Preregistration at public research registries is considered a promising solution to the credibility crisis in science, but empirical evidence of its actual benefit is limited. Guaranteeing research integrity is especially vital in clinical research, where human lives are at stake and investigators might suffer from financial pressure. This paper analyzes the distribution of p-values from pre-approval drug trials reported to ClinicalTrials.gov, the largest registry for research studies in human volunteers, conditional on the preregistration status. The z-score density of non-preregistered trials displays a significant upward discontinuity at the salient 5% threshold for statistical significance, indicative of p-hacking or selective reporting. The density of preregistered trials appears smooth at this threshold. With caliper tests, we establish that these differences between preregistered and non-preregistered trials are robust when conditioning on sponsor fixed effects and other design features commonly indicative of research integrity, such as blinding and data monitoring committees. Our results suggest that preregistration is a credible signal for the integrity of clinical trials, as far as it can be assessed with the currently available methods to detect p-hacking.

The causes of multiple sclerosis (MS) remain unknown. Smoking has been associated with MS in observational studies and is often thought of as an environmental risk factor. We used two-sample Mendelian Randomization (MR) to examined whether this association is causal using genetic variants identified in genome-wide association studies (GWAS) as associated with smoking. We assessed both smoking initiation and lifetime smoking behaviour (which captures smoking duration, heaviness and cessation). There was very limited evidence for a meaningful effect of smoking on MS susceptibility was measured using summary statistics from the International Multiple Sclerosis Genetics Consortium (IMSGC) meta-analysis, including 14,802 cases and 26,703 controls. There was no clear evidence for an effect of smoking on the risk of developing MS (smoking initiation: odds ratio [OR] 1.03, 95% confidence interval [CI] 0.92-1.61; lifetime smoking: OR 1.10, 95% CI 0.87-1.40). These findings suggest that smoking does not have a detrimental consequence on MS susceptibility. Further work is needed to determine the causal effect of smoking on MS progression.

Antimicrobial resistance (AMR) poses an urgent public health challenge. To improve patient outcomes and design interventions we must identify patient characteristics which predict the presence of AMR pathogens. One potential and commonly collected patient characteristic is host age, consensus remains elusive regarding its impact on the probability of infecting pathogens being resistant to antimicrobials. Here, we employ a meta-analysis to consolidate and compare these previous studies and examine the relationship between antibiotic resistance and host age across bacteria and antibiotics. We show that although the probability that infecting bacteria are antimicrobial resistant increases with host age on average, diverse patterns exist across antibiotic classes and bacterial genera, including negative, humped, and U-shaped relationships. We further illustrate, using a compartmental epidemiological model, that this variation is likely driven by differences in antibiotic consumption or incidence of bacterial infection/carriage between age groups, combined with age assortative transmission. These findings imply that empirical antibiotic therapy could be improved by considering age-specific local resistance levels (compared with overall local resistance levels), resulting in improved treatment success and reduced spread of antibiotic resistance. They additionally display consequences of assuming population homogeneity in epidemiological models. Finally, they indicate that the landscape of the already severe resistance crisis is likely to change as the age distribution of the human population shifts.

Failure to routinely assess the state of knowledge as new studies accumulate results in 1) non-use of effective interventions, 2) continued use of ineffective or harmful interventions, and 3) unnecessary research. We use a published cumulative meta-analysis of interventions to reduce the harms of acute myocardial infarctions (1966-1992), and applied population attributable risk to assess the mortality consequences of the failure to cumulatively assess the state of knowledge. Failure to use knowledge that would have been available with cumulative meta-analysis may have resulted in annual estimated mortality: 41,000 deaths from non-use of intravenous dilators, 35,000 deaths from non-use of aspirin, and 37,000 deaths annually from non-use of {beta}-blockers. Continued use of Class 1 anti-arrhythmic drugs, which would have been found to be harmful in 1981, resulted 39,000 deaths annually. Failure to routinely update the state knowledge can have large health consequences. The process of building knowledge and practice in medicine and public health needs fundamental revision.

