United States National Library of Medicine

The SARS-CoV-2 pandemic has highlighted the need for society, as a whole, to be prepared against potential pandemics caused by a variety of different viral families of concern. Here, we describe a roadmap towards the identification and validation of conserved T cell epitope regions from Viral Families of Pandemic Potential (VFPP). For each viral family, we select a prototype virus, the sequence of which could be utilized in epitope identification screens. Examples of viral families considered and their respective prototypes (species/ subspecies) are Coronaviridae (Severe Acute Respiratory Syndrome-related Coronavirus/ SARS-CoV-2), Flaviviridae (Dengue virus/DENV2), Togaviridae (Chikungunya virus/ CHIKV), Paramyixoviridae (Morbillivirus/measles), Arenaviridae (Mammarenavirus/ Lassa), and Picornaviridae (Enterovirus C/poliovirus). The peptide sequences encoded in each prototype virus are then analyzed to determine their conservation across different viral taxonomic groups and viral variants derived from each of the VFPP. We outline available methodologies for epitope discovery based on panels of overlapping peptides and bioinformatics- based predictions of HLA-peptide binding, along with high-throughput in vitro assays, with emphasis on addressing coverage of the general worldwide population. Validation can be achieved by a variety of methodologies, including determining HLA restriction and recognition in samples from volunteers convalescent from previous infections or immunized with approved or experimental vaccines, and immunophenotyping of responding T cells. The capacity of these regions to induce crossreactive T cell responses can be tested experimentally with homologous peptides derived from the various viral species of interest. Importantly, they could be considered as a component of pan-viral family vaccines. Conversely, immunogenic regions that are highly specific to a given virus could be of interest for diagnostic applications.

Surface adhesion is critical to the survival of pathogenic bacteria both in natural niches and during infections, often via forming matrix-embedded communities called biofilms. Vibrio cholerae, the causal agent of pandemic cholera, is capable of forming biofilms adhering to both biotic and abiotic surfaces and the biofilm lifestyle has been implicated in promoting the survival of V. cholerae both in the natural reservoir and during host colonization. Previously, a 57-amino acid loop in the biofilm-specific adhesin Bap1 (Bap1-57aa) has been identified as a key contributor to the adhesion of V. cholerae biofilms to various surfaces including lipid membranes. However, the mechanism underlying its interaction with lipids, as well as its secondary structures, remain unresolved. Here, we combined biophysical, computational, and genetic approaches to elucidate the molecular mechanism of how this adhesive peptide interacts with lipids and lipid-coated surfaces. We found that a central aromatic-rich motif anchors the peptide to lipid bilayers while peripheral pseudo repeats enhance binding through avidity. Surprisingly, the core motif undergoes a lipid-induced conformational transition into a β-hairpin, enabling robust membrane insertion. We confirmed these findings both in vitro and in the biofilm context. Moreover, we demonstrated that the adhesive peptide can adhere to model host surfaces and is sensitive to membrane curvature. Finally, we show that the biofilm-derived peptide is found in several other Vibrio species, and its sequence is well-conserved. Our results provide molecular insight into biofilm adhesion and may lead to new strategies for targeted biofilm removal, as well as the design of bioinspired underwater adhesives.

Transformer models, neural networks that learn context by identifying relationships in sequential data, underpin many recent advances in artificial intelligence. Nevertheless, their inner workings are difficult to explain. Here, we find that a transformer model within the AlphaFold architecture uses simple, sparse patterns of amino acids to select protein conformations. To identify these patterns, we developed a straightforward algorithm called Conformational Attention Analysis Tool (CAAT). CAAT identifies amino acid positions that affect AlphaFold’s predictions substantially when modified. These effects are corroborated by experiments in several cases. By contrast, modifying amino acids ignored by CAAT affects AlphaFold predictions less, regardless of experimental ground truth. Our results demonstrate that CAAT successfully identifies the positions of some amino acids important for protein structure prediction, narrowing the search space required to predict effective mutations and suggesting a framework that can be applied to other transformer-based neural networks.

