AUTHOR=Ju Yanxiu , Li Songtao , Kong Xiangyi , Zhao Qing 

TITLE=EBF1 is a potential biomarker for predicting progression from mild cognitive impairment to Alzheimer's disease: an in silico study

JOURNAL=Frontiers in Aging Neuroscience

VOLUME=Volume 16 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/aging-neuroscience/articles/10.3389/fnagi.2024.1397696

DOI=10.3389/fnagi.2024.1397696

ISSN=1663-4365

ABSTRACT=Introduction: Prediction of mild cognitive impairment (MCI) to Alzheimer's disease (AD) progression is an important clinical challenge. This study aimed to identify the independent risk factors and develop a nomogram model that can predict progression from MCI to AD.Methods: Data of 141 MCI patients were obtained from ADNI database. We set a follow-up time of 72 months, and defined patients as stable MCI and progressive MCI according to whether or not the conversion of MCI to AD occurred. By utilizing weighted gene co-expression network analysis, machine learning, and Cox proportionalhazards model , we identified and screened independent risk factors. Subsequently, we developed a nomogram model for predicting progression from MCI to AD. The performance of our nomogram was evaluated by C-index, calibration curve and decision curve analysis. Bioinformatics analysis and immune infiltration analysis were used to clarify the function of early B cell factor 1 (EBF1).Results: First, the results show that 40 differentially expressed genes related to the prognosis of MCI were generated by weighted gene co-expression network analysis. Second, 5 hub variables were obtained by machine learning. Third, low MoCA score (hazard ratio: 4.258, 95% confidence interval: 1.994-9.091) and low EBF1 expression (hazard ratio: 3.454, 95% confidence interval: 1.813-6.579) were the independent risk factors by Cox proportional-hazards regression analysis. Last, we developed a nomogram model with MoCA score, EBF1, and potential confounders (age and gender). By evaluating our nomogram model and validating it in both the internal and external validation sets, we demonstrated that our nomogram model exhibits excellent predictive performance. Through Gene ontology, Kyoto Encyclopedia of Genes Genomes functional enrichment analysis, and immune infiltration analysis, we found that the role of EBF1 in MCI was closely related to B cells.  Conclusion: EBF1, as a B cell-specific transcription factor, may be a key target for predicting progression from MCI to AD, and our nomogram model was able to provide personalized risk for progression from MCI to AD after evaluation and validation.