Supplementary material 3 from: Orosco F (2025) Pan-viral efficacy profile of ribavirin: quantitative potency and safety landscape across virus families. Pharmacia 72: 1-11. https://doi.org/10.3897/pharmacia.72.e159854

*% detected as AI AI detection includes the possibility of false positives. Although some text in this submission is likely AI generated, scores below the 20% threshold are not surfaced because they have a higher likelihood of false positives. Caution: Review required. It is essential to understand the limitations of AI detection before making decisions about a student’s work. We encourage you to learn more about Turnitin’s AI detection capabilities before using the tool. Disclaimer Our AI writing assessment is designed to help educators identify text that might be prepared by a generative AI tool. Our AI writing assessment may not always be accurate (it may misidentify writing that is likely AI generated as AI generated and AI paraphrased or likely AI generated and AI paraphrased writing as only AI generated) so it should not be used as the sole basis for adverse actions against a student. It takes further scrutiny and human judgment in conjunction with an organization's application of its specific academic policies to determine whether any academic misconduct has occurred. Frequently Asked Questions How should I interpret Turnitin's AI writing percentage and false positives? The percentage shown in the AI writing report is the amount of qualifying text within the submission that Turnitin’s AI writing detection model determines was either likely AI-generated text from a large-language model or likely AI-generated text that was likely revised using an AI-paraphrase tool or word spinner. False positives (incorrectly flagging human-written text as AI-generated) are a possibility in AI models. AI detection scores under 20%, which we do not surface in new reports, have a higher likelihood of false positives. To reduce the likelihood of misinterpretation, no score or highlights are attributed and are indicated with an asterisk in the report (*%). The AI writing percentage should not be the sole basis to determine whether misconduct has occurred. The reviewer/instructor should use the percentage as a means to start a formative conversation with their student and/or use it to examine the submitted assignment in accordance with their school's policies. What does 'qualifying text' mean? Our model only processes qualifying text in the form of long-form writing. Long-form writing means individual sentences contained in paragraphs that make up a longer piece of written work, such as an essay, a dissertation, or an article, etc. Qualifying text that has been determined to be likely AI-generated will be highlighted in cyan in the submission, and likely AI-generated and then likely AI-paraphrased will be highlighted purple. Non-qualifying text, such as bullet points, annotated bibliographies, etc., will not be processed and can create disparity between the submission highlights and the percentage shown. Page 2 of 13 - AI Writing Overview Submission ID trn:oid:::29034:97165091 Page 2 of 13 - AI Writing Overview Submission ID trn:oid:::29034:97165091 Pan-viral efficacy profile of Ribavirin: Quantitative potency and safety landscape across virus families Abstract Marked variation in antiviral potency, selectivity, and clinical efficacy is observed across seven major virus families for Ribavirin. In-vitro analyses reveal highest potency against flaviviruses and paramyxoviruses, while coronaviruses and astroviruses display reduced susceptibility, consistent with the presence of viral proofreading polymerases. Comparative profiling shows Remdesivir excels in coronavirus and filovirus models, while Favipiravir demonstrates exceptional selectivity in orthomyxovirus assays. Clinical data align with laboratory findings, confirming robust therapeutic effects for hepatitis C virus and Lassa fever, but less consistent outcomes for respiratory viruses. Trait-based regression modeling highlights the impact of viral polymerase fidelity on drug susceptibility. The results emphasize the necessity for pan-viral benchmarking in antiviral development and the rational deployment of broad-spectrum agents in clinical and outbreak settings. Keywords: Cytotoxicity, ribavirin, selectivity index, trait analysis, viral taxonomy Introduction Pandemic preparedness and the ongoing challenge of emerging infectious diseases have highlighted the importance of broad-spectrum antivirals as foundational tools in modern medical therapeutics (Thomas et al. 2012, Loustaud-Ratti et al. 2016). Unlike pathogen-specific therapies or vaccines, which may require months or years to develop and distribute, broad-spectrum antiviral agents can offer immediate, flexible response options against novel or re-emerging viral threats. Among these, ribavirin has stood out for decades as one of the most extensively deployed and investigated compounds, demonstrating inhibitory activity