The known diversity of CRISPR-Cas systems continues to expand. To encompass new discoveries, here we present an updated evolutionary classification of CRISPR-Cas systems. The updated CRISPR-Cas classification includes 2 classes, 7 types and 46 subtypes, compared with the 6 types and 33 subtypes in our previous survey 5 years ago. In addition, a classification of the cyclic oligoadenylate-dependent signalling pathway in type III systems is presented. We also discuss recently characterized alternative CRISPR-Cas functionalities, notably, type IV variants that cleave the target DNA and type V variants that inhibit the target replication without cleavage. Analysis of the abundance of CRISPR-Cas variants in genomes and metagenomes shows that the previously defined systems are relatively common, whereas the more recently characterized variants are comparatively rare. These low abundance variants comprise the long tail of the CRISPR-Cas distribution in prokaryotes and their viruses, and remain to be characterized experimentally.

Clinical calculators play a vital role in healthcare, but their utilization is often hindered by usability and dissemination challenges. We introduce AgentMD, a novel language agent capable of curating and applying clinical calculators across various clinical contexts. As a tool builder, AgentMD first uses PubMed to curate a diverse set of 2,164 executable clinical calculators with over 85% accuracy for quality checks and over 90% pass rate for unit tests. As a tool user, AgentMD autonomously selects and applies the relevant clinical calculators. Our evaluations show that AgentMD significantly outperforms GPT-4 for risk prediction (87.7% vs. 40.9% in accuracy). Results on 698 real-world emergency department notes confirm that AgentMD accurately computes medical risks at an individual level. Moreover, AgentMD can provide population-level insights for institutional risk management. Our study illustrates the capabilities of language agents to curate and utilize clinical calculators for both individual patient care and at-scale healthcare analytics.

This article briefly documents the history and significance of PubMed Central (PMC) Journal Backfiles Digitization, 2004-2024 to raise awareness of this open access project among researchers who will find much to discover to advance understanding about the human condition across time and place. The success of PMC Journal Backfiles Digitization—including the interdisciplinary teamwork and partnerships underpinning it—provides a blueprint for future efforts to make the globally appreciated collections of the National Library of Medicine (NLM) accessible to all. By continuing to prioritize open access, teamwork, and partnerships, NLM and likeminded institutions can ensure that knowledge and data inform the advancement of medicine and public health.

Autonomous and multimodal AI agents have emerged as transformative tools in biomedical and biochemical research. These AI coscientists are capable of autonomously performing a range of complex tasks in the scientific research lifecycle, while interact

Concerns about potential prostate-related complications associated with testosterone replacement therapy (TRT) often lead hypogonadal men to remain untreated. This large-scale retrospective cohort study aimed to address these concerns by analyzing Medicare enrollment and claims data (Parts A/B/C/D) from 2007 to 2020 for men aged ≥65 with diagnosed primary or secondary hypogonadism. Cox regression analysis and 1:1 propensity score matching, both stratified by age group, were employed to evaluate the association between prostate outcomes and TRT use. TRT use was further categorized by treatment duration, administration route, and aromatase inhibitor (AI) coadministration. Following propensity score matching, the prostate cancer (PCa) and benign prostatic hyperplasia (BPH) risk analyses included 546 964 and 412 782 hypogonadal men, respectively, with comparable baseline characteristics at the time of hypogonadism diagnosis. TRT use was linked to a significant 16% reduction in the hazard of PCa (HR = 0.84; 95% CI: 0.82-0.86), including reductions observed across long-term (15%), short-term (16%), parenteral (12%), and topical (13%) use. Sensitivity analyses excluding individuals with elevated prostate-specific antigen and family history of PCa similarly demonstrated a 18% reduction in the hazard of PCa among TRT users. Conversely, TRT use exhibited a 13% increase in the hazard of BPH (HR = 1.13; 95% CI: 1.11-1.14), with a larger increase for parenteral (19%) vs topical TRT (12%). Notably, long-term topical use was linked to a modest 9% increase. TRT use was associated with reduction in PCa hazard and modest increase in BPH hazard. These associations varied by treatment duration, route, and AI coadministration. These findings may inform clinical decision-making regarding TRT use in hypogonadal men and support a more individualized approach to care.