In this paper, we compare the inference regarding the effectiveness of the various non-pharmaceutical interventions (NPIs) for COVID-19 obtained from three SIR models, all developed by the Imperial College COVID-19 Response Team. One model was applied to European countries and published in Nature1 (model 1), concluding that complete lockdown was by far the most effective measure, responsible for 80% of the reduction in Rt, and 3 million deaths were avoided in the examined countries. The Imperial College team applied a different model to the USA states2 (model 2), and in response to our original submission, the Imperial team has proposed in a referee report a third model which is a hybrid of the first two models (model 3). We demonstrate that inference is highly nonrobust to model specification. In particular, inference regarding the relative effectiveness of NPIs changes substantially with the model and decision makers who are unaware of, or ignore, model uncertainty are underestimating the risk attached to any decisions based on that model. Our primary observation is that by applying to European countries the model that the Imperial College team used for the USA states (model 2), complete lockdown has no or little effect, since it was introduced typically at a point when Rt was already very low. Moreover, using several state-of-the-art metrics for Bayesian model comparison, we demonstrate that model 2 (when applied to the European data) is better supported by the data than the model published in Nature1. In particular, serious doubt is cast on the conclusions in Flaxman et al.1, whether we examine the data up to May 5th (as in Flaxman et al.1) or beyond the point when NPIs began to be lifted. Only by objectively considering a wide variety of models in a statistically principled manner, can one begin to address the effectiveness of NPIs such as lockdown. The approach outlined in this paper provides one such path.

IntroductionClinical and population decision-making relies on the systematic evaluation of extensive regulatory evidence. The FDA drug reviews provide detailed information on clinical trial design, enrollment criteria, sample size, randomization, comparators, endpoints, and indications. However, extracting these data is resource-intensive and time-consuming. Generative Artificial Intelligence large language models (LLMs) may accelerate the extraction and synthesis of such information. This study compares the performance of three LLMs, ChatGPT-4o, Gemini 2.5 Pro, and DeepSeek R1, in extracting and synthesizing regulatory and clinical information using antibiotics approved for complicated urinary tract infections (cUTIs) between 2010 and 2025 as a case study. MethodsLLM models were evaluated using general (short, direct) and detailed (structured, guidance-referencing) prompts across five domains including accuracy (precision and recall), explanation quality, error type (hallucination rate, misclassification, and omission), efficiency (response time, correct answers per second, and seconds per correct answer), and consistency with responses generated in duplicate runs. Two investigators independently reviewed outputs against FDA guidance, resolving discrepancies by consensus. Statistical analyses included {chi}{superscript 2}, Wilcoxon, and Kruskal-Wallis tests with false discovery rate correction. ResultsAmong 324 responses, accuracy differed significantly across models ({chi}{superscript 2}, p<0.001) with Gemini 2.5 Pro achieving the highest accuracy (66.7%), followed by ChatGPT-4o (51.9%) and DeepSeek R1 (37.0%). General prompts outperformed detailed prompts (59.3% vs 44.4%; p=0.011). Gemini 2.5 Pro showed highest explanation quality and most consistent outputs, while ChatGPT-4o had shortest response times and highest efficiency. Hallucination was the most frequent error type across models. ConclusionLLMs showed variable capability in extracting regulatory and clinical trial information. Gemini 2.5 Pro showed the strongest overall performance, while ChatGPT-4o was faster but less accurate, and DeepSeek R1 underperformed across most domains. These findings highlight both the promise and limitations of LLMs in regulatory science and support complementary use alongside human review to streamline evidence synthesis. Author SummaryOur research addresses a critical question in artificial intelligence for healthcare: how well do generative Generative Artificial Intelligence (GenAI) tools extract and synthesize regulatory and clinical information to inform decision-making? We assessed ChatGPT-4o, Gemini 2.5 Pro, and DeepSeek R1 performance in extracting and synthesizing information from regulatory documents and clinical trial data using all FDA approved antibiotics for the treatment of complicated urinary tract infections. We compared LLMs outputs directly with the original data sources. We assessed the models performance using both broad and detailed prompts across several areas, including accuracy of the information (precision and recall), quality of explanation, type of errors (hallucination, misclassification, and omission), efficiency and speed (response time, correct answers per second, and seconds per correct answer), and consistency of responses across repeated runs. The results suggest that while the models were generally fast and efficient extracting large volumes of information, they also produced errors and omissions that could limit their reliability. These findings highlight both the promise and the current limitations of GenAI, underscoring its potential value as a human supervised tool for safely supporting regulatory science and clinical decision-making.

Calls for eliminating prioritization for SARS-CoV-2 vaccines are growing amid concerns that prioritization reduces vaccination speed. We use an SEIR model to study the effects of vaccination distribution on public health, comparing prioritization policy and speed under mitigation measures that are either eased during the vaccine rollout or sustained through the end of the pandemic period. NASEMs recommended prioritization results in fewer deaths than no prioritization, but does not minimize total deaths. If mitigation measures are eased, abandoning NASEM will result in about 134,000 more deaths at 30 million vaccinations per month. Vaccination speed must be at least 53% higher under no prioritization to avoid increasing deaths. With sustained mitigation, discarding NASEM prioritization will result in 42,000 more deaths, requiring only a 26% increase in speed to hold deaths constant. Therefore, abandoning NASEMs prioritization to increase vaccination speed without substantially increasing deaths may require sustained mitigation.

