<?xml version="1.0" encoding="utf-8"?>
    <rss version="2.0">
      <channel xmlns:content="http://purl.org/rss/1.0/modules/content/">
        <title>Experimental Biology and Medicine | AI in Biology and Medicine section | New and Recent Articles</title>
        <link>https://www.ebm-journal.org/journals/experimental-biology-and-medicine/sections/ai-in-biology-and-medicine</link>
        <description>RSS Feed for AI in Biology and Medicine section in the Experimental Biology and Medicine journal | New and Recent Articles</description>
        <language>en-us</language>
        <generator>Frontiers Feed Generator,version:1</generator>
        <pubDate>2026-04-06T22:18:44.663+00:00</pubDate>
        <ttl>60</ttl>
        <item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10601</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10601</link>
        <title><![CDATA[Role of NAD metabolism-related genes in diabetic nephropathy: subtype classification, biomarker identification, and association with renal function]]></title>
        <pubdate>2026-01-26T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Shengnan Zeng</author><author>Yuhong Tao</author><author>Hui Guo</author>
        <description><![CDATA[Diabetic nephropathy (DN) remains a major complication of diabetes, significantly impacting renal function. Emerging evidence suggests that NAD metabolism plays a crucial role in DN pathogenesis. This study investigates the roles of NAD metabolism-related genes in DN and how there are associated with different disease subtypes. We analyzed gene expression data from DN-associated datasets (GSE30528 and GSE30529) to identify differences in NAD metabolism-related genes between normal and DN samples. We classified DN into subtypes based on NAD gene expression and evaluated NAD scores using ssGSEA. Immune cell infiltration and pathway analyses were assessed using ssGSEA, Microenvironment Cell Populations-counter (MCPcounter), and Gene Set Variation Analysis (GSVA). Key biomarker genes were identified using machine learning algorithms and validated across multiple datasets. We further explored the relationship between gene expression and kidney function using the Nephroseq V5 tool. Thirteen differentially expressed NAD metabolism-related genes were identified, with distinctive expression patterns observed between normal and DN samples. Two distinct NAD-related subtypes were classified, demonstrating significant differences in gene expression, immune cell infiltration, and pathway activities. Immune-related pathways and cellular processes exhibited varied enrichment between subtypes. Six key NAD metabolism-related genes (FMO3, ALDH1A3, FMO5, TKT, LBR, HPGD) were identified as potential biomarkers. Higher levels of FMO3, ALDH1A3, TKT, and LBR were linked to worse kidney function, while FMO5 and HPGD were associated with better kidney function. The study highlights the significant involvement of NAD metabolism-related genes in DN pathogenesis and their association with disease subtypes and renal function. The identified biomarkers could be targets for new treatments and provide insight for future DN research.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10612</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10612</link>
        <title><![CDATA[Bioinformatics-based screening and validation of ferroptosis-related genes in sepsis and type 2 diabetes mellitus]]></title>
        <pubdate>2025-10-21T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Heng Xiao</author><author>Zhonghua Ding</author><author>Cheng Liu</author><author>Xu He</author><author>Yanyan Tao</author>
        <description><![CDATA[Emerging clinical evidence underscores a bidirectional epidemiological linkage between sepsis and type 2 diabetes mellitus (T2DM). This study mechanistically investigates the underlying pathogenesis of this comorbidity, specifically focusing on the role of ferroptosis-related genes in its pathogenesis. A total of 1204 shared genes between sepsis and T2DM were screened using datasets from sepsis (GSE65682) and T2DM (GSE76894). GO and KEGG enrichment analyses, combined with WGCNA, were performed to identify key pathways and hub genes. Three signaling pathways—MAPK, adherens junction, and peroxisome—were significantly associated with the sepsis-T2DM interaction. Subsequent Pearson correlation analysis implicated ferroptosis as a critically involved process. Five core ferroptosis-related genes, including CDC25B, DPP7, FBXO31, PTCD3, and CNPY2, were were identified and experimentally validated using qRT-PCR. Furthermore, based on cMAP, we screened eight candidate drugs targeting these genes. Echinacea and Ibudilast were predicted to possess the greatest preclinical potential among them. This study provides a deeper insight into the shared pathogenesis of sepsis and T2DM, highlighting the pivotal role of ferroptosis in the development and progression of this comorbidity. Our findings offer preliminary insights into the sepsis-T2DM comorbidity, highlighting ferroptosis as a potential key pathological mechanism and identifying candidate targets for future therapeutic exploration.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10756</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10756</link>