against a wide variety of both RNA and DNA viruses (Graci and Cameron 2006, Thomas et al. 2012, Loustaud-Ratti et al. 2016). Initially introduced in the 1970s as a synthetic guanosine analog, ribavirin has earned clinical approval or emergency use authorization for several high-consequence viral infections. Its applications encompass the treatment of chronic hepatitis C virus (HCV) infection (Lawitz et al. 2013, Zeuzem et al. 2014), Lassa fever (McCormick et al. 1986, Bausch et al. 2010), and respiratory syncytial virus (RSV) infections (Rodriguez et al. 1987). More recently, ribavirin has also been utilized in the management of severe coronavirus outbreaks, including SARS and COVID-19, sometimes as part of combination regimens or compassionate use protocols (Koren et al. 2003, Tong et al. 2020). The breadth of these indications attests to the compound’s broadspectrum potential, but also underscores the complex interplay of viral, host, and pharmacological factors that can influence clinical outcomes. Despite ribavirin’s longstanding clinical presence, the full spectrum of its antiviral mechanisms remains a subject of ongoing research and debate. Multiple modes of action have been described, Page 3 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Page 3 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 including inhibition of inosine monophosphate dehydrogenase (IMPDH), which leads to depletion of intracellular guanosine pools; direct interference with viral RNA-dependent RNA polymerase; the induction of lethal mutagenesis through increased replication errors; and modulation of host immune responses (Crotty et al. 2001, Tam et al. 2001, Leyssen et al. 2005, Graci and Cameron 2006). Notably, the dominant mechanism appears to vary among viral families. In cell culture systems for flaviviruses and paramyxoviruses, IMPDH inhibition is frequently observed as the key pathway (Leyssen et al. 2005), while in coronaviruses, the presence of a proofreading exoribonuclease can mitigate the mutagenic effects of ribavirin, reducing its efficacy (Smith et al. 2013, Agostini et al. 2018, Ferron et al. 2018). This mechanistic diversity is reflected in the marked variability of ribavirin’s potency across different viruses, as well as among strains within a given family(Graci and Cameron 2006, Thomas et al. 2012, Loustaud-Ratti et al. 2016). Comparative studies involving other nucleoside analogs, such as favipiravir and remdesivir, have further clarified ribavirin’s position within the antiviral pharmacopeia (Furuta et al. 2013, Sheahan et al. 2017, Xu et al. 2021, Radoshitzky et al. 2023). While ribavirin demonstrates moderate activity in vitro against a number of viruses, its therapeutic application is often limited by the attainable plasma drug concentrations and the selectivity index, which defines the margin between antiviral efficacy and cytotoxicity. Hemolytic anemia, a well-recognized dose-limiting toxicity, remains a key concern in clinical management, particularly during prolonged therapy (De Franceschi et al. 2000, Xu et al. 2021). This risk-benefit scenario must be carefully balanced, especially when alternative broad-spectrum agents are available. Recent advances in viral genomics and molecular virology have brought increasing attention to the influence of viral traits on drug susceptibility. Features such as genome polarity, replication site, the presence or absence of viral proofreading mechanisms, and genome segmentation are now recognized as important determinants of antiviral response (De Franceschi et al. 2000, Smith et al. 2013, Ferron et al. 2018). The systematic analysis of these traits may provide a foundation for rational drug development and facilitate the prediction of antiviral efficacy in both established and emerging pathogens (Minskaia et al. 2006, Bouvet et al. 2012). While a substantial literature exists describing the in vitro effects, mechanisms, and clinical experience with ribavirin, no comprehensive synthesis has yet quantified its pan-viral efficacy in a way that integrates potency, selectivity, viral determinants, and therapeutic outcomes across the major human viral pathogens. Such an analysis is necessary to benchmark ribavirin’s true position among current and future broad-spectrum antivirals, to identify the most relevant viral predictors of response, and to inform public health strategies for outbreak management and pandemic response. The present work addresses this gap by assembling a harmonized dataset of in vitro potency and selectivity data for ribavirin across seven viral families, with direct benchmarking against favipiravir and remdesivir. The analysis further investigates the viral trait determinants of susceptibility and synthesizes clinical outcome data to contextualize laboratory findings. Thus, this study aims to provide an up-to-date, quantitative profile of ribavirin’s strengths and limitations, with implications for both mechanistic understanding and translational antiviral