Molnupiravir is an antiviral medicine that induces lethal copying errors during SARS-CoV-2 RNA replication. Molnupiravir reduced hospitalization in one pivotal trial by 50% and had variable effects on reducing viral RNA levels in three separate trials. We used mathematical models to simulate these trials and closely recapitulated their virologic outcomes. Model simulations suggest lower antiviral potency against pre-omicron SARS-CoV-2 variants than against omicron. We estimate that in vitro assays underestimate in vivo potency 7-8 fold against omicron variants. Our model suggests that because polymerase chain reaction detects molnupiravir mutated variants, the true reduction in non-mutated viral RNA is underestimated by [~]0.5 log10 in the two trials conducted while omicron variants dominated. Viral area under the curve estimates differ significantly between non-mutated and mutated viral RNA. Our results reinforce past work suggesting that in vitro assays are unreliable for estimating in vivo antiviral drug potency and suggest that virologic endpoints for respiratory virus clinical trials should be catered to the drug mechanism of action.

ImportanceIt is important to monitor changes in biomedical literature and its funding. China has surpassed the USA in publications and, in some analyses, also in some impact indicators. ObjectiveTo evaluate changes over time in the profiles of the most highly cited biomedical papers. DesignThe 100 top-cited biomedical papers (based on Scopus) published in each of three time periods (2003-4, 2013-4, and 2023-4) were assessed for corresponding authors, types of publications represented, and funding sources, with emphasis on funding from the US National Institutes of Health (NIH) that has been traditionally considered the major funder of biomedical research. SettingGlobal. Participantsnot relevant Exposuresnot relevant Main outcome measuresProvenance and funding sources. Main findingsCorresponding authors from the USA decreased overtime (59/100 papers in 2003-4, 58/100 in 2013-4, 45/100 in 2023-4). China had corresponding authors in 0,1, and 4 top cited papers in the three time periods, respectively. There was a marked increase in consensus items (10/100 in 2003-4 versus 24/100 in 2023-4) and in reference statistics papers (1/100 in 2003-4, 10/100 in 2013-4, 11/100 in 2023-4). Reviews remained common among top cited papers, but almost always they were non-systematic. NIH funding was listed in 45/100, 50/100, and 23/100 papers in the three time periods, respectively. All other countries combined surpassed US public funding in 2023-4. Funding by NIH alone decreased sharply in the last decade (32/100, 28/100, and 2/100 in the three time periods, respectively). More commonly listed funding from non-profit organizations, societies, and institutions complemented the NIH funding decline. The first authors of 7/45 and the corresponding author(s) of 14/45 top cited USA-based papers of 2023-4 were listed as leaders of active NIH grants in RePORTER as of February 2025. Citation gaming became more obvious in 2023-4. Conclusions and relevanceOverall, the USA remains a world leader regarding the most highly cited biomedical research and NIH funding retains a substantial presence among top cited papers. However, NIH influence has shrunk overall, and top cited papers funded exclusively by NIH have almost disappeared. Strengthening public funding is essential to secure research serves the common good.

