(1) Background: Artificial Intelligence (AI) models have shown potential in various educational contexts. However, their utility in explaining complex biological phenomena, such as Intrinsically Disordered Proteins (IDPs), requires further exploration. This study empirically evaluated the performance of various Large Language Models (LLMs) in the educational domain of IDPs. (2) Methods: Four LLMs, GPT-3.5, GPT-4, GPT-4 with Browsing, and Google Bard (PaLM 2), were assessed using a set of IDP-related questions. An expert evaluated their responses across five categories: accuracy, relevance, depth of understanding, clarity, and overall quality. Descriptive statistics, ANOVA, and Tukey's honesty significant difference tests were utilized for analysis. (3) Results: The GPT-4 model consistently outperformed the others across all evaluation categories. Although GPT-4 and GPT-3.5 were not statistically significantly different in performance (p>0.05), GPT-4 was preferred as the best response in 13 out of 15 instances. The AI models with browsing capabilities, GPT-4 with Browsing and Google Bard (PaLM 2) displayed lower performance metrics across the board with statistically significant differences (p<0.0001). (4) Conclusion: Our findings underscore the potential of AI models, particularly LLMs such as GPT-4, in enhancing scientific education, especially in complex domains such as IDPs. Continued innovation and collaboration among AI developers, educators, and researchers are essential to fully harness the potential of AI for enriching scientific education.

Invasive fungal sinusitis (IFS) is associated with high rates of morbidity and mortality and often presents with orbital apex syndrome. Prompt diagnosis and management are crucial to prevent irreversible visual loss. We report a case of an immunosuppressed patient with rapidly progressive severe visual loss associated with frontal lobe cerebritis and leptomeningitis related to IFS, causing an adjacent compressive inflammatory optic neuropathy, which was treated successfully by large-dose corticosteroids. A 29-year-old woman with acute myeloid leukemia status post chemotherapy presented with right-sided headaches and periorbital swelling. Her examination was significant for subjective red desaturation and trace right eyelid edema and ptosis. The remainder of her initial ocular examination was normal. Her labs demonstrated neutropenia and thrombocytopenia. Imaging of the brain and orbits was concerning for extensive sinus disease with intracranial extension. An urgent multi-sinus and optic nerve decompression was performed given concern for compressive optic neuropathy, and the biopsy was consistent with invasive fungal infection. Despite aggressive antifungal treatment, vision in her right eye decreased rapidly to counting fingers. No optic nerve abnormalities were observed on serial MRIs, but adjacent inferior frontal lobe enhancement was present. After a vigorous debate in a multidisciplinary meeting, her severe vision loss was attributed to cerebritis causing an adjacent compressive inflammatory optic neuropathy, and large-dose intravenous (IV) steroid treatment was initiated while maintaining systemic antifungal therapy. Remarkably, she had a full recovery of her vision. Severe vision loss in IFS can occur due to compressive inflammatory optic neuropathy without direct fungal invasion as a contributing factor. Timely and effective intervention is crucial in preventing vision loss. Large-dose steroid therapy may be a potential treatment option for immunocompromised patients with invasive fungal sinusitis and intracranial invasion, provided strict fungal infection control measures are in place.

Introduction: The human retina relies on a complex network of proteins, many of which exhibit intrinsic disorder and liquid‐liquid phase separation (LLPS), enabling dynamic interactions for retinal function. Disruptions in these properties, along with missense mutations, have been linked to retinal diseases. This study aims to characterize and compare retinal proteins categorized by their expression specificity and tissue distribution using bioinformatics tools to explore relationships between intrinsic protein disorder, phase separation potential, and mutation pathogenicity.Methods: We analyzed retinal proteins classified by the Human Protein Atlas (HPA) into two major groups based on gene expression specificity (degree of unique retinal expression) and gene expression distribution (extent of expression across tissues). We analyzed nine retinal proteomes categorized by gene expression specificity and distribution. Intrinsic protein disorder was assessed using per‐residue and global disorder predictors from the Rapid Intrinsic Disorder Analysis Online (RIDAO) platform, LLPS potential was evaluated with ParSe v2, and missense mutation pathogenicity was predicted using AlphaMissense.Results: Significant differences in per‐residue intrinsic protein disorder were found within the specificity and distribution subgroups (p < 0.0001). In addition, global disorder predictions from the RIDAO platform demonstrated non‐random distributions of protein species across the proteomes analyzed in both subgroups (p < 0.0001). Furthermore, proteins specifically elevated in the retina exhibited higher intrinsic disorder and greater phase separation propensity (ParSe v2, AUC up to 0.650), compared to those more broadly expressed. Lastly, AlphaMissense analysis showed significant variations in the average pathogenicity scores of missense mutations within subgroups (p < 0.0001).Conclusion: Our results show that intrinsic disorder, LLPS, and mutational tendencies are not evenly distributed among retinal proteomes. Our study demonstrates a link between intrinsic disorder, LLPS potential, and pathogenic vulnerability among retinal proteins, underscoring the unique structural and functional landscape of retinal proteomes. Proteins with higher specificity to the retina exhibit greater disorder and phase separation potential, highlighting their potential role in dynamic cellular processes that support retinal function. Conversely, proteins widely distributed across multiple tissues tend to be more ordered, suggesting a need for structural stability in their broader functional roles.

BACKGROUND: Artificial Intelligence (AI) models have shown potential in various educational contexts. However, their utility in explaining complex biological phenomena, such as Intrinsically Disordered Proteins (IDPs), requires further exploration. This study empirically evaluated the performance of various Large Language Models (LLMs) in the educational domain of IDPs. METHODS: Four LLMs, GPT-3.5, GPT-4, GPT-4 with Browsing, and Google Bard (PaLM 2), were assessed using a set of IDP-related questions. An expert evaluated their responses across five categories: accuracy, relevance, depth of understanding, clarity, and overall quality. Descriptive statistics, ANOVA, and Tukey's honestly significant difference tests were utilized for analysis. RESULTS: The GPT-4 model consistently outperformed the others across all evaluation categories. Although GPT-4 and GPT-3.5 were not statistically significantly different in performance (p>0.05), GPT-4 was preferred as the best response in 13 out of 15 instances. The AI models with browsing capabilities, GPT-4 with Browsing and Google Bard (PaLM 2), displayed lower performance metrics across the board with statistically significant differences (p<0.0001). CONCLUSION: Our findings underscore the potential of AI models, particularly LLMs such as GPT-4, in enhancing scientific education, especially in complex domains such as IDPs. Continued innovation and collaboration among AI developers, educators, and researchers are essential to fully harness the potential of AI for enriching scientific education.