        <title><![CDATA[High-fidelity, personalized cardiac modeling via AI-driven 3D reconstruction and embedded silicone rubber printing]]></title>
        <pubdate>2025-10-01T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Xuefang Wang</author><author>Yixin Li</author><author>Zhiqi Liang</author><author>Ruxu Du</author><author>Ting Song</author>
        <description><![CDATA[The burgeoning clinical demand for patient-specific cardiac modeling encounters significant challenges. The current clinical cardiac models are either difficult to manufacture or lack of detailed geometric structures and hence, often fail to incorporate important patient-specific characteristics. Moreover, most 3D-printable soft materials, such as Thermoplastic Poly-Urethane (TPU) or elastic resins, exhibit insufficient flexibility and biocompatibility to accurately mimic cardiac tissues, therefore limiting their ability to truly replicate patient-specific cardiac conditions. To address these limitations, we propose an innovative method for patient-specific cardiac substructure reconstruction based on the integration of Artificial Intelligence (AI) and embedded 3D printing. First, by combining medical imaging data (CT scan) with AI-driven high-precision 3D reconstruction algorithms, the new method segments the patient-specific cardiac structure into 10 substructures. The average Dice coefficient across the ten substructures is 0.87. Second, it uses an embedded 3D printing technique which utilizes silicone rubber matrix as supporting structure and uses diluted catalyst ink to extrude onto the supporting matrix. Through precise regulation of the matrix composition, material deposition rate and curing time, it can fabricate high-fidelity, complex 3D patient-specific silicone heart models with the average dimensional error less than 0.5 mm. The proposed method can substantially reduce manual intervention and post-processing time. The fabricated models provide valuable morphological insights for cardiovascular diagnosis and treatment planning. It is believed that many clinic applications will follow.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10555</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10555</link>
        <title><![CDATA[Pharmacovigilance in the digital age: gaining insight from social media data]]></title>
        <pubdate>2025-05-27T00:00:00Z</pubdate>
        <category>Review</category>
        <author>Fan Dong</author><author>Wenjing Guo</author><author>Jie Liu</author><author>Tucker A. Patterson</author><author>Huixiao Hong</author>
        <description><![CDATA[Pharmacovigilance is essential for protecting patient health by monitoring and managing medication-related risks. Traditional methods like spontaneous reporting systems and clinical trials are valuable for identifying adverse drug events, but face delays in data access. Social media platforms, with their real-time data, offer a novel avenue for pharmacovigilance by providing a wealth of user-generated content on medication usage, adverse drug events, and public sentiment. However, the unstructured nature of social media content presents challenges in data analysis, including variability and potential biases. Advanced techniques like natural language processing and machine learning are increasingly being employed to extract meaningful information from social media data, aiding in early adverse drug event detection and real-time medication safety monitoring. Ensuring data reliability and addressing ethical considerations are crucial in this context. This review examines the existing literature on the use of social media data for drug safety analysis, highlighting the platforms involved, methodologies applied, and research questions explored. It also discusses the challenges, limitations, and future directions of this emerging field, emphasizing the need for ethical principles, transparency, and interdisciplinary collaboration to maximize the potential of social media in enhancing pharmacovigilance efforts.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10445</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10445</link>
        <title><![CDATA[Optimal transport reveals immune perturbation and fingerprints over time in COVID-19 vaccination]]></title>
        <pubdate>2025-05-21T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Zexuan Wang</author><author>Jiong Chen</author><author>Matei Ionita</author><author>Qipeng Zhan</author><author>Zhuoping Zhou</author><author>Li Shen</author>