OBJECTIVETo explore the nature of genetic and environmental susceptibility to multiple sclerosis (MS) and to define the limits of this nature based on the statistical uncertainties regarding the various epidemiological observations that have been made. BACKGROUNDCertain parameters of MS-epidemiology are directly observable (e.g., the risk of MS-recurrence in siblings and twins of an MS proband, the proportion of women among MS patients, the population-prevalence of MS, and the time-dependent changes in the female-to-male (F:M) sex-ratio. By contrast, other parameters can only be inferred from observed parameters (e.g., the proportion of the population that is genetically susceptible, the proportion of women among susceptible individuals, the probability that a susceptible individual will experience an environment sufficient to cause MS given their genotype, and if they do, the probability that they will develop the disease). DESIGN/METHODSThe "genetically-susceptible" subset (G) of the population (Z) is defined to include everyone with any non-zero life-time chance of developing MS under some environmental conditions. For the observed parameters, acceptable ranges are assigned values such that they always include their 95% confidence intervals. By contrast, for the non-observed parameters, the acceptable ranges are assigned such that they cover the entire "plausible" range for each parameter. Using both a Cross-sectional Model and a Longitudinal Model, together with established parameter relationships, we explore, iteratively, trillions of potential parameter combinations and determine those combinations (i.e., solutions) that fall within the acceptable range for the observed and non-observed parameters. RESULTSBoth Models and all analyses are consistent and converge to demonstrate that genetic-susceptibitly is limited to 52% or less of the population and to 30% or less of women. Consequently, most individuals (particularly women) have no chance whatsoever of developing MS, regardless of their environmental exposure. Also, currently, the penetrance of MS in susceptible women is greater than it is in men. Moreover, as expected, the probability that susceptible individuals will develop MS increases with an increased likelihood of these individuals experiencing an environment sufficient to cause MS, given their genotype. Nevertheless, although it is conceivable that these response-curves plateau at 100% for both women and men, this possibility requires extreme conditions and seems remote. Rather, at least men, seem to plateau well below this level and, if so, it is this difference, rather than any differences in the genetic and environmental determinants of disease, that primarily accounts both for the difference in penetrance between women and men and for the increasing proportion of women among of MS patients worldwide. CONCLUSIONSThe development of MS (in an individual) requires both that they have an appropriate genotype (which is uncommon in the population) and that they have an environmental exposure sufficient to cause MS given their individual genotype. Nevertheless, even when the necessary genetic and environmental factors, sufficient for MS pathogenesis, co-occur for an individual, this still insufficient for that person to develop MS. Thus, disease pathogenesis, even in this circumstance, seems not to be deterministic but, rather, to involve an important element of chance. Author SummaryCertain parameters of MS-epidemiology can be directly observed. These parameters include the risk of MS recurrence in siblings and twins of an MS proband, the proportion of women among MS patients, the population-prevalence of MS, and the time-dependent changes in the female-to-male (F:M) sex-ratio. By contrast, there are other parameters of MS-epidemiology, which cant be observed, but which must be inferred based on the values of the observable parameters. These parameters include the proportion of the general population (Z) that is genetically susceptible to MS, the proportion of women among susceptible individuals, the probability that a susceptible individual will experience an environment sufficient to cause MS, and if they do, the likelihood that they will, in fact, develop the MS. We define the subset (G) - i.e., the genetically-susceptible subset - to include everyone in (Z) who has any non-zero chance of developing MS over their life-time, under some environmental circumstances. For the observed parameters, plausible ranges are assigned acceptable values such that they always include their 95% confidence interval. By contrast, for the non-observed parameters, the acceptable ranges are assigned such that they cover the entire "plausible" range for each parameter. Then, using both a Cross-sectional Model and a Longitudinal Model, together with established parameter relationships, we explore iteratively trillions of potential parameter combinations and determine those combinations (i.e., solutions) that are allowed by the observed and non-observed parameter ranges. The Cross-sectional Model makes two assumptions, commonly made in studies of monozygotic twins, to establish certain relationships between the observed and non-observed parameters. By contrast, the Longitudinal Model makes neither of these assumptions but, rather, this Model utilizes the observed changes in the female-to-male (F:M) sex-ratio and the disease prevalence, which have taken place over the past 4-5 decades, to determine the response curves for susceptible individuals, relating their probability of developing MS to their probability of experiencing an environment sufficient to cause MS. Both Models and all analyses are consistent with each other and converge to demonstrate that genetic-susceptibitly is limited to 52% or less of the population and 30% or less of women. Consequently, most individuals have no chance whatsoever of developing MS, regardless of their environmental experiences. Thus, MS is a genetic disease in the sense that, if an individual does not have the correct genetic makeup, they cant develop the disease. However, the probability that susceptible individuals will develop MS increases with an increased likelihood of these individuals experiencing an environment sufficient to cause MS, given their genotype. Thus, MS is also and environmental disease in the sense that the development of MS (in an individual), in addition to their having an appropriate genotype, requires that they experience an environmental exposure sufficient to cause MS given their individual genotype. Nevertheless, there must be another factor involved in disease pathogenesis because, although it is conceivable that these response-curves plateau at 100% for both women and men, this possibility requires extreme conditions and seems remote. Rather, at least men, seem to plateau well below this and, if so, it is this difference, rather than differences in the genetic and environmental determinants of disease, that primarily accounts both for the difference in penetrance between women and men and for the increasing proportion of women among of MS patients worldwide. Consequently, even when the necessary genetic and environmental factors, sufficient for MS pathogenesis, co-occur for an individual, this still seems to be insufficient for that person to develop MS. Thus, disease pathogenesis, even in this circumstance, seems not to be deterministic but, rather, to involve an important element of chance.