strategy. Page 4 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Page 4 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Materials and Methods 1. Literature Screening and Data Extraction A systematic review of the published literature was conducted to identify original studies assessing the antiviral activity of Ribavirin, as well as comparator data for Favipiravir and Remdesivir, against major human virus families. Bibliographic databases searched included PubMed, Scopus, and Web of Science, supplemented by reference mining from relevant reviews and prior meta-analyses. The search encompassed articles published up to March 2024, without language restrictions at the initial stage. The search strategy combined the terms “Ribavirin,” “broad-spectrum antiviral,” and the names of targeted virus families or genera. Studies were considered eligible if they reported original, experimentally determined in vitro potency data (EC₅₀ or IC₅₀), cytotoxicity (CC₅ ₀ ), or selectivity index (SI), or if they presented clinical outcomes following Ribavirin administration in human viral infections. Inclusion required the presence of at least one quantitative EC₅ ₀ or equivalent measure for a virus-drug pair. Exclusion criteria comprised the absence of experimental or clinical outcome data, lack of relevant quantitative measures, duplicate or overlapping publications, and non-English articles at the fulltext review stage if no translation was available. The literature screening proceeded through multiple phases. Initial retrieval produced 48 records. After deduplication, 42 unique titles and abstracts were screened for relevance. Full-text assessment was performed for 25 articles, with 22 studies meeting all eligibility criteria and included for data extraction. Extraction of data from in vitro studies included virus species, viral family, assay platform, host cell or animal model, EC₅ ₀ or IC₅₀ value, CC₅ ₀ , SI, assay conditions (temperature, MOI, incubation time), number of replicates, and study reference. For clinical studies, data fields encompassed virus species, study design, patient population, primary clinical endpoint, effect size (e.g., risk ratio, odds ratio, response rate), confidence intervals or p-values, dosing regimen, and sample size. Discrepancies and ambiguities were addressed by consulting supplementary materials, cross-referencing with other datasets, or contacting study authors if necessary. 2. In Vitro Potency and Safety Data In-vitro antiviral potency was characterized by the half-maximal effective concentration (EC₅ ₀ ), defined as the concentration of drug required to inhibit 50% of virus replication relative to untreated controls. When reported, the half-maximal inhibitory concentration (IC₅ ₀ ) was considered equivalent to EC₅ ₀ unless otherwise specified (Leyssen et al. 2005, Graci and Cameron 2006, Loustaud-Ratti et al. 2016). Cytotoxicity was measured by the half-maximal cytotoxic concentration (CC₅ ₀ ), representing the drug concentration that reduces host cell viability by 50%. The selectivity index (SI), a widely used metric of therapeutic window, was calculated as the ratio of CC₅ ₀ to EC₅ ₀ (SI = CC₅ ₀ /EC₅ ₀ ) (Graci and Cameron 2006, Loustaud-Ratti et al. 2016). Page 5 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Page 5 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Assay platforms included cytopathic effect (CPE) reduction assays, plaque reduction assays, RTqPCR quantification of viral RNA, luciferase or fluorescence-based reporter assays, and repliconbased systems, depending on virus and laboratory protocol. Host cell models varied according to virus family, encompassing Vero, Huh-7, HeLa, MDCK, Caco-2, and RD cells among others. All EC₅ ₀ and CC₅ ₀ values were standardized to micromolar (µM) units. Where results were reported as “greater than” (>) or “less than” (<) a threshold value, these were flagged as censored observations and retained at the reported boundary for visualization purposes, but excluded from calculation of medians and interquartile ranges. Records lacking numeric or complete EC₅ ₀ or CC₅ ₀ data were excluded from potency and selectivity analyses. For each drug (Ribavirin, Favipiravir, Remdesivir), a drug-virus-family matrix was constructed to summarize the availability of potency and safety data across the seven major virus families included in this study. 