        <description><![CDATA[Mass cytometry enables high-throughput characterization of heterogeneous cell populations at single-cell resolution, using metal isotopes to capture cellular signals and avoiding the spectral overlap common in flow cytometry. Despite advancements, conventional data analysis often focuses on manual gating or clustering within specific samples, overlooking disparities across subjects or biological samples. To address this gap, we propose a novel framework that treats the cell-by-protein matrix as a high-dimensional distribution, using Quantized Optimal Transport (QOT) to quantify distances between samples based on their cellular protein expression profiles. This approach allows for a direct comparison of distributions without relying on predefined gating strategies, capturing subtle variations in the data. We validated our method through two experiments using real-world time-series Coronavirus Disease 2019 (COVID-19) cytometry data. First, we conducted a leave-one-out analysis to identify immunologically unstable proteins over time, revealing CD3 and CD45 as the proteins changing the most during the vaccine response. Second, we aimed to capture individual immune fingerprints over time by calculating pairwise Wasserstein distances between samples and applying hierarchical clustering. Using silhouette scores to evaluate clustering effectiveness, we identified optimal combinations of immunological markers that effectively grouped samples from the same participant across different time points. Our findings demonstrate that the QOT framework provides a robust and flexible tool for cohort-level analysis of mass cytometry data, enabling the identification of unstable immunological markers and capturing immune response heterogeneity among vaccinated cohorts.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10374</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10374</link>
        <title><![CDATA[A refined set of RxNorm drug names for enhancing unstructured data analysis in drug safety surveillance]]></title>
        <pubdate>2025-05-02T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Wenjing Guo</author><author>Fan Dong</author><author>Jie Liu</author><author>Aasma Aslam</author><author>Tucker A. Patterson</author><author>Huixiao Hong</author>
        <description><![CDATA[Adverse drug events are harms associated with drug use, whether the drug is used correctly or incorrectly. Identifying adverse drug events is vital in pharmacovigilance to safeguard public health. Drug safety surveillance can be performed using unstructured data. A comprehensive and accurate list of drug names is essential for effective identification of adverse drug events. While there are numerous sources for drug names, RxNorm is widely recognized as a leading resource. However, its effectiveness for unstructured data analysis in drug safety surveillance has not been thoroughly assessed. To address this, we evaluated the drug names in RxNorm for their suitability in unstructured data analysis and developed a refined set of drug names. Initially, we removed duplicates, the names exceeding 199 characters, and those that only describe administrative details. Drug names with four or fewer characters were analyzed using 18,000 drug-related PubMed abstracts to remove names which rarely appear in unstructured data. The remaining names, which ranged from five to 199 characters, were further refined to exclude those that could lead to inaccurate drug counts in unstructured data analysis. We compared the efficiency and accuracy of the refined set with the original RxNorm set by testing both on the 18,000 drug-related PubMed abstracts. The results showed a decrease in both computational cost and the number of false drug names identified. Further analysis of the removed names revealed that most originated from only one of the 14 sources. Our findings suggest that the refined set can enhance drug identification in unstructured data analysis, thereby improving pharmacovigilance.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10238</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10238</link>
        <title><![CDATA[Artificial intelligence for children with attention deficit/hyperactivity disorder: a scoping review]]></title>
        <pubdate>2025-04-24T00:00:00Z</pubdate>
        <category>Review</category>
        <author>Bo Sun</author><author>Fei Cai</author><author>Huiman Huang</author><author>Bo Li</author><author>Bing Wei</author>
        <description><![CDATA[Attention deficit/hyperactivity disorder is a common neuropsychiatric disorder that affects around 5%–7% of children worldwide. Artificial intelligence provides advanced models and algorithms for better diagnosis, prediction and classification of attention deficit/hyperactivity disorder. This study aims to explore artificial intelligence models used for the prediction, early diagnosis and classification of attention deficit/hyperactivity disorder as reported in the literature. A scoping review was conducted and reported in line with the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines. Out of the 1994 publications, 52 studies were included in the scoping review. The included articles reported the use of artificial intelligence for 3 different purposes. Of these included articles, artificial intelligence techniques were mostly used for the diagnosis of attention deficit/hyperactivity disorder (38/52, 79%). Magnetic resonance imaging (20/52, 38%) were the most frequently used data in the included articles. Most of the included articles used data sets with a size of <1,000 samples (28/52, 54%). Machine learning models were the most prominent branch of artificial intelligence used for attention deficit/hyperactivity disorder in the studies, and the support vector machine was the most used algorithm (34/52, 65%). The most commonly used validation in the studies was k-fold cross-validation (34/52, 65%). A higher level of accuracy (98.23%) was found in studies that used Convolutional Neural Networks algorithm. This review provides an overview of research on artificial intelligence models and algorithms for attention deficit/hyperactivity disorder, providing data for further research to support clinical decision-making in healthcare.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10359</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10359</link>