Stockpiling and rapid use of oral influenza antiviral drugs such as oseltamivir and baloxavir marboxil (BXM) can reduce the disease burden of a nascent influenza pandemic before vaccines are available. Current estimates for pandemic preparedness drug stockpile size vary widely depending on modeling assumptions and do not account for heterogeneities in healthcare-seeking behaviour or drug-specific transmission risk reduction. Here, we developed a novel transmission model that accounts for heterogeneous healthcare-seeking behaviour and recent estimates of transmission risk reduction by antivirals to estimate country-specific demand and impact of distributing oseltamivir and BXM in 186 countries. Due to its transmission reducing properties, BXM could maximally double the median percentage of mean pandemic deaths averted (37%-68%) relative to oseltamivir with [~]5%-10% with smaller stockpile size (7%-34% per-capita). Under limited drug availability, age-based rationing does not meaningfully lower total antiviral demand and drug priority should be given to treatment over post-exposure prophylaxis.

ObjectiveMajor histocompatibility complex (MHC) locus carries a significant genetic risk burden for multiple sclerosis (MS). Here we investigate the structurally diverse complement component 4 (C4) alleles within the MHC in MS development and disease progression. MethodsWe imputed and examined C4 alleles based on data from two case-control cohorts (N1= 3252 cases and 5725 controls; N2= 8978 cases and 6976 controls), a clinical MS cohort (N3= 2387 cases) and a cohort with immune cell gene expression data (N4= 33 cases and 33 controls). We have performed gene-level analysis to investigate the shared genetic landscape between MS susceptibility (N= 14802 cases and 26703 controls) and plasma C4 protein (N= 68716). ResultsOur data showed that C4 genetic structural variants were associated with significant changes in MS susceptibility and disability progression. For instance, higher C4AL copy number burden was associated with lower MS susceptibility (fixed effect meta-analysis odds ratio= 0.89, P= 5.65x10-6) independent of established MHC risk variants such as HLA-DRB1*15:01. Higher C4AL copy number was also associated with reduced hazard of reaching MS disability milestones such as Expanded Disability Status Scale 3 (hazard ratio= 0.79, P= 9.0x10-15). In addition, we found C4 alleles may also modulate C4 expression in disease-relevant immune cell types such as CD8+ T cells. Further, we identified that candidate genes shared between MS susceptibility and plasma C4 protein level were enriched in biological pathways of immune regulation, Epstein-Barr virus infection and other autoimmune diseases such as lupus. InterpretationThese findings support future investigations of the C4 genetic structural variants as potential mechanistic and therapeutic targets in MS pathogenesis and disease progression.

BackgroundDespite representing approximately 85% of the global population, low- and middle-income countries (LMICs) contribute less than 10% of participants to genome-wide association studies and biobank research. This disparity has profound implications for the generalisability of precision medicine. However, no standardised framework exists to quantify research equity at the biobank level or track progress over time. MethodsWe developed the Health Equity Informative Metrics (HEIM) framework to quantify alignment between biobank research output and global disease burden. We analysed 75,356 PubMed-indexed publications (2000-2025) from 27 biobanks across 19 countries, mapping each to 179 disease categories from the Global Burden of Disease Study 2021. We calculated disease-specific Gap Scores measuring the mismatch between burden (disability-adjusted life years, DALYs) and research attention, biobank-level Equity Alignment Scores (EAS), and regional equity ratios comparing high-income (HIC) to LMIC research intensity. FindingsWithin our 27-biobank sample, HIC biobanks produced substantially higher research output per DALY compared to LMIC biobanks (ratio: 322:1; sensitivity analyses: >100:1 across methodological variations). Regional concentration was marked: the Americas and Europe accounted for 97.8% of publications, while Africa, Eastern Mediterranean, and South-East Asia combined contributed <1%. Of 179 disease categories, 23 (13%) exhibited critical or high-severity research gaps despite substantial global burden. Only 4 of 27 biobanks (15%) achieved Strong or Moderate equity alignment scores; 19 (70%) were rated Poor. Six disease categories showed critical gaps, including drowning (15.7 million DALYs, 0 mapped publications) and iodine deficiency (2.3 million DALYs, 10 publications). InterpretationThe HEIM framework reveals substantial disparities in how biobank research capacity is distributed relative to global disease burden. While the precise equity ratio varies with sample selection and methodology, the fundamental pattern of profound inequity is robust across reasonable analytical choices. These findings provide baseline measurements for tracking progress toward more equitable genomic research and identify high-priority targets for intervention.