3. Viral Trait Annotation Virus species were annotated for genomic and replication traits expected to influence antiviral susceptibility. Traits included genome type (positive-sense single-stranded RNA [+ssRNA], negative-sense single-stranded RNA –ssRNA]), site of replication (cytoplasm or nucleus), presence of a proofreading polymerase (e.g., ExoN in coronaviruses), envelope status (enveloped or non-enveloped), and genome segmentation (monopartite or segmented) (Smith et al. 2013, te Velthuis 2014, Ferron et al. 2018). Annotation sources included primary literature, virus taxonomy databases, and established reviews on viral replication and structure (Strauss and Strauss 1994, Koonin et al. 2015). Trait variables were encoded as follows: genome type as binary categorical (+ssRNA vs –ssRNA), replication site as binary categorical (cytoplasm vs nucleus), proofreading polymerase as binary (yes/no), envelope status as binary (yes/no), and genome segmentation as numeric (number of segments, standardized to zero mean and unit variance for regression). Where ambiguity existed, majority consensus among recent reviews was used. 4. Quantitative Analyses Potency and selectivity metrics for each drug-virus-family combination were calculated as described above. Breadth of coverage was defined as the number of virus families with at least one EC₅ ₀ value reported for a given drug. Family-wise summary statistics (median EC₅ ₀ , interquartile range, mean, standard error) were computed for each virus family, with only numeric (non-censored) EC₅₀ values included. Radar plots were constructed to visualize and compare each drug’s spectrum in terms of breadth, median potency, and median selectivity index. For visual comparability, each axis was normalized to its global maximum, and polygons for each drug were overlaid with 40% fill transparency. Trait-based predictors of Ribavirin potency were evaluated using ridge regression. The response variable was log-transformed EC₅ ₀ (log₁₀[EC₅ ₀ , µM]), and the five annotated viral traits served as predictors. Predictors were standardized and, where appropriate, encoded using onehot or binary variables. Ridge regression models were fit with regularization parameter Page 6 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Page 6 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 alpha = 1.0. To assess coefficient stability and confidence intervals, a nonparametric bootstrap was applied with 1,000 resamplings of the dataset. Point estimates correspond to model coefficients from the full dataset, with 95% confidence intervals derived from the 2.5th and 97.5th percentiles of the bootstrap distribution. Scatterplots were generated to visualize the relationship between EC₅ ₀ and CC₅ ₀ for all assays, plotted on log₁ ₀ –log₁ ₀ axes. Points were colored according to SI (capped at the 95th percentile to avoid saturation), and marker shapes were used to differentiate Ribavirin, Favipiravir, and Remdesivir. Diagonal dashed lines (SI = 1 and SI = 10) were included for reference. 5. Clinical Outcomes Data Clinical studies reporting outcomes of Ribavirin therapy in human infections were extracted for primary efficacy endpoints (such as sustained virologic response, mortality, or disease progression), effect size estimates (risk ratio, odds ratio, response rate), associated confidence intervals, p-values, dosing regimens, and sample size. Where effect size was reported as a percent or absolute difference, transformation to risk-ratio scale was performed for consistency. Odds ratios and hazard ratios were converted to risk ratios using established formulas when required. In cases where only a p-value was available, the 95% confidence interval was approximated using the log-risk ratio and the z-score for the reported significance level. Studies lacking sufficient detail for effect size harmonization were excluded from quantitative synthesis. 6. Statistical and Visualization Methods All data cleaning, analysis, and visualization were performed using Python (pandas, numpy, matplotlib, seaborn, and scikit-learn libraries) and R (tidyverse and ggplot2 packages) as appropriate. Log-transformation was applied to EC₅ ₀ , CC₅₀, and SI variables to reduce skew and facilitate visualization. Groupwise statistics and confidence intervals were calculated using standard nonparametric approaches. Main-text figures and tables present the primary dataset, summary statistics, trait analysis, and clinical outcomes. Extended datasets (including Favipiravir and Remdesivir raw data, virus trait tables, and supplementary family-wise summaries) are provided in the Supplementary Material. Figure legends and supplementary information describe exact plotting parameters, normalization procedures, and additional data sources where relevant. Results 1. Literature Screening and Dataset Composition A systematic literature review was performed to identify published studies examining the antiviral effects of Ribavirin, as well as Favipiravir and Remdesivir, across a wide spectrum of viruses. The search strategy retrieved 48 unique records, which were reduced to 42 after deduplication. Abstract and title screening led to the selection of 40 studies for further consideration. Full-text Page 7 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Page 7 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 review was completed for 25 studies, ultimately resulting in the inclusion of 22 that met all eligibility criteria and contained extractable quantitative data for either in vitro potency, cytotoxicity, selectivity index, or clinical efficacy. The flow of study