        <title><![CDATA[Developing predictive models for µ opioid receptor binding using machine learning and deep learning techniques]]></title>
        <pubdate>2025-03-19T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Jie Liu</author><author>Jerry Li</author><author>Zoe Li</author><author>Fan Dong</author><author>Wenjing Guo</author><author>Weigong Ge</author><author>Tucker A. Patterson</author><author>Huixiao Hong</author>
        <description><![CDATA[Opioids exert their analgesic effect by binding to the µ opioid receptor (MOR), which initiates a downstream signaling pathway, eventually inhibiting pain transmission in the spinal cord. However, current opioids are addictive, often leading to overdose contributing to the opioid crisis in the United States. Therefore, understanding the structure-activity relationship between MOR and its ligands is essential for predicting MOR binding of chemicals, which could assist in the development of non-addictive or less-addictive opioid analgesics. This study aimed to develop machine learning and deep learning models for predicting MOR binding activity of chemicals. Chemicals with MOR binding activity data were first curated from public databases and the literature. Molecular descriptors of the curated chemicals were calculated using software Mold2. The chemicals were then split into training and external validation datasets. Random forest, k-nearest neighbors, support vector machine, multi-layer perceptron, and long short-term memory models were developed and evaluated using 5-fold cross-validations and external validations, resulting in Matthews correlation coefficients of 0.528–0.654 and 0.408, respectively. Furthermore, prediction confidence and applicability domain analyses highlighted their importance to the models’ applicability. Our results suggest that the developed models could be useful for identifying MOR binders, potentially aiding in the development of non-addictive or less-addictive drugs targeting MOR.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10389</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10389</link>
        <title><![CDATA[AI-powered topic modeling: comparing LDA and BERTopic in analyzing opioid-related cardiovascular risks in women]]></title>
        <pubdate>2025-02-28T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Li Ma</author><author>Ru Chen</author><author>Weigong Ge</author><author>Paul Rogers</author><author>Beverly Lyn-Cook</author><author>Huixiao Hong</author><author>Weida Tong</author><author>Ningning Wu</author><author>Wen Zou</author>
        <description><![CDATA[Topic modeling is a crucial technique in natural language processing (NLP), enabling the extraction of latent themes from large text corpora. Traditional topic modeling, such as Latent Dirichlet Allocation (LDA), faces limitations in capturing the semantic relationships in the text document although it has been widely applied in text mining. BERTopic, created in 2022, leveraged advances in deep learning and can capture the contextual relationships between words. In this work, we integrated Artificial Intelligence (AI) modules to LDA and BERTopic and provided a comprehensive comparison on the analysis of prescription opioid-related cardiovascular risks in women. Opioid use can increase the risk of cardiovascular problems in women such as arrhythmia, hypotension etc. 1,837 abstracts were retrieved and downloaded from PubMed as of April 2024 using three Medical Subject Headings (MeSH) words: “opioid,” “cardiovascular,” and “women.” Machine Learning of Language Toolkit (MALLET) was employed for the implementation of LDA. BioBERT was used for document embedding in BERTopic. Eighteen was selected as the optimal topic number for MALLET and 23 for BERTopic. ChatGPT-4-Turbo was integrated to interpret and compare the results. The short descriptions created by ChatGPT for each topic from LDA and BERTopic were highly correlated, and the performance accuracies of LDA and BERTopic were similar as determined by expert manual reviews of the abstracts grouped by their predominant topics. The results of the t-SNE (t-distributed Stochastic Neighbor Embedding) plots showed that the clusters created from BERTopic were more compact and well-separated, representing improved coherence and distinctiveness between the topics. Our findings indicated that AI algorithms could augment both traditional and contemporary topic modeling techniques. In addition, BERTopic has the connection port for ChatGPT-4-Turbo or other large language models in its algorithm for automatic interpretation, while with LDA interpretation must be manually, and needs special procedures for data pre-processing and stop words exclusion. Therefore, while LDA remains valuable for large-scale text analysis with resource constraints, AI-assisted BERTopic offers significant advantages in providing the enhanced interpretability and the improved semantic coherence for extracting valuable insights from textual data.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10444</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10444</link>