selection is summarized narratively and corresponds to the standard approach for transparent data synthesis. These 22 studies encompassed a broad diversity of experimental settings, viral taxa, and assay platforms. In total, 41 unique in vitro antiviral assays were included, representing seven major virus families of human relevance: Flaviviridae, Paramyxoviridae, Coronaviridae, Filoviridae, Orthomyxoviridae, Astroviridae, and Picornaviridae. Each family was represented by at least one viral species with standardized EC₅₀ and CC₅ ₀ data, enabling quantitative cross-family comparisons. In addition to Ribavirin, Favipiravir and Remdesivir were included as comparator drugs wherever sufficient data could be obtained. Datasets for Favipiravir and Remdesivir, along with additional detailed results by family, are provided in Supplementary Tables S1 and S2. The composition of the dataset allowed for robust analysis of drug-virus interactions. For most virus families, multiple independent assays using distinct platforms or host cell lines were available, improving the reliability of potency and selectivity estimates. The studies contributing clinical outcome data included randomized controlled trials, open-label studies, and retrospective cohort analyses, with a total of five studies spanning Flaviviridae, Arenaviridae, Paramyxoviridae, and Coronaviridae. 2. In-Vitro Pan-Viral Potency of Ribavirin Table 1 summarizes the in vitro potency and selectivity profiles of Ribavirin across all included assays. Potency, as measured by the half-maximal effective concentration (EC₅ ₀ ), ranged widely between virus families and individual assays. Flaviviridae and Paramyxoviridae consistently showed the lowest median EC₅₀ values, with several replicon-based hepatitis C virus assays reporting EC₅ ₀ values below 1 µM, and the majority of CPE and RT-qPCR-based assays yielding values below 50 µM. Conversely, assays targeting SARS-CoV-2, West Nile virus, and human astrovirus often returned EC₅ ₀ values exceeding 100 µM or were right-censored due to lack of inhibition at the highest tested concentration. The selectivity index (SI), calculated as CC₅ ₀ /EC₅₀, provided further insight into the therapeutic window of Ribavirin across different viral systems. Highest SI values were observed in assays with low EC₅ ₀ and high CC₅ ₀ (often exceeding 1000 µM in mammalian cell lines), while lower SI values reflected either increased cytotoxicity or diminished antiviral activity. Supplementary Table S4 details the family-wise means and standard errors for log EC₅ ₀ , as well as the complete SI distributions for each virus group. Fig. 1 presents the distribution of log₁₀ EC₅ ₀ values for Ribavirin in a family-wise manner using violin plots. An overlay of pooled family means and standard errors highlights central tendencies and the degree of dispersion within each family. In Flaviviridae and Paramyxoviridae, the density of points is tightly clustered, reinforcing the observation of consistent, potent activity across multiple viral species within these families. Notably, Flaviviridae assays, primarily those involving hepatitis C, dengue, and yellow fever viruses, displayed the most favorable potency profiles, Page 8 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Page 8 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 supporting the established clinical role of Ribavirin for chronic hepatitis C and severe flaviviral infections (McCormick et al. 1986, Lawitz et al. 2013, Zeuzem et al. 2014). In contrast, Coronaviridae and Astroviridae displayed a broader spread and higher median EC₅ ₀ values. Individual SARS-CoV-2 and human astrovirus assays demonstrated significant variability, with some data points approaching or exceeding the upper detection threshold, indicating reduced susceptibility. Filoviridae, Orthomyxoviridae, and Picornaviridae showed intermediate potency, often reflecting the influence of both viral and assay-specific factors, such as the choice of host cell, the duration of incubation, and the MOI. This family-wise analysis highlights the non-uniform activity of Ribavirin across the virosphere, with a clear gradient from highest potency in Flaviviridae and Paramyxoviridae to lowest in Coronaviridae and Astroviridae. Such variability likely reflects both intrinsic differences in viral biology and the impact of cellular and experimental contexts. 