        <title><![CDATA[Artificial intelligence in the diagnosis of uveal melanoma: advances and applications]]></title>
        <pubdate>2025-02-19T00:00:00Z</pubdate>
        <category>Review</category>
        <author>Albert K. Dadzie</author><author>Sabrina P. Iddir</author><author>Sanjay Ganesh</author><author>Behrouz Ebrahimi</author><author>Mojtaba Rahimi</author><author>Mansour Abtahi</author><author>Taeyoon Son</author><author>Michael J. Heiferman</author><author>Xincheng Yao</author>
        <description><![CDATA[Advancements in machine learning and deep learning have the potential to revolutionize the diagnosis of melanocytic choroidal tumors, including uveal melanoma, a potentially life-threatening eye cancer. Traditional machine learning methods rely heavily on manually selected image features, which can limit diagnostic accuracy and lead to variability in results. In contrast, deep learning models, particularly convolutional neural networks (CNNs), are capable of automatically analyzing medical images, identifying complex patterns, and enhancing diagnostic precision. This review evaluates recent studies that apply machine learning and deep learning approaches to classify uveal melanoma using imaging modalities such as fundus photography, optical coherence tomography (OCT), and ultrasound. The review critically examines each study’s research design, methodology, and reported performance metrics, discussing strengths as well as limitations. While fundus photography is the predominant imaging modality being used in current research, integrating multiple imaging techniques, such as OCT and ultrasound, may enhance diagnostic accuracy by combining surface and structural information about the tumor. Key limitations across studies include small dataset sizes, limited external validation, and a reliance on single imaging modalities, all of which restrict model generalizability in clinical settings. Metrics such as accuracy, sensitivity, and area under the curve (AUC) indicate that deep learning models have the potential to outperform traditional methods, supporting their further development for integration into clinical workflows. Future research should aim to address current limitations by developing multimodal models that leverage larger, diverse datasets and rigorous validation, thereby paving the way for more comprehensive, reliable diagnostic tools in ocular oncology.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2025.10399</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2025.10399</link>
        <title><![CDATA[MONet: cancer driver gene identification algorithm based on integrated analysis of multi-omics data and network models]]></title>
        <pubdate>2025-02-04T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Yingzan Ren</author><author>Tiantian Zhang</author><author>Jian Liu</author><author>Fubin Ma</author><author>Jiaxin Chen</author><author>Ponian Li</author><author>Guodong Xiao</author><author>Chuanqi Sun</author><author>Yusen Zhang</author>
        <description><![CDATA[Cancer progression is orchestrated by the accrual of mutations in driver genes, which endow malignant cells with a selective proliferative advantage. Identifying cancer driver genes is crucial for elucidating the molecular mechanisms of cancer, advancing targeted therapies, and uncovering novel biomarkers. Based on integrated analysis of Multi-Omics data and Network models, we present MONet, a novel cancer driver gene identification algorithm. Our method utilizes two graph neural network algorithms on protein-protein interaction (PPI) networks to extract feature vector representations for each gene. These feature vectors are subsequently concatenated and fed into a multi-layer perceptron model (MLP) to perform semi-supervised identification of cancer driver genes. For each mutated gene, MONet assigns the probability of being potential driver, with genes identified in at least two PPI networks selected as candidate driver genes. When applied to pan-cancer datasets, MONet demonstrated robustness across various PPI networks, outperforming baseline models in terms of both the area under the receiver operating characteristic curve and the area under the precision-recall curve. Notably, MONet identified 37 novel driver genes that were missed by other methods, including 29 genes such as APOBEC2, GDNF, and PRELP, which are corroborated by existing literature, underscoring their critical roles in cancer development and progression. Through the MONet framework, we successfully identified known and novel candidate cancer driver genes, providing biologically meaningful insights into cancer mechanisms.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2024.10341</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2024.10341</link>