3. Comparative Activity of Broad-Spectrum Antivirals Fig. 2 provides a comprehensive heatmap depicting the median EC₅ ₀ values of Ribavirin, Favipiravir, and Remdesivir for each virus family. Data for Favipiravir and Remdesivir, derived from both shared and distinct experimental sources, are detailed in Supplementary Tables S1 and S2. Potency is encoded by a color scale mapped to log₁₀ EC₅ ₀ , with darker shades indicating higher antiviral activity. The heatmap reveals clear, drug-specific patterns. Ribavirin is most potent against Flaviviridae and Paramyxoviridae, aligning with clinical indications and in vitro findings. Favipiravir exhibits its greatest efficacy against Orthomyxoviridae, especially influenza A and B viruses, with median EC₅ ₀ values less than 1 µM. This is consistent with Favipiravir’s clinical development for influenza and other negative-sense RNA viruses (Furuta et al. 2013). Remdesivir demonstrates superior potency against Coronaviridae and Filoviridae, supporting its recent clinical deployment for COVID-19 and Ebola virus disease (Warren et al. 2016, Sheahan et al. 2017). Blank cells in the heatmap denote virus-drug-family combinations where insufficient or no standardized EC₅ ₀ data were available. The visual summary provided by the heatmap underscores the concept that no single broadspectrum antiviral achieves optimal potency across all virus families. Instead, each agent’s efficacy is shaped by both viral biology and the unique mechanism of action, supporting the rationale for tailored, virus family-specific, or combination antiviral strategies. 4. Breadth, Potency, and Selectivity: Integrated Comparison A more holistic comparison of the three drugs is provided in Fig. 3, which displays a radar plot integrating three core performance metrics: family coverage (taxonomic breadth), normalized potency score (1/log EC₅ ₀ ), and normalized median selectivity index (log SI). For each axis, values are normalized to the highest observed value among the three drugs, allowing direct shape and size comparisons of their pan-viral performance. Supplementary Tables S1, S2, and S4 provide the raw and summary statistics underlying each axis. Page 9 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Page 9 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Ribavirin stands out for its pan-viral breadth, with assay data available for all seven families and a generally favorable selectivity profile. Remdesivir, while covering fewer virus families, delivers the highest normalized potency, particularly for coronaviruses and filoviruses. Favipiravir achieves the highest selectivity, attributable to its generally low cytotoxicity, but its overall potency is lower than that of Remdesivir. The geometric shapes of the polygons emphasize the inherent trade-offs: for example, a wide breadth is not always accompanied by the highest potency or selectivity. This multidimensional view reveals the practical limitations and advantages of each agent in the context of pandemic and emerging pathogen preparedness. 5. Viral Trait Determinants of Ribavirin Potency To elucidate the viral determinants influencing Ribavirin’s efficacy, regularized regression analysis was performed using trait data from Supplementary Table S3. Fig. 4 presents the ridge regression coefficients and 95% confidence intervals for five core viral traits: genome type, replication site, proofreading polymerase, envelope status, and genome segmentation. The most prominent finding is the significant positive association between the presence of a proofreading polymerase and elevated EC₅ ₀ (reduced susceptibility). This supports existing mechanistic data showing that coronaviruses, which encode a viral exonuclease, are less susceptible to nucleoside analog mutagenesis due to their enhanced polymerase fidelity (Smith et al. 2013, Agostini et al. 2018, Ferron et al. 2018). Genome type (–ssRNA vs +ssRNA) also emerges as a meaningful predictor, with negative-sense RNA viruses generally showing higher EC₅ ₀ values. Envelope status, site of replication, and segmentation contribute marginal or inconsistent effects. Regression analysis not only confirms the biological observations seen in the raw data but also quantitatively demonstrates the importance of trait-based predictors for understanding and anticipating antiviral susceptibility. These insights highlight the role of viral genomics and replication mechanisms in shaping responses to ribavirin and related agents. 6. Potency vs. Cytotoxicity Landscape Fig. 5 visualizes the therapeutic window of each drug by plotting every assay as a point on a log– log scale, with EC₅ ₀ on the x-axis and CC₅₀ on the y-axis. Data for all three drugs are sourced from Tables 1, Supplementary Tables S1 and S2, and SI is encoded by point color (from red to blue, low to high SI). The majority of Ribavirin assays are located in the upper-right quadrant, indicating moderate potency (EC₅₀: 1–200 µM) and relatively high cytotoxicity thresholds (CC₅ ₀ : 100–1000 µM), resulting in intermediate SI. Favipiravir points cluster towards the top of the plot, reflecting extremely high CC₅₀ values and, thus, a wider therapeutic margin, despite more variable potency. Remdesivir points are found towards the lower left, reflecting outstanding potency in selected assays but also somewhat lower CC₅ ₀ values. Dashed reference lines for SI = 1 and SI = 10 make it easy to discern which assays are within a desirable therapeutic range. Page 10 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091 Page 10 of 13 - AI Writing Submission Submission ID trn:oid:::29034:97165091