        <title><![CDATA[Leveraging AI to improve disease screening among American Indians: insights from the Strong Heart Study]]></title>
        <pubdate>2025-01-08T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Paul Rogers</author><author>Thomas McCall</author><author>Ying Zhang</author><author>Jessica Reese</author><author>Dong Wang</author><author>Weida Tong</author>
        <description><![CDATA[Screening tests for disease have their performance measured through sensitivity and specificity, which inform how well the test can discriminate between those with and without the condition. Typically, high values for sensitivity and specificity are desired. These two measures of performance are unaffected by the outcome prevalence of the disease in the population. Research projects into the health of the American Indian frequently develop Machine learning algorithms as predictors of conditions in this population. In essence, these models serve as in silico screening tests for disease. A screening test’s sensitivity and specificity values, typically determined during the development of the test, inform on the performance at the population level and are not affected by the prevalence of disease. A screening test’s positive predictive value (PPV) is susceptible to the prevalence of the outcome. As the number of artificial intelligence and machine learning models flourish to predict disease outcomes, it is crucial to understand if the PPV values for these in silico methods suffer as traditional screening tests in a low prevalence outcome environment. The Strong Heart Study (SHS) is an epidemiological study of the American Indian and has been utilized in predictive models for health outcomes. We used data from the SHS focusing on the samples taken during Phases V and VI. Logistic Regression, Artificial Neural Network, and Random Forest were utilized as in silico screening tests within the SHS group. Their sensitivity, specificity, and PPV performance were assessed with health outcomes of varying prevalence within the SHS subjects. Although sensitivity and specificity remained high in these in silico screening tests, the PPVs’ values declined as the outcome’s prevalence became rare. Machine learning models used as in silico screening tests are subject to the same drawbacks as traditional screening tests when the outcome to be predicted is of low prevalence.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2024.10215</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2024.10215</link>
        <title><![CDATA[Integrating machine learning with bioinformatics for predicting idiopathic pulmonary fibrosis prognosis: developing an individualized clinical prediction tool]]></title>
        <pubdate>2024-12-23T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Hongmei Ruan</author><author>Chunnian Ren</author>
        <description><![CDATA[Idiopathic pulmonary fibrosis (IPF) is a chronic interstitial lung disease with a poor prognosis. Its non-specific clinical symptoms make accurate prediction of disease progression challenging. This study aimed to develop molecular-level prognostic models to personalize treatment strategies for IPF patients. Using transcriptome sequencing and clinical data from 176 IPF patients, we developed a Random Survival Forest (RSF) model through machine learning and bioinformatics techniques. The model demonstrated superior predictive accuracy and clinical utility, as shown by the concordance index (C-index), the area under the operating characteristic curve (AUC), Brief scores, and decision curve analysis (DCA) curves. Additionally, a novel prognostic staging system was introduced to stratify IPF patients into distinct risk groups, enabling individualized predictions. The model’s performance was validated using a bleomycin-induced pulmonary fibrosis mouse model. In conclusion, this study offers a new prognostic staging system and predictive tool for IPF, providing valuable insights for treatment and management.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2024.10393</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2024.10393</link>
        <title><![CDATA[Enhancing pharmacogenomic data accessibility and drug safety with large language models: a case study with Llama3.1]]></title>
        <pubdate>2024-12-03T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Dan Li</author><author>Leihong Wu</author><author>Ying-Chi Lin</author><author>Ho-Yin Huang</author><author>Ebony Cotton</author><author>Qi Liu</author><author>Ru Chen</author><author>Ruihao Huang</author><author>Yifan Zhang</author><author>Joshua Xu</author>
        <description><![CDATA[Pharmacogenomics (PGx) holds the promise of personalizing medical treatments based on individual genetic profiles, thereby enhancing drug efficacy and safety. However, the current landscape of PGx research is hindered by fragmented data sources, time-consuming manual data extraction processes, and the need for comprehensive and up-to-date information. This study aims to address these challenges by evaluating the ability of Large Language Models (LLMs), specifically Llama3.1-70B, to automate and improve the accuracy of PGx information extraction from the FDA Table of Pharmacogenomic Biomarkers in Drug Labeling (FDA PGx Biomarker table), which is well-structured with drug names, biomarkers, therapeutic area, and related labeling texts. Our primary goal was to test the feasibility of LLMs in streamlining PGx data extraction, as an alternative to traditional, labor-intensive approaches. Llama3.1-70B achieved 91.4% accuracy in identifying drug-biomarker pairs from single labeling texts and 82% from mixed texts, with over 85% consistency in aligning extracted PGx categories from FDA PGx Biomarker table and relevant scientific abstracts, demonstrating its effectiveness for PGx data extraction. By integrating data from diverse sources, including scientific abstracts, this approach can support pharmacologists, regulatory bodies, and healthcare researchers in updating PGx resources more efficiently, making critical information more accessible for applications in personalized medicine. In addition, this approach shows potential of discovering novel PGx information, particularly of underrepresented minority ethnic groups. This study highlights the ability of LLMs to enhance the efficiency and completeness of PGx research, thus laying a foundation for advancements in personalized medicine by ensuring that drug therapies are tailored to the genetic profiles of diverse populations.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2024.10320</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2024.10320</link>
        <title><![CDATA[Integrating artificial intelligence in strabismus management: current research landscape and future directions]]></title>
        <pubdate>2024-11-25T00:00:00Z</pubdate>
        <category>Review</category>
        <author>Dawen Wu</author><author>Xi Huang</author><author>Liang Chen</author><author>Peixian Hou</author><author>Longqian Liu</author><author>Guoyuan Yang</author>
        <description><![CDATA[Advancements in artificial intelligence (AI) are transforming strabismus management through improved screening, diagnosis, and surgical planning. Deep learning has notably enhanced diagnostic accuracy and optimized surgical outcomes. Despite these advancements, challenges such as the underrepresentation of diverse strabismus types and reliance on single-source data remain prevalent. Emphasizing the need for inclusive AI systems, future research should focus on expanding AI capabilities with large model technologies, integrating multimodal data to bridge existing gaps, and developing integrated management platforms to better accommodate diverse patient demographics and clinical scenarios.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.ebm-journal.org/articles/10.3389/ebm.2024.10279</guid>
        <link>https://www.ebm-journal.org/articles/10.3389/ebm.2024.10279</link>
        <title><![CDATA[Development of a comprehensive open access “molecules with androgenic activity resource (MAAR)” to facilitate risk assessment of chemicals]]></title>
        <pubdate>2024-09-19T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Fan Dong</author><author>Barry Hardy</author><author>Jie Liu</author><author>Tomaz Mohoric</author><author>Wenjing Guo</author><author>Thomas Exner</author><author>Weida Tong</author><author>Joh Dohler</author><author>Daniel Bachler</author><author>Huixiao Hong</author>
        <description><![CDATA[The increasing prevalence of endocrine-disrupting chemicals (EDCs) and their potential adverse effects on human health underscore the necessity for robust tools to assess and manage associated risks. The androgen receptor (AR) is a critical component of the endocrine system, playing a pivotal role in mediating the biological effects of androgens, which are male sex hormones. Exposure to androgen-disrupting chemicals during critical periods of development, such as fetal development or puberty, may result in adverse effects on reproductive health, including altered sexual differentiation, impaired fertility, and an increased risk of reproductive disorders. Therefore, androgenic activity data is critical for chemical risk assessment. A large amount of androgenic data has been generated using various experimental protocols. Moreover, the data are reported in different formats and in diverse sources. To facilitate utilization of androgenic activity data in chemical risk assessment, the Molecules with Androgenic Activity Resource (MAAR) was developed. MAAR is the first open-access platform designed to streamline and enhance the risk assessment of chemicals with androgenic activity. MAAR’s development involved the integration of diverse data sources, including data from public databases and mining literature, to establish a reliable and versatile repository. The platform employs a user-friendly interface, enabling efficient navigation and extraction of pertinent information. MAAR is poised to advance chemical risk assessment by offering unprecedented access to information crucial for evaluating the androgenic potential of a wide array of chemicals. The open-access nature of MAAR promotes transparency and collaboration, fostering a collective effort to address the challenges posed by androgenic EDCs.]]></description>
      </item>
      </